Download Document79388 (79388.pdf)
Text Digital, Culture, Media and Sport
Oral evidence: Fake News, HC 363
Tuesday 27 February 2018
Ordered by the House of Commons to be published on 27 February 2018.
Watch the meeting
Members present: Damian Collins (Chair); Paul Farrelly; Simon Hart; Julian
Knight; Ian C. Lucas; Christian Matheson; Rebecca Pow; Giles Watling.
I: Alexander Nix, Chief Executive, Cambridge Analytica
Written evidence from witnesses:
- Cambridge Analytica
Examination of witnesses
Alexander Nix, Chief Executive, Cambridge Analytica
Q621 Chair: Good morning. Welcome to Alexander Nix to this further session
of the Digital, Culture, Media and Sport Committee and our inquiry on
fake news and disinformation. During the course of our investigation so
far, it has been clear to us that understanding of data analytics and
behavioural patterns online is key to understanding the way that
messaging works. That is why we were particularly keen to talk to
Cambridge Analytica, which is one of the leading companies in the
world—I think it would be fair to say—in understanding the way in which
data analytics and behaviour activity works online.
We have a range of questions that we want to ask you, Mr Nix, about
that. We are also this morning publishing the letter you sent to me on 23
February following the evidence session we had in Washington, in which
you raised some concerns about things that were said about Cambridge
Analytica at that hearing that you wanted to correct for the record. I am
publishing that letter this morning, but I wanted to start by asking a
couple of questions relating to that letter and to clear up a few things.
One of the issues you raised in the letter in response to Mr Matheson's
question was to state that Cambridge Analytica had never had any
involvement in the EU referendum campaign. To quote for the record for
people who may not have seen the letter, you say, "Cambridge Analytica
had no involvement in the referendum, was not retained by any
campaign, and did not provide any services (paid or unpaid) to any
campaign." That is what you said in your letter. We are publishing that
today but I wanted to be clear. You will understand why this confusion
has arisen about Cambridge Analytica's role, because there have been
public statements made by you that did associate you with the
referendum campaign and with Leave.EU in particular. Why are those
previous statements not correct but what you say now correct instead?
Alexander Nix: Let me start by saying thank you for the invitation to
come and speak to this Committee. Fake news is a credible threat to the
public and indeed to the industry that we work in, and we are pleased to
try to help in any way that we can.
We have been very consistent for the last two years about our
involvement, or lack of involvement, in the EU referendum. There was
one statement only, which was put out erroneously, that indicated that
we were involved in the campaign. It was drafted by a slightly
overzealous PR consultant who worked for us, and referenced work that
we hoped and intended to undertake for the campaign. Subsequently,
work was never undertaken. The moment that that statement went out
we were absolutely crystal clear to all the media outlets that we were not
involved and that it had been released in error, and we tried to correct
the press again and again and again. Unfortunately, and somewhat
ironically, this was an example of fake news that was disseminated and
spun out virally. By the time it had penetrated the internet it became a
matter of fact even though there was no fact behind it.
Q622 Chair: So that I am clear, is this the statement? I will read from it. Tell
me if this is the statement you are referring to: "Recently Cambridge
Analytica has teamed up with Leave.EU—the UK's largest group
advocating for a British exit... from the European Union—to help them
better understand and communicate with UK voters. We have already
helped supercharge Leave.EU's social media campaign by ensuring the
right messages get to the right voters online". Is that the statement you
are referring to?
Alexander Nix: That is the statement, and this was a statement that
was prepared in anticipation of working with that organisation and was
released, unfortunately, ahead of any work starting. Again, it was an
error. We were very vocal about that at the time and we addressed it
head-on immediately when we realised that it had been put out.
Q623 Chair: I have seen that quotation in an article for Campaign magazine,
which is in your name.
Alexander Nix: That is where it was put out.
Chair: Yes. It is an article in your name and it is still on their website
today, so why have you not asked them to withdraw it if that statement
was put out in error?
Alexander Nix: I cannot speak to that personally, but I am sure that we
have asked them. I can double-check for you.
Q624 Chair: I think we all understand that sometimes an over-eager press
officer might say the wrong thing. It is quite different when the leading
figure in a company signs off an article that goes out in their name and
the key fact in that argument is wrong. It does not just refer to an
anticipated relationship. It says that you have already worked for them,
"We have already helped supercharge Leave.EU's social media"
messaging and, in particular, references the growth in the Facebook page
for the campaign. Presumably that refers to work that has been done not
just things you hope to do.
Alexander Nix: While your point is valid—we have addressed it head-on
again and again—the facts of the matter are that we did no work on that
campaign or any campaigns. We were not involved in the referendum.
While we could dwell on this, I think we should probably look at the facts
of the matter, which are that we were not involved, period.
Q625 Chair: What you are saying is clear. Unfortunately, the question will still
keep coming up because people will reference this and think it is odd that
a statement was put out that was totally untrue, when it refers to not just
work you hope to do but work that you have already done. This will not
be news to you, but in the Newsnight programme that you were
interviewed in, they had had footage of a Cambridge Analytica employee
sitting in a press conference with Leave.EU. It was Brittany Kaiser, and
she talked about the fact that she would be working on running large-
scale research for Leave.EU. That was in 2015. Was that work
undertaken? Did she do that as a Cambridge Analytica employee or was
that done in a personal capacity?
Alexander Nix: It is not unusual, when you are exploring a working
relationship with a client, to speak in public together about the work that
you hope to undertake. That was simply an example of that.
Q626 Chair: She was talking about work that they hoped to do, but that work
was not done.
Alexander Nix: Exactly right.
Q627 Chair: When we talk about you or your organisation, when you say that
work was not done and there was never any work done, does that apply
to not just Cambridge Analytica but all your affiliate companies and
companies in your group as well?
Alexander Nix: That is absolutely right. No company that falls under any
of the group vehicles in Cambridge Analytica or SCL or any other
company that we are involved with has worked on the EU referendum.
Chair: Any associates or anyone?
Alexander Nix: Or any associates.
Q628 Chair: In April 2017, Andy Wigmore, the communications director at
Leave.EU, put out a tweet in response to some news from the
Conservative party about the people it had hired to advise it on the last
general election for its digital campaign. He says, "You should use
Cambridge Analytics—we did apparently can highly recommend them".
Why would he have said that?
Alexander Nix: You are going to have to speak to Andy about that. I
cannot begin to second-guess why he would have said that. My
understanding is that he subsequently changed that statement, but,
again, you would have to speak to him.
Q629 Chair: He also put out another tweet saying, "Leave.EU campaign brings
in US voter data and messaging firm Cambridge Analytica". That was a
Alexander Nix: I do not know the date of that, but I can only assume
that at the time he was vying to be the designated leave campaign and
that by associating himself with a data analytics firm such as ours, which
had quite a high profile for our work in the United States in the US
presidential primaries, he was hoping that would give him additional
credibility through association.
Q630 Chair: That was in November 2015 and that was presumably after the
press conference that Brittany Kaiser took part in as well. Again, that
would suggest that there was a working relationship between Cambridge
Analytica and Leave.EU at that time.
Alexander Nix: I do not know how to explain this to you more clearly:
we did not work with them. However you look at this or however it
appears to you or whatever tweets other people have said about the
situation, we did no paid or unpaid work. We had no formalised
relationship with them. We did not work on the EU referendum with that
organisation or any other organisation.
Q631 Chair: The reason I think it is important that we ask these questions is
that we are publishing a written statement from you that seems to
correct the record on this point. The reason the questions keep coming
up is that what you have said today is clearly challenged by what you
have said in the past, or statements that have gone out in your name in
the past, and what people like Andy Wigmore have said and what other
employees at Cambridge Analytica have said in the past as well. We are
now being asked to believe that the version of reality that was portrayed
at the end of 2015 and 2016 is false, and the current statement is that
there was no work of any kind done by either Cambridge Analytica or any
associates during the referendum. They are at such odds it is not
unreasonable that these questions keep coming up.
Alexander Nix: You are looking at that in isolation. As I said before, that
press release went out in error. After it went out, we were very quick to
go to the press and to correct it and to say to them, "This was a mistake.
For the record, we are not doing any work. We have not been retained or
contracted by any of these organisations". We consistently put this
message out over a two-year period. One press release you are referring
to was instantly corrected, and we have been consistent in our messaging
ever since, so I do not think your line of inquiry is entirely fair.
Q632 Chair: It is an article, not a press release, in your name, and it is still on
the website of the organisation that published it, in your name. It has not
been taken down.
Alexander Nix: That is out of our control, clearly.
Chair: You could have made a request to them.
Alexander Nix: We have made several requests to leading newspaper
publications to retract statements that we have been involved with this.
We have told news outlets, and we have put out our own press releases,
but unfortunately we are not always successful in these entreaties.
Q633 Chair: It is normal for companies, when they are pitching for work—and
from what you said it sounded like you were in the process of pitching for
work for Leave.EU even if it did not come about. Probably a fair
interpretation of the article we have been discussing is that you
anticipated that you were going to be hired to do some work for them
and that did not happen. What sort of work was done in order to pitch?
Normally you go out and see prospective clients and you pitch to them
and show them what you can do and the value you could add if you were
Alexander Nix: That is exactly right. We have a political division. It is
not uncommon for us to go and speak to political parties. Indeed, in this
country I think I have spoken with every political party—or at least been
approached by Labour, Liberal Democrats, UKIP, SNP, Conservatives—on
how we might be able to help them with their campaigns, various
different campaigns, and to present our services, talk about our track
record, our extensive 27-year history in managing election campaigns
around the world, the technologies that we have developed to help
campaigning and make it more efficient and then to talk about how our
services might be most relevant to the clients that we are seeking to
assist. I think that is pretty common practice.
Q634 Chair: I used to work in the advertising industry many years ago, part of
the "Mad Men" style mass messaging industry that you say is now dead.
What was normal there is that you would produce draft campaigns. You
would say, "If you hired us, these are some of the advertisements that
we would run for you." Given that what you do is for many people, and
probably for many of your clients, quite a new area of activity, do you
create demonstrations, saying, "This is a sample of the sorts of work we
would do for you based on our understanding of this issue and the
understanding of your audience. Here are some examples of the work we
would do if we were hired"?
Alexander Nix: Unlike the "Mad Men" days of advertising where it is
creative-led, so you can draw on the imagination to come up with these
sorts of examples, our communications are rooted in data and in science.
As a result, in order to produce these things there is considerably more
time and effort and work involved and we also need access to the
appropriate datasets. It would be almost impossible for us to provide a
client with a meaningful demonstration of what we might be able to do
for them unless we have access to their data and have spent a lot of time
modelling. More often—I would say that this is always is the case—we will
simply show them work from other projects that we have worked on, to
give them an understanding of the sort of work that we might be able to
deliver to them.
Q635 Chair: In your discussions with Leave.EU, did they say that they had a
dataset that they could make available to you in order to assist targeting
in that campaign?
Alexander Nix: I am not sure that they did have a huge dataset or any
dataset. I would have to revert to you on that. I think the idea was that
we would help them go out and capture their data for them.
Q636 Ian C. Lucas: I have a quotation in front of me, dated 8 February 2017,
from Bloomberg Businessweek: "We did undertake some work with
Leave.eu, but it's been significantly over-reported". Are you saying that
that is not correct, you never said that?
Alexander Nix: What I am saying is that the work we undertook was
exploring a business relationship together.
Ian C. Lucas: You explored a business relationship but you did not begin
a business relationship?
Alexander Nix: That is correct, sir.
Q637 Ian C. Lucas: Do you know who Arron Banks is?
Alexander Nix: I do know who Arron Banks is.
Q638 Ian C. Lucas: Have you read this book?
Alexander Nix: I know I have not.
Q639 Ian C. Lucas: It is called The Bad Boys of Brexit and it was sent to me
by Arron Banks. Do you have a copy in your office?
Alexander Nix: That is correct. I was given a copy as well.
Q640 Ian C. Lucas: Can I suggest you read it, Mr Nix, because on 22 October
2015, according to this book, Mr Bank says, "We have hired Cambridge
Analytica, an American company that uses 'big data and advanced
psychographics' to influence people". Are you saying that is incorrect?
Alexander Nix: I am saying that is incorrect.
Q641 Ian C. Lucas: Were you aware of that statement?
Alexander Nix: I saw the statement in the book.
Ian C. Lucas: You said you had not read it.
Alexander Nix: I have not read the book. I have seen the pages
relevant to Cambridge Analytica.
Q642 Ian C. Lucas: You are aware of that statement.
Alexander Nix: Yes, I am aware of that statement.
Q643 Ian C. Lucas: Do you think that improves the business reputation of
Alexander Nix: Unfortunately, that is something that is out of our
control. We have spoken to Mr Banks about this statement, and we spoke
to Mr Wigmore about some of the statements that he made. We told
them that we disagreed with them and that they were not true. I believe
that they retracted some of their statements. The book came out, and it
was already published by time I knew that that statement was going to
be included in it. There was very little that I could do at the time to
Q644 Ian C. Lucas: You could have sued, couldn't you? You could have sued if
it was damaging to the reputation of Cambridge Analytica.
Alexander Nix: I could have but I did not think that was adequate use
of time and resources.
Q645 Ian C. Lucas: What he says is not true?
Alexander Nix: That is not true.
Q646 Ian C. Lucas: He is a liar.
Alexander Nix: It is not true.
Q647 Ian C. Lucas: He not only says that he used Cambridge Analytica; he
said, specifically, that he hired you.
Alexander Nix: That is not true.
Q648 Ian C. Lucas: There are no financial payments from Leave.EU to
Cambridge Analytica or any of associates?
Alexander Nix: Let me be absolutely crystal clear about this. I do not
know how many ways I can say this. We did not work for Leave.EU. We
have not undertaken any paid or unpaid work for them, okay? There is
nothing else I can add to that that is going to clarify that statement in
any more detail.
Q649 Ian C. Lucas: Mr Nix, I am sorry, but I am going to quote back to you
what you said, which is, "We did undertake some work with Leave.EU". It
is in the quotation, and you have just said exactly the opposite. Which is
Alexander Nix: I was using the word "work" to mean that we met with
them to discuss an opportunity. That is working. Unfortunately, having
meetings, even if they do not lead anywhere, is still work but it does not
entail the sort of relationship that you are trying to suggest existed
between their organisation and our company.
Q650 Ian C. Lucas: Would you disclose your bank statements to show that no
payments have been made from Leave.EU to Cambridge Analytica?
Alexander Nix: Yes, I would. I would be pleased to do that.
Ian C. Lucas: I would be very grateful if you would send those to the
Committee so that we can check them.
Q651 Simon Hart: Why are you so desperate to distance yourself from
Alexander Nix: I am not. I am desperate to make sure that the facts of
the matter are crystal clear, because that is the purpose of this inquiry,
although I thought the purpose of this inquiry was that I could help
inform the Committee on how data and targeting are used in
Q652 Chair: Absolutely, and believe me we do want to come on to that. It is
just that because you raised this in your letter to us we feel this is
something we have to bottom out with you.
Simon Hart: Keep going. You were just getting to the end of that.
Alexander Nix: I was simply saying that we were trying to establish the
Q653 Simon Hart: You suggested that the work that was involved was around
preparatory discussions that might or might not have led to some form of
contract. As a way of expanding on the answers you gave to Mr Lucas,
what went wrong? Have you any idea why you did not get the job? Have
you any idea why Arron Banks is apparently so determined to argue that
you did? I do not understand how something so simple could become so
Alexander Nix: Deals fall down or transactions fall down for all manner
of reasons. It could be price, or it could be personalities.
Q654 Simon Hart: What was it in this instance?
Alexander Nix: There simply was not the appetite to move forward.
Q655 Simon Hart: By you or by them?
Alexander Nix: I think by both parties. We did not feel that the
marriage value of Cambridge Analytica working with Leave.EU, and
clearly vice versa, was going to bear a fruitful and successful relationship.
Q656 Simon Hart: Yet it would seem that Leave.EU are, according to you,
making claims now that suggested that that relationship did exist. Why, if
there was not the will go forward and if there was not the will to enter
into any sort of contract, do you think that they are misrepresenting the
truth or, as you put it, commenting inaccurately?
Alexander Nix: I cannot possibly speculate on Arron Banks or Andy
Wigmore or anyone else's motivations. That would be an unfair question.
Q657 Simon Hart: A final point on this, and I think we will come back to the
data element. There is a sense of irony in the way you seem to have
found yourself to be the victims of misinformation being peddled online,
which is arguably one of the accusations that is made about your
company since you assist people in playing to the fears of vulnerable
sections of the electorate in order to alter their voting plans. Do you set a
moral compass anywhere in the manner in which you advise clients on
vulnerable-voter sections in order to try to move them from one position
to another? Do you see that as a positive contribution to society or do
you just say, "They are paying the bills, therefore we will provide
whatever it is they want"? Where does the social responsibility sit in all
Alexander Nix: I think that is another entirely unfair question that
stems from a total misunderstanding about what it is that we are trying
to do and how we help our clients. We are trying to use data and
technology to allow campaigns to engage with voters in a more informed
and relevant way. We are trying to make sure that voters receive
messages on the issues and policies that they care most about, and we
are trying to make sure that they are not bombarded with irrelevant
materials. That can only be good. That can only be good for politics, it
can only be good for democracy and it can be good in the wider realms of
communication and advertising.
Q658 Simon Hart: It is not an unfair question to simply report the fact that
some people consider the manner in which the data is used for electoral
purposes is quite subliminal. It is arguably manipulative. I am simply
asking for comment; I am not expressing a view myself. Quite a lot of
political parties wish they could afford your services, I suspect, but they
do not. I am simply asking whether there is any element of this that
causes you concern. If you are trying to nudge—we were all watching the
presentation you made yesterday, where you are trying to help people
move from one voting position to another. It is not anything that is
particularly drastic; it is just moving a couple of notches on the dial. Do
you have any comments about whether it is unusual when you see a
political party using the advice that you have given perhaps to alarm
certain sections of the voting community into taking a position on the
basis of what, in the old-fashioned term, would be subliminal advertising?
Is that an unfair accusation?
Alexander Nix: Let's start by establishing the fact that the use of big
data and predictive analytics in political campaigns was something that
was really championed by Obama's campaign in 2008. They were the
ones who made the significant advances in what is known now as micro-
targeting—the use of data to start to look at the electorate as very small
groups of people, hopefully, ideally as individuals as opposed to
homogeneous masses, and to start to serve them most relevant
messages. Again in 2012, the Democrats pioneered the use of
addressable advertising technology in order to improve the way that they
use this data to target people as individuals.
As Mr Collins well knows, they have been using these sorts of techniques
in the realm of advertising to personalise advertising for many years—
decades even—as they seek to build relationships between brands and
their consumers such that you do not get blanketed with generic
messaging but everything becomes more relevant to you. That is an
entire industry that is moving in this direction. It is not Cambridge
Analytica. All we have simply done is look at the industry—the advertising
industry—and at what is going on in the political industry, and we have
taken the best practices and in a very short of time we have replicated
them and, I would like to say, improved on some of these techniques and
methodologies and served them up to a different political party in order
to help them have an equal chance of competing in a free and fair
I think part of the issue is that our candidate is somewhat polarising and
so people see the work that we did in a negative light, and they refuse to
accept the fact that Clinton's machine was twice the size or three times of
anything that we were doing for Trump. She had hundreds of data
scientists and digital practitioners working for her. They were using very
similar techniques, and they were targeting the audience in a very similar
way, yet they do not come under the limelight and they do not get the
scrutiny that we get simply because of the candidate involved.
I think if you look at the industry and you say to yourself, "Is it good for
politics that you can make communications more relevant, that you can
start to run a national campaign that involves millions or tens of millions
or even hundreds of millions of voters and you can start to treat that
campaign as you would a small mayoral election or a local election in the
UK and you can start to speak to press releases about very local concerns
that are relevant to them?" Whether it is speeding cameras or regulation
of parking permits or whatever it is—things that matter as opposed to
blanket messaging—that has to be good to make politics more personal,
more individualised and more engaging.
Q659 Simon Hart: Last question, and I should know the answer to this: were
you involved in the 2017 election here?
Alexander Nix: No.
Q660 Simon Hart: 2015?
Alexander Nix: As a rule of thumb, we do not involve ourselves in
politics in the UK.
Q661 Chair: You said as a rule of thumb, but have you?
Alexander Nix: I have been with the company for about 14 years and I
have never worked on a campaign in the UK, simply because, as a
predominantly British campaign, we think it would be complex and
possibly divisive to ask our employees and staff to support a particular
political party in the country that they reside in.
Q662 Chair: Following up, you took umbrage at one of Simon's questions
about playing on people's fears, but you gave a presentation about your
work for the Ted Cruz campaign where you demonstrated that, based on
the psychological profile of the audience, you might use an advertisement
that played on a woman's fear of being attacked in her own home to
support the gun lobby. You might say that techniques like that are used
by other people, but is that not a good example of the sort of campaign
that Mr Hart was referring to?
Alexander Nix: Both sides used fear of spending and fear of economic
exclusion as arguments for staying and remaining in Europe. I think
presenting a fact that is underpinned by an emotion is not
fearmongering. If you believe that yourself, it is very sensible. I think
there is an argument to say that, in the particular instance you are
talking about, there are people who look to the second amendment for
self-protection. In fact, I would say there are quite a lot of people who
fall into that bucket.
Q663 Chair: In that example there that you gave, fear was the emotion that
you were playing on.
Alexander Nix: You are looking at the drivers that are going to influence
the decision making.
Chair: In that case, the driver that was selected in that example for that
decision maker was fear.
Alexander Nix: The fear of being unable to protect yourself.
Chair: The answer to that question is yes?
Alexander Nix: Yes, in that case.
Q664 Christian Matheson: Who is Brittany Kaiser?
Alexander Nix: Brittany Kaiser is an employee of Cambridge Analytica.
Q665 Christian Matheson: Is she still an employee?
Alexander Nix: As of three years, I believe—three or four years.
Christian Matheson: Is she still an employee?
Alexander Nix: She is still an employee.
Q666 Christian Matheson: She spoke, representing Cambridge Analytica, at a
panel on the launch of Leave.EU, did she not?
Alexander Nix: I believe so.
Q667 Christian Matheson: Representing Cambridge Analytica.
Alexander Nix: Representing our proposed involvement as a company
that was going to support Leave.EU.
Q668 Christian Matheson: She said at the time that, "The most important
part of this referendum is appealing to first time and apathetic voters".
Alexander Nix: Yes.
Q669 Christian Matheson: We have had the press release put out by the
junior press officer—that was scotched straight away—but your
involvement with Leave.EU continued up until the very launch and her
speaking at that launch.
Alexander Nix: She was not speaking as a consultant to Leave.EU, she
was speaking as a representative of Cambridge that was seeking to do
some work for Leave.EU.
Q670 Christian Matheson: Did she get paid for being on that panel?
Alexander Nix: No, she did not.
Q671 Christian Matheson: We had the tweet from 29 November, which again
was quickly being scotched by Andy Wigmore, but a couple of months
later, on 10 February 2016, you were quoted in Campaign magazine as
saying, "Recently Cambridge Analytica has teamed up with Leave.EU... to
help them better understand and communicate with UK voters. We have
already helped supercharge Leave.EU's social media campaign". I know
you are unhappy with the line of questioning, but it is yet another piece
of evidence, is it not, Mr Nix, that is contradictory to the statement that
you have given in your letter to the Chairman?
Alexander Nix: I think it is the same piece of evidence that has already
been brought up, so rather than go round the houses and have exactly
the same conversation that we had 20 minutes ago, we have probably
addressed this one.
Q672 Christian Matheson: My fear is that there are several individual pieces
of contradictory evidence that provide a weight to each other.
Alexander Nix: No, there are two pieces of evidence that suggested an
association, and we have addressed them both.
Q673 Paul Farrelly: I want to try to close this opening line of questioning in
my own mind, because I fear that I am hearing the English language
changing in my ears as this session has gone on. You firstly described
that you were not working for someone, but by "work" you meant that
you had meetings about working for someone, which to my mind does
not count as working for someone, so that rather confused me. We have
two sets of characters: you and Mr Banks. I use the book as a coffee mat
in my office, because we were all sent unsolicited copies of it during the
election. Mr Banks is saying, "Hey, we are hiring Cambridge Analytica",
and you are wanting to be Cambridge Analytica working with Leave.EU,
so you are both going around professing love for each other and your
intention to get hitched. Then you say there was no marriage value in
this. What did you mean by that?
Alexander Nix: That we did not get hitched, to use your metaphor. To
use your metaphor, we dated each other, we had a couple of dinners but
we did not get married. Again, how can I spell this out to you? It is pretty
Q674 Paul Farrelly: I am continuing your metaphor. I do not know what a
marriage value is, so perhaps you could help me. There was no marriage
value in it for you. What do you mean?
Alexander Nix: The idea that when two parties come together, the sum
of the relationship is better than the individuals staying on their own.
Q675 Paul Farrelly: I am still confused as to why your relationship broke
Alexander Nix: I am sure, as experienced businesspeople, you
understand that there are often situations where you engage in
conversations about working together with clients and they do not lead to
a relationship being formed. Unfortunately, this is the nature of business.
Q676 Paul Farrelly: Could you spell it out? Did they not think you could
deliver or were they not prepared to pay the rate that you wanted? Could
you be a little bit clearer?
Alexander Nix: I cannot be more clear because I cannot recall. This was
four years ago or three years ago. It was one meeting three years ago
that did not lead to business. We do dozens of meetings every day and
some of them lead to contracts and some of them do not, so I cannot be
more clear. All I know is that we met some representatives from
Leave.EU, we had some discussions, but no business was taken forward.
Q677 Paul Farrelly: It also led to a presence on a launch platform for
something that is pretty seminal in the recent history of this country, but
your memory is not very clear.
Alexander Nix: At the time we were preoccupied with some fairly
important work in the United States and other countries as well.
Q678 Chair: Mr Nix, you are very clear in saying that Cambridge Analytica
received no payment for any work relating to the referendum. Is that also
the case for SCL, your parent company?
Alexander Nix: It is not our parent company, but that is also the case,
Q679 Rebecca Pow: I want to look at the system that you used—I think you
might describe it as a trait-profiling system, the OCEAN system—and at
how you gather data and what you include. Could you very briefly explain
the OCEAN method to us?
Alexander Nix: Obviously, depending on which territory you are
operating in, there are different means to gather data depending on the
legislative environment available. In a country such as the United States,
we are able to commercially acquire large datasets on citizens across the
United States—on adults across the United States—that comprise of
consumer and lifestyle data points. This could include anything from their
hobbies to what cars they drive to what magazines they read, what
media they consume, what transactions they make in shops and so forth.
These data are provided by data aggregators as well as by the big brands
themselves, such as supermarkets and other retailers. We are able to
match these data with first-party research, being large, quantitative
research instruments, not dissimilar to a poll. We can go out and ask
audiences about their preferences, their preference for a particular
purchases—whether they prefer an automobile over another one—or
indeed we can also start to probe questions about personality and other
drivers that might be relevant to understanding their behaviour and
Q680 Rebecca Pow: I think the stated commercial aim of the SCL Group said
that you then collate all this information to micro-target people with all
your analysis in order to influence their long-term behaviour. Can you
give an example or a couple of examples of where this has been very
Alexander Nix: Let me try to route this into something that is a bit more
relatable. If you were an automotive company and you were seeking to
advertise your product to an audience, just knowing whether that
audience was more interested in the engine and performance of the
vehicle, as opposed to the safety features or the boot space or anything
else, is going to be very relevant to how you communicate with them.
That is an example of one or two data points. If you can expand on that
and start to really understand what it is that you, as an individual, care
about in purchasing decisions—purchasing a car for instance—you can
start to tailor the product to the individual and start to tailor the
communication in a similar way. Then you can talk about, in the case of
somebody who cares about the performance of a vehicle, how it handles
and its metrics for speeding up and braking and torque and all those
Q681 Rebecca Pow: I assume you are gathering all this data on the British
population as well.
Alexander Nix: Obviously there is a different set of regulations in the EU
as opposed to the US. The EU is an opt-in data culture as opposed to an
opt-out data culture, as is the case in the United States, so the datasets
that we have in the UK, for instance, are not the same as those that we
have in the US.
Q682 Rebecca Pow: Does any of the data come from Facebook? I have read
that you have said that within so many "likes" you can almost predict
what somebody is going to think about something, or indeed possibly
how somebody might vote, and that you might know more about them
than, say, their partner or spouse or work colleague does within a few
simple steps. Is that right?
Alexander Nix: I have read a similar article. It was not published by us
or written by us, I should say. It was written by an academic active in the
space, so I cannot comment on whether that is true or not. We do not
work with Facebook data, and we do not have Facebook data. We do use
Facebook as a platform to advertise, as do all brands and most agencies,
or all agencies, I should say. We use Facebook as a means to gather
data. We roll out surveys on Facebook that the public can engage with if
they elect to.
Q683 Rebecca Pow: But you can put your micro-targeted messages, as you
were saying, on Facebook as advertisements to try to persuade people or
nudge them in one direction or another.
Alexander Nix: We are platform-agnostic. We will match our offline data
segments with any platform out there. Facebook obviously is an
extremely prevalent platform and has an incredible global reach so it is a
go-to platform of choice for many or most agencies, but if there are other
more-targeted platforms, we would use those.
Q684 Rebecca Pow: We had a gentleman before our panel called David
Carroll, who was an associate professor of media design in the States. He
said that there is no indication of where Cambridge Analytica obtained its
data for any of your rankings. Do you not feel people ought to know
where you are getting your data from and then what you are doing with
it, how you are sharing it, whether you are processing it or even whether
people ought to have a right to be able to delete it?
Alexander Nix: In the United Kingdom, individuals, as governed by EU
law and data protection regulation, are entitled to make a subject access
requests and, as they will be able to under GDPR with all companies, they
will be able to ask for their data and have that data removed from those
companies' databases. We are fully complicit with the law and the
legislation that is currently in place.
Q685 Rebecca Pow: Do you see yourselves as being an all-powerful presence
with all the knowledge and data that you have and that it is not
surprising people are trying to find out whether you are doing anything
perhaps you should not do in the way of influencing elections? You do
seem to be in a position where, with all your knowledge and your
powerful data, you could do that.
Alexander Nix: It is very flattering that you suggest that people might
see us as having these incredible powers. What we are doing is no
different from what the advertising industry at large is doing across the
commercial space. We are a small technology company that is trying to
develop best-in-practice technologies. We are not a political agency, and
we do not have a political ideology. We work on as many elections each
year that are left of centre as are right of centre. We only work for
mainstream political parties; we do not work for fringe actors. We only
work in free and fair democracies. The science of political campaigning
goes back hundreds of years and what we are doing is a very natural
evolution to what has been done before, and what is being done by many
other people as well.
Q686 Rebecca Pow: Doesn't the very fact that you are working on political
campaigns mean that you must be influencing them, given that your
remit is to influence people's long-term behaviour?
Alexander Nix: All campaign management consultancies or agencies are
there to help their customers or their clients, as a good advertising
agency is there to help the brands that it represents. We are there to
make sure that our candidates are able to communicate with the
electorate in the most relevant and effective way. That is what campaign
consultancies do, as most of the people in this room should well know
because of their involvement in politics.
Q687 Ian C. Lucas: Do you share data between, for example, SCL and
Alexander Nix: SCL is a very different company to Cambridge Analytica.
It is a different company that has different employees who sit in a
different office. It has a different board and a different board of advisers.
It has different datasets, and it has different clients. The short answer is
no. The only relationship between Cambridge Analytica and SCL is some
shareholders. Apart from that, they are completely separate entities.
Q688 Ian C. Lucas: There would never be circumstances when you would
transfer data from SCL to Cambridge Analytica?
Alexander Nix: We could transfer data from Cambridge Analytica to
SCL, but because SCL is a company that operates in the government and
defence space, it acts as company that has secret clearance—X-list
accreditation in the UK—so we could not transfer data the other way.
Q689 Ian C. Lucas: You could transfer data from Cambridge Analytica to SCL,
Alexander Nix: Certain data, yes.
Q690 Ian C. Lucas: Are there any individuals who work for both organisations?
Alexander Nix: There are individuals like myself who, at a high level, sit
on a board of both organisations, but there are no employees who work
for both organisations.
Q691 Ian C. Lucas: Is it your understanding that if I lawfully give one of those
businesses information about me, another one of those businesses can
use that information?
Alexander Nix: I said certain data. There are certain data that we can
go out and commercially—I am talking about the United States, by the
Ian C. Lucas: I am talking about the UK.
Alexander Nix: In the UK that is not our practice.
Q692 Ian C. Lucas: It is not your practice. Is that because it is unlawful?
Alexander Nix: We do not share data. No, it is simply because there is a
different legislative environment here.
Ian C. Lucas: So it is because it is unlawful.
Alexander Nix: In America, we can go out and acquire data. In the UK,
we can still work with data, but as a data processor not as data
controller. We can work with client data. I am sure you are familiar with
the distinction between the two. We never own these data; we are simply
processing these data on behalf of our clients.
Q693 Christian Matheson: Very briefly on that—going back to the answer you
gave to Mr Lucas about not having any common employees, just common
shareholders—the registered representative with the Information
Commissioner for Cambridge Analytica is Jordanna Zetter. Does that
Alexander Nix: Sorry, the registered representative for—
Christian Matheson: With the Information Commissioner.
Alexander Nix: Right.
Q694 Christian Matheson: Is it not the case that she is also publicly named
as the Operations Executive for SCL Elections Ltd and Cambridge
Alexander Nix: Jordanna has been acting as the liaison in an
administrative capacity for helping the ICO with some of their inquiries
into data and data protection. We have a data compliance team who are
undertaking the work. Her role is more about co-ordination and
Q695 Christian Matheson: You would stand by the position that there are no
common employees or employees who spend time working for both?
Alexander Nix: Jordanna's employment is with Cambridge Analytica.
Q696 Ian C. Lucas: Can I come in on that? That is a very important role, full
stop, but particularly in an analytics company—the person in charge who
is lawfully responsible to the Information Commissioner.
Alexander Nix: There is a misunderstanding. She is not the person in
charge. Ultimately, the CEO is in charge and our data compliance team is
in charge. She is simply the liaison who passes messages between the
Q697 Christian Matheson: What is the difference between owning the data
and processing it?
Alexander Nix: In the UK, if we were to undertake work for a big
corporate, their data would be owned by them and they would always be
the data controller. They would have control and responsibility for their
data. They could bring us in to work on their data, but we would never
take receipt of that data and we would never own that data. We would
simply come in and perform analytic function on that data. Their data
would remain their data.
Q698 Rebecca Pow: This is just a small point related to data. I believe I asked
you whether you gathered the data from Facebook and whether you were
using all that information. I think you said you did some surveys. Could
you expand a bit more on what those surveys are, what you are asking
people and how you are gathering the data? Do you keep that data on
surveys carried out on Facebook or does Facebook keep it?
Alexander Nix: I cannot speak to Facebook, but as far as I am aware
the process works a bit like an opinion survey. If I want to find out how
many people prefer red cars or yellow cars, I can post that question on
Facebook and people can agree. They can opt in to answer a survey and
they give their consent and they say, "I prefer a yellow car" and then we
can collect that data. That is no different to running a telephone poll or a
digital poll or a mail poll or any other form of poll. It is just a platform
that allows you to engage with communities.
Q699 Rebecca Pow: Are they a big part of your data-gathering service?
Alexander Nix: When we work for brands, whether it is in the UK or in
the US or elsewhere, we often feel the need to probe their customers and
find out what they think about particular products or services. We might
use Facebook as a means to engage with the general public to gather this
Q700 Simon Hart: Let me ask a very quick question on the Facebook survey
opt-in option that you were describing. If you are asking somebody what
kind of car they prefer and they opt in, does that facilitate access to other
data that may be held by Facebook, which is irrelevant to car colour, or is
it only the data you collect on car colour that is relevant? Nothing else
that is part of the data held by Facebook would be available to you.
Alexander Nix: You are absolutely right—no other data. As far as I am
aware, Facebook does not share any of its data. It is what is known as a
walled garden, which keep its data—
Q701 Simon Hart: People are not in any way accidently giving you consent to
access data other than that that you specifically asked for.
Alexander Nix: That is correct. People are not giving us consent and
Facebook does not have a mechanism that allows third parties such as us
to access its data on its customers.
Q702 Simon Hart: Even with its customers' consent.
Alexander Nix: Even with its customers' consent.
Q703 Chair: I have taken one of your surveys. It was found through your
website. I think it was a profiling survey that is linked to the OCEAN
model. The incentive to take the survey is to understand more about your
psychological profile. When you complete the survey—this is where I
bailed out of the process—it invites you to complete the survey and get
the information back you want by logging in with your Facebook log-in at
the end of the process. If someone does that, what data are they
allowing you to share from their Facebook process? What is the purpose
of the Facebook log-in at the end of the survey?
Alexander Nix: Logging in with Facebook is a fairly common practice in
the digital realm. It simply saves you the time of putting in your name
and e-mail address and so forth, such that we can then send you that
report on the survey that you have just taken.
Q704 Chair: Does that give you the right to access any other data points from
my Facebook profile?
Alexander Nix: No, absolutely not. Absolutely not.
Q705 Paul Farrelly: I just want to clear up two things—I am sure we want to
pursue the use of Facebook. You mentioned that SCL Group, or whichever
of the companies it is—you can perhaps be more precise—has X-list
accreditation for work with Government and defence. Can you explain
that? I have not come across an X list before. Forgive my ignorance.
Alexander Nix: Not at all. I am sure you are not ignorant at all. SCL is a
behavioural communications agency that was set up specifically to
service the government space. We do an awful lot of work in the UK, but
it is principally in the US, working with Government Departments such as
the DOD, State, Pentagon and so forth. We are trying to use an
understanding of group audience behaviour to address often hostile
actions. Specifically, that includes programmes of counter-terrorism and
counter-radicalisation. We are looking at how we can use an
understanding of audiences to address problems in the drug trade, in
children and women trafficking, programmes of social change,
Government information programmes, a huge number of health
programmes—trying to understand how we can encourage people to live
more healthy lives.
Q706 Paul Farrelly: Is this all in the US, not in the UK?
Alexander Nix: It is in the US, the UK and globally. Obviously, the US is
a larger market for this kind of work.
Q707 Paul Farrelly: Are all the various SCL entities involved in this?
Alexander Nix: That is correct.
Q708 Paul Farrelly: What does SCL stand for?
Alexander Nix: Historically, going back some 14 or 15 years, it stood for
Strategic Communication Laboratories. It has been abbreviated now.
Q709 Paul Farrelly: One more loose end. To Rebecca Pow's question about
Facebook, where she said that with a certain number of likes you could
know someone better than this, this or this, you attributed that to some
Alexander Nix: I do not think I attributed to anyone by name.
Paul Farrelly: No, you said it was something that had been written by a
Alexander Nix: I think I said by an academic, but that is my
Q710 Paul Farrelly: You said that yourself in a speech to the fabulously titled
Online Marketing Rockstars conference in 2017. You made those claims
about Cambridge Analytica's capability.
Alexander Nix: Could you read that out, please?
Paul Farrelly: You claimed, "With 10 Facebook likes, Cambridge
Analytica can predict an individual's behaviour better than their work
colleague might. They only need 70 to make"—that is you—"behavioural
predictions better than a friend, 150 to understand a voter better than
their parents". With 300 likes you claimed your organisation can, "predict
a person's actions, thoughts and feelings better than their spouse".
Alexander Nix: Those are not my words. I am familiar with that text.
That was the text that your colleague Rebecca was quoting from but
those were not Cambridge's words. That was a statement that was made
by an academic who spent a number of years, I believe at Stanford
University, looking into this area. That was his work, and that was his
statement. I do not know why that has been attributed to Cambridge.
Q711 Paul Farrelly: You have not made those claims on Cambridge
Alexander Nix: That is not my statement, period.
Q712 Paul Farrelly: Did you quote it?
Alexander Nix: No, I did not quote it. I have never memorised those
statistics in order to include them in a quotation.
Q713 Ian C. Lucas: Is Cambridge Analytica a Facebook developer?
Alexander Nix: No.
Q714 Ian C. Lucas: What is your relationship with Facebook?
Alexander Nix: We are a client of Facebook. We purchase advertising
through Facebook, as every other digital agency does.
Q715 Chair: Mr Nix, I would like to clarify one or two things that you have
said, and I have one or two things about Facebook before we move on to
other topics. You were asked earlier about Jordanna Zetter's dual role
working with SCL and with Cambridge Analytica. Was that dual role
created in response to the inquiries from the Information Commissioner
or was she working on that dual role before those inquiries were made to
Alexander Nix: As far as I am aware—and I would be pleased to circle
back to the Committee to confirm this—Jordanna Zetter is employed by
our office in London to work for Cambridge Analytica. She has no formal
role with SCL Group, which is based out of Arlington, Virginia, in the
United States. I think, as far as I am aware, she has never even visited
those offices and has no relationship with them. She was simply asked to
help with the ICO's inquiries into data and data protection. She kindly
agreed to act as a liaison in that respect. Should they have any
questions—I am not sure that they do or did, but again I would need to
confirm that for you—for SCL in the United States, she would pass them
on and be the conduit of that, simply because they do not have a
relationship with those people. It is purely an administrative role. She is
not a senior member of staff.
Q716 Chair: If you were able to confirm that in writing we would be grateful
for that, and also whether that liaison role was in place before the
Information Commissioner's investigation commenced.
Moving on to Facebook, in response to other questions you drew a
distinction between being a data controller and a data processor. Could
you explain a little bit more to us about that distinction between those
Alexander Nix: Again, this is certainly not my area of expertise. I run
the leadership team, not the data compliance team. My understanding of
how this works in the United Kingdom is that the brands that we work
with remain the controllers of their own data. That is, if you are a large
retail brand and you have collected a lot of data on consumers, you own
and are responsible for that data and for looking after it and securitising
it and protecting it and all the regulation that governs it. These brands
are allowed to engage with companies and agencies such as Cambridge
Analytica in order to help them to process or model this data. We perform
an analytic function as a subcontractor or as a contractor on this data.
We do that work for them, often on their own servers within their own
data ecosystem and then we leave. We do not control that data, and we
do not have a copy of that data.
Q717 Chair: You said in your letter to me that, "Cambridge Analytica does not
gather" data from Facebook.
Alexander Nix: From Facebook?
Alexander Nix: That is correct.
Q718 Chair: The actual quote from the letter is: "On 8 February 2018 Mr
Matheson implied that Cambridge Analytica 'gathers data from users on
Facebook.' Cambridge Analytica does not gather such data." But from
what you said you do, do you not, through the surveys?
Alexander Nix: Yes, I think I can see what has happened here. What we
were trying to say in our letter is that we do not gather Facebook data
from Facebook users. We can use Facebook as an instrument to go out
and run large-scale surveys of the users, but we do not gather Facebook
Q719 Chair: By that do you mean that you do not have access to data that is
owned by Facebook?
Alexander Nix: Exactly.
Q720 Chair: You acquire data from Facebook users through them engaging
with surveys and other things.
Alexander Nix: Exactly right.
Q721 Chair: Is your engagement, either directly or through any associate
companies you may have, just through the placing of surveys or are
there other tools or games or things that are on Facebook that you use to
gather data from Facebook users?
Alexander Nix: No, simply through surveys.
Q722 Chair: We referenced earlier the presentation you gave at the Concordia
Conference in 2016 about your work for the Ted Cruz campaign. Were
you hired to work for Donald Trump on the basis of your work with Ted
Cruz? I imagine you could not have been working for two candidates in
the same race, so did you start—
Alexander Nix: Actually we were working for two candidates in the
primaries. We were working for Dr Ben Carson's campaign and for Ted
Cruz's campaign, and we were seeking to work for Donald Trump's
campaign throughout the primaries as well but the Trump campaign did
not wish to engage with us other than on an exclusive basis. We had two
clients who we were well established with and we were not willing to give
those up, so we said to the Trump campaign that, "In the event that you
win the primary let's reopen the discussions that we started almost a
Q723 Chair: The work that you did for the Trump campaign was obviously after
Ted Cruz's campaign had ended. Is that right?
Alexander Nix: Immediately after Trump won the primary—Cruz coming
to second to him—we reopened a dialogue with the Trump campaign
about how we could take all the technology that we developed, largely for
the Cruz campaign, and pivot it across to Trump to give him the same or
Q724 Chair: Is that what you did?
Alexander Nix: That is exactly what we did.
Q725 Chair: In your presentation you said that the Cruz campaign relied on a
sort of tripartite strategy where the campaigns were based on
understanding behavioural communications, which is where the OCEAN
survey and psychological profiling comes in, data analytics-data from
multiple sources where it is available—and then ad data to place the
messaging. Is that the same approach you used for Donald Trump's
Alexander Nix: We would have liked it to have been, but it was not the
same approach, simply because when we joined the Trump campaign we
had about five and a half months before polling to build or re-engender
the entire analytics capability that Clinton's team had and that we had
been giving to Cruz. We simply did not have the time and resources to be
able to go into the same depth of services that we provided to the Cruz
campaign, which we had been working on for very many more months. It
was a very extended and protracted programme. We made the decision
to focus on the data and analytics elements of the campaign and the
tech, digital and data-driven TV elements of the campaign. We did not
have time to bake in or to incorporate the behavioural approach, the
psychographics that we had used on the Cruz campaign.
Q726 Chair: Why not? You said that the psychographic information is based on
surveys you have done, which means you believe you have an accurate
model for understanding every voter in America. Presumably that
database could be migrated to support other campaigns as well. It is not
data specific to a particular campaign, is it? It is generic profiling of
Alexander Nix: Yes, but then you have to take these data and
contextualise them into the campaign that you are working on.
Everything that we did for Cruz in that regard was centric around Cruz.
We would have to then replicate that for Trump and that would have just
been the most enormous piece of work.
Q727 Chair: In that case, I don't understand why you were hired. You have
made it quite clear in your presentation on your work for Ted Cruz that
your model is based on the combination of these different elements. The
ad data is about message targeting—the media placement bit of it—but
the smart bit is the merging of psychological profiling and data analytics.
In fact, what you do here in this country is based on that too. That seems
to be your USP. Why would none of that psychological profiling have been
used to augment the data in the Trump campaign when that is the way
Alexander Nix: As I said before, the reason it was not used was because
we simply did not have the time and resources to include it. If we had
had that opportunity, we would have. We did not and we are just trying
to be transparent about that fact. To say that the smart bit is simply that
would be doing an injustice to the 40 or so PhD data scientists who spent
100 hours a week for five months crunching data and numbers in order
to develop the very accurate and insightful models that they did for the
Q728 Chair: The way you sold yourself and what your company did was based
on the combination of these elements together. If I was someone in the
room that day—the way these conferences work is that you are
effectively there pitching yourself and your company in the hope of
winning new business and new clients on the back of it. You have a very
clear model and it is quite interesting that one of the three supports of
that stall has been taken away to go and work on another high profile
campaign. It seems very strange.
Alexander Nix: You can only provide a client, whether it is political or
brand, with the services that you are able to deliver within the constraints
of the project timelines that you are presented with. You are absolutely
correct that, in an ideal world, not only would we have liked to have
delivered the services that I spoke about in that presentation but many
other services that we have developed across the engagement space, as
digital and television in particular, but we simply did not have time.
Unfortunately, unlike running a brand campaign where you are selling
automotives or toothpaste or something where you don't have the same
time pressures, we have a finite amount of time. You have to choose
which technologies in your arsenal are going to be the most important
and that can be deployed most effectively, and we made a decision.
Q729 Chair: In that presentation I think there is a slide on data analytics
where you describe that data is sourced from multiple sources and any
marketing company will know that there are companies that specialise in
data analytics to analyse consumer behaviour. I think on your chart you
had logos of different companies. I think Experian was one and Nielsen
was one. You had Facebook on there as well. Again, just to confirm on
this, is that because you are highlighting the fact that you can gather
data from Facebook?
Alexander Nix: Collect data through Facebook—that is exactly right,
Q730 Chair: Does any of your data comes from Global Science Research
Alexander Nix: GSR?
Alexander Nix: We had a relationship with GSR. They did some research
for us back in 2014. That research proved to be fruitless and so the
answer is no.
Q731 Chair: They have not supplied you with data or information?
Alexander Nix: No.
Q732 Chair: Your datasets are not based with information you have received
Alexander Nix: No.
Chair: At all?
Alexander Nix: At all.
Q733 Ian C. Lucas: Can I go back to an answer I think you gave earlier? I just
want you to confirm something. Did you say that you had never worked
on any political campaigns in the UK?
Alexander Nix: I said that I personally had not been involved in any
political campaigns in the UK. I have been with the company for about 14
years and, as far as I am aware, in the last 14 years we have never
worked on any campaigns in the UK.
Ian C. Lucas: That is Cambridge Analytica and—
Alexander Nix: Cambridge Analytica was only formed in 2012, so this
would have been a company that I worked for prior to forming Cambridge
Analytica, which was called SCL.
Q734 Ian C. Lucas: By political campaigns, that means general elections.
What about for candidate elections for political parties within the UK?
Alexander Nix: Again, as far as I am aware, since I have been in the
company we have never worked—we don't seek to work in the UK, for
the reasons I discussed earlier. We don't see the UK as a commercial
market of interest for SCL or—
Q735 Ian C. Lucas: That really puzzles me because there are lots of other
businesses that might do work for different political organisations within
the advertising sector or within the information sector. Why is it that you
are so involved in politics in the US but you are not involved in the UK?
Alexander Nix: I think I have addressed this already but let me explain
it again for you. First, we only entered the US market in 2012. We have
been running election campaigns since 1994. We take on a number of
national elections every year. That could be three, four, five, six, seven
elections across the world in every single year for prime ministers and
presidents. That could be in Asia, Latin America, Europe, Africa or
beyond. We entered the US market relatively recently. We saw a
commercial opportunity to bring some of the technologies that we had
developed to that market, particularly to help with Republican politics
because they were losing the tech arms race, if you like, to the
Democrats and that is where the opportunity existed. We have no more
interest in servicing the US political market than we do helping political
parties in Africa or Asia. Specifically in regards to the UK, as I said
before, originally we were a British company with most of our staff based
in the UK, and we felt that it would be unfair on our staff for senior
management to make a decision about which political party the company
supported, especially if that was at odds with the political views of our
employees, so we did not want to put them in that awkward position.
Q736 Ian C. Lucas: That is very thoughtful, because lots of other businesses
don't do that in the UK. Could you tell me who your clients are in the UK?
Alexander Nix: We work for brands in the UK. Although the current
impression is that we are a political consultancy, I have tried to explain to
this Committee that we are not a political consultancy. We are a
technology-driven marketing firm, and the majority of our business is in
the brand space. We work for big and small brands, trying to help them
market and sell their products and services to consumers around the
world. We also have a government and defence division, which I am very
proud of and which does enormously important work in saving lives all
over the world in campaigns about issues that really matter. We have a
political division, but our political division is only, say, 20% or 25% of our
Q737 Ian C. Lucas: Given your thoughtful approach with your employees
about not getting involved in political campaigns, what led you to start to
discuss with Leave.EU about getting involved in probably the most
contentious political campaign in British history?
Alexander Nix: It was only an exploratory discussion and, as I have said
to the Committee already, we have these sorts of discussions with all
parties in the UK. Without exception, every single party has approached
us, and every single party has asked about our services. We have had
these discussions. Some of them have been more protracted than others
and we have never engaged with anyone—not on my watch—but it does
not mean we are not interested to understand more about their needs
and concerns, to understand more about the technology they are
embracing and to see what the market is like. This was an example of
that. In the end the decision was made not to move forward.
Q738 Ian C. Lucas: You are not a partisan political outfit.
Alexander Nix: I don't think we would be particularly good at our job if
we were partisan. We try to be objective. We try to walk into a country
and service our clients with the best, most cutting edge technology and
methodologies available for communications and campaigning.
Q739 Ian C. Lucas: Why was Steve Bannon on your board?
Alexander Nix: Steve Bannon was on our board to help a British
company to understand a new market that it was trying to penetrate. I
can't think of many people who would be better to help a company enter
a market, particularly into the Republican space, than somebody who had
extensive experience of the commercial and business landscape through
his time at Goldman Sachs and the like, who understood the media
landscape through other experiences, and who also had a very attuned
Q740 Ian C. Lucas: I can't think of anyone who is a more partisan political
Alexander Nix: I can't speak to your—
Ian C. Lucas: Okay, but it does not sit easily with the fact that you are
non-partisan in one country but you are massively partisan in another.
Alexander Nix: In the United States, the practice is not to switch sides.
You either work for the Republicans or you work for the Democrats. You
don't do one election for one and then change sides. That is how the
convention is. In other countries, we always give our previous clients first
bite and first sight of our services, but if for any reason they do not wish
to engage with us, we are at liberty to go and work for opposition parties
and sell our services to them.
Q741 Chair: You might be interested to know, Mr Nix, that various people are
watching the evidence session. It is being broadcast and people are
tweeting about it. Julian Assange has tweeted about it, sharing a link to
the session where people can watch it. Has Cambridge Analytica or any of
its associate companies ever worked on campaigns to distribute
information that has been sourced from WikiLeaks?
Alexander Nix: We have no relationship with WikiLeaks. We have never
spoken to anyone at WikiLeaks. We have never done any business with
WikiLeaks. We have no relationship with them, period.
Q742 Chair: That was not quite the question I asked, which was whether you
had ever been involved in advising on or organising campaigns to
distribute information that has been sourced from WikiLeaks.
Alexander Nix: We have never been involved in organising or advising
on campaigns that distribute data or information from WikiLeaks.
Q743 Chair: Arron Banks has also been following the session and tweeting
about it. He has invited himself to come and give evidence to the
Committee, which we might well take him up on—Mr Lucas suggested
that earlier on. If Mr Banks is still watching, if he wants to keep his diary
free over the next couple of weeks, we may well be in touch. What he
says in his tweet about your negotiations—this potential marriage or
courtship that failed—is that, "CA wanted a fee of £1m to start work &
then said they would raise £6m in the states. We declined the offer
because it was illegal." Is what he is saying correct?
Alexander Nix: Absolutely incorrect.
Q744 Chair: That is the second time he has lied, according to you.
Alexander Nix: Mr Banks is at liberty to say whatever he likes, but I
don't have to agree with it.
Q745 Chair: What he has said in that message is totally untrue?
Alexander Nix: That is totally untrue.
Q746 Simon Hart: I have a quick question on the attitude to your staff and not
wishing to put them in a difficult position when it comes to choosing
campaigns. Do you apply the same principle to which brands you choose
to represent? Do your staff have a say? They may feel more comfortable
with some brands than others. Do you consult them over that?
Alexander Nix: We always give our staff a choice about which projects
they would like to work on. If anyone for any reason does not feel
comfortable working on a particular brand, we are happy to offer them
the opportunity to work on a different brand. If, for instance—and I keep
using the example—you don't believe that the automotive industry is
necessarily good for the environment, you do not have to work on it. You
can go and work on selling bicycles.
Q747 Simon Hart: Likewise, UK employees working on the Trump campaign
could also choose—
Alexander Nix: All our employees have that opportunity.
Q748 Chair: I wanted to ask you something about data that I meant to ask you
about earlier. You gather data from various sources, including from
Facebook through the survey tools you have on the platform. Do you
retain that data and information unless you receive a request to hand it
back or to destroy it?
Alexander Nix: In a country like the United States, we retain that data.
Some of that data is purchased and some of that data is licensed, so then
you need to either return it or delete it or refresh it.
Q749 Chair: When you talk about data being purchased, who are the people
you have purchased it from in that case?
Alexander Nix: It is possible to go out and purchase commercially
available datasets. There are people who make a living from selling data
in the United States.
Q750 Chair: Could you give us an example of a company?
Alexander Nix: This would be people who collect data off their
customers, for instance.
Q751 Chair: Given that Facebook is such a major platform, would that include
customer Facebook data that has been gathered by other people that you
Alexander Nix: I don't know. My understanding—you would have to
speak to Facebook and I know you have spoken to Facebook—is that they
do not share any of their data, and it would be bad for their business
model, I assume, if they did.
Q752 Chair: Yes. That is certainly what they said to us in their evidence
session. You would hold data that you have acquired on people in
America in particular. That data would be used in different campaigns.
You might gather that data as part of one piece of work you are doing for
one client, and you might even use that data in another campaign if it
was relevant to that campaign.
Alexander Nix: Yes. At a high level, yes, but it depends on the data
agreement that is signed with the person that you acquire or license that
data from. Different datasets might have different licensing regulations.
Q753 Chair: In that presentation, you cited the Iowa caucus as being a
campaign you have worked on. Could data that you used working for Ted
Cruz in Iowa have been used in the Trump campaign as well?
Alexander Nix: Hypothetically, it could have been, yes.
Q754 Chair: We have talked a lot about profiling, both psychologically and
through data analysis. The third pillar of your presentation was about ad
data and ad placement. Do you advise clients not just on how to reach
people and what message they should see, but also how frequently they
need to be contacted about a message in order for it to be persuasive?
Alexander Nix: Yes, we do.
Q755 Chair: Do you advise on the architecture for distributing that information
as well? We have taken a lot of evidence about networks of accounts on
Facebook and Twitter in particular that are used and set up to reach
audiences and target audiences with information. Would you advise
people on how those should be set up?
Alexander Nix: Do you mean physical architecture or are you talking
about the balance between different channels and how they are
Chair: I suppose it is a combination of those things. If you are saying
that you give advice on how frequently someone needs to see a message
and, let's say, that Facebook is going to be one of the chosen media in
order for that message to be seen, do you advise on how that message
should be delivered by Facebook? You could just set up a Facebook page
with some information on, you could use advertising to supplement the
reach of that, or you could use the interaction of other accounts and
other pages to bring audiences to that page. Is that the sort of advice
you give clients as part of advising them on not just what they should
say, but how frequently they need to say it?
Alexander Nix: It is either the advice that we would give them or the
work that we would undertake on their behalf.
Q756 Chair: We have looked at the role of bot accounts, in particular on
Twitter. Is that something that you would advise clients on, how to set up
networks of bot accounts to augment messages and make sure that
people see them frequently?
Alexander Nix: Absolutely not. That is not something we engage with. It
is not something that we would engage with. It goes against everything
that we are trying to achieve. What we are trying to do is make sure that
the most relevant messages hit the right audiences. The idea that you
could use bot accounts to spread messages is contrary to everything we
are set up to do. That goes back to the area of blanket advertising and
spamming people with irrelevant information. That is not what we do.
Q757 Chair: Twitter were very clear when they gave evidence to us that bot
accounts can be used for good and bad. There is a bot account in my
constituency that tweets the weather every day, which is a perfectly
harmless service that someone is providing. The question is what they
are for, but clearly any one of these accounts is used to target
information at people you want to receive it. It is a targeting tool, not
just a broadcasting tool. Is that the sort of advice you give people on how
to use these networks of accounts in order to reach the right people?
Alexander Nix: No, it is not. It really isn't, because there would be no
need to invest in our services if that was the implementation or the
engagement process that you were seeking to use. What we are trying to
do is get away from everything that could be construed as mass
communication, spamming or large-scale media engagement. We are
trying to make our communications more personal—really personal. We
are trying to build the individual relationship between the brands and
their customers. I agree with you that bots can be used for good, but it is
generally not a technology that aligns with what we have been set up to
do and what we have been doing for the last 10 years.
Q758 Chair: If you had a client and you were advising them on how to reach
their audience and the frequency with which they need to reach their
audience with a message, and Twitter and Facebook were the chosen
platforms through which that communication would take place, how
would you go about planning and developing that campaign and
Alexander Nix: What we are trying to do is look at an audience and
segment that audience into as many different groups as possible such
that we can begin to identify what each of those subgroups in the target
audience care about most in relation to a specific product or service or,
indeed, political candidate or campaign. Then we start to tailor messaging
that can be made most relevant to the concerns, issues, hopes and fears
of those particular people, so that we can give them the facts of the
matter in the most relevant and personal way. It is about breaking
messages down into multiple messages and then nuancing them to make
them more relevant.
Q759 Chair: I understand the message creation; I am asking about the
message delivery. How do you deliver the message to somebody with the
media that you use?
Alexander Nix: In the instance of Facebook, we are able to take these
offline segments, and we can match them to cookies and target the
specific adverts that we have made for that group of people, whether it is
100 people or 100,000 people, and serve that message to them through
Q760 Chair: Through cookies on any site?
Alexander Nix: On Facebook or any other platform.
Q761 Chair: Given that Facebook is a closed platform, how do you do that? Do
you go to Facebook and say, as other advertisers would do, "We have got
this campaign. We want to target Facebook users. This is the profile of
the people we want to target"?
Alexander Nix: Exactly that. It is an anonymised match based on a
target profile that we have developed through our data analytics.
Q762 Chair: Then you just pay Facebook in the ordinary way for targeting
Alexander Nix: Exactly that.
Q763 Chair: Do you gather data that enables you to plan campaigns that might
be targeted at people who have certain religious beliefs or come from a
certain ethnic background or hold certain political opinions?
Alexander Nix: In our campaigns we are trying to get away from
demographics. We are trying to look at people based on their
fundamental drivers and, in the case of politics, this is about who is likely
to be supporting a Republican versus a Democrat and what issues are
most relevant to those audiences. I don't think that targeting in the way
that you have implied is going to help us.
Q764 Chair: If someone came along and said, "I want to run a campaign that
is going to be particularly relevant to people that I think are voting
Republican but have very strong religious beliefs", would you say, "We
can create that campaign for you because we not only understand
people's political motivations but we also can identify people with strong
religious beliefs because of the profiling we have done on them"?
Alexander Nix: I think that would depend very much on the data that
we have access to or have already gathered. I would not be able to
answer that question specifically, but hypothetically it would be possible if
you had enough data, say on evangelical Christians in America, to have a
look at that audience and see if there is a correlation between that and
some political agenda.
Q765 Chair: These sorts of data are gathered that theoretically makes it
Alexander Nix: It certainly is not gathered by us, but obviously there
are very large church organisations and religious organisations that might
have access to these types of data.
Q766 Chair: We have talked a lot about America and I appreciate the data laws
in America are different and, in some ways, I guess that makes what you
do easier in America than it does in the UK. What steps do you take as a
company to make sure that you are always fully compliant with UK data
Alexander Nix: Data is the core of our business and so we take data
incredibly seriously. We have an in-house data compliance team who are
working continually with the legislators not only to help understand the
laws, but to help inform on data legislation and how it could be updated
and kept forward. We anonymise and encrypt all data that we receive
from clients, and we would like to believe that we are very much at the
cutting edge of the technologies behind both of those services. We do not
store data locally on devices in order to mitigate the possibility of data
breaches and so forth. Internally, we have our policies on how we treat
data and so forth. It is something that we have given a great deal of
consideration to and something we take incredibly seriously.
Q767 Chair: When you say you are at the cutting edge of this industry and the
way you use that data is part of being at the cutting edge of that, what
sort of data processing do you do? You say you are receiving raw data
but then you are organising it in some way to make it more relevant. Can
you explain a bit more about how that process works? What would you
consider to be an ethical use of that data being processed in such a way?
Alexander Nix: Certainly. We are not a data miner or a company like
that. We are a data analytics company, so our job is to turn data into
insight—to take very large datasets and try to identify patterns in that
data and to use the data to make predictions about audiences. We are
just trying to run algorithms on the data to try to find meaning in it.
Q768 Chair: Is it fair to say that there would be many people who have given
their data to you who are not aware that their data is being used in this
way and that they could be targeted in a way that they would never have
expected, because they do not understand that that is how their data can
be not only gathered but processed and then used to support other
Alexander Nix: There are several answers to that. The first answer is
that these are not particularly intrusive data. This is not like someone has
given up their health data or their financial data or their private data.
These data are commercially available, as I have said. These are data on
your consumer and lifestyle habits: what car you drive, what magazines
you read, whether you have Weetabix for breakfast and the like. I think
most people understand that there is a reciprocity with large brands
whereby they agree, for instance, to receive a loyalty card and get
discounts and offers on products and services from that company. Most
people understand that their data is being taken in return to help that
brand to drive its marketing. Let's say that large UK supermarkets—Tesco
or Sainsbury's—all have these types of things. People are not naive. They
understand that reciprocity and they say, "If people find out whether I
buy a loaf of bread and a pint of milk and I get 10% discount off, that is
a fair trade-off".
What is this going to look like in the future? I think that the landscape is
changing, clearly with GDPR coming in. What we are seeing is that people
are going to want to have more sovereignty over their data and are going
to want to see a greater reciprocity of how their data is used and greater
control. I think that is very healthy and something that we are investing
very heavily in and look forward to. It is going to improve the data
landscape, improve how data can be used where people say, "Okay, I
recognise that my data has a value and why should other companies
simply benefit from that? Why should I not be a participant in receiving
some of that remuneration?"
Q769 Chair: Do you not see there is a big difference here between saying, "I
understand that you like a particular brand of car and, therefore, I am
going to send you information about other brands of car that are similar
and you might be interested in because we know you like a certain type
of car"—which I think people understand as it has been a marketing
technique or a direct marketing technique for very many years—and
saying, "Because of the information I have gathered about you, I know
how to make you frightened"? Isn't that a very different proposition?
Alexander Nix: These are only opinions we are looking at. You can go
and speak to people yourselves and you can form an opinion based on
those conversations about what might be the most relevant information
to them. All we are doing is looking at data and making our own personal
opinion about what we think is going to be important insights from that
Q770 Ian C. Lucas: Can you help me a little with these surveys that elicit
information from the people who fill them in? I have filled in a couple of
these surveys online over the years but to my knowledge I have never
filled in a Cambridge Analytica survey. When you are eliciting
information, what does it look like on a platform? What does it look like
on a Facebook platform?
Alexander Nix: It might start with basic demographic information: your
name, your age, your gender.
Q771 Ian C. Lucas: Does it say who is asking?
Alexander Nix: That will obviously depend—I am sure you have seen an
opinion survey in your life. They are all fairly similar. They follow a fairly
standard structure, which is generally establishing who you are speaking
to and then asks you some questions.
Q772 Ian C. Lucas: Does it say it is from Cambridge Analytica? Would you
ever present a survey on Facebook as saying, "This is a Cambridge
Analytica survey. Please give us this information"?
Alexander Nix: We have done and we do, but it depends what the
purpose of the survey is.
Ian C. Lucas: I have never had one.
Alexander Nix: I think we rolled out 350,000 to 400,000 surveys a
month for the Trump campaign in the United States over a five-month
period. These were being done on and behalf of the Trump campaign and
that was the label of the survey. The fact that we were helping them to
gather these data was less relevant.
Q773 Ian C. Lucas: I think it is important that people know who is asking. You
have just said that was for the Trump campaign, and I think that is
entirely legitimate and fair, but when individuals fill in a survey, do you
think they should be told who is the client, who is the person asking them
and where the data is going?
Alexander Nix: I can see no reason to obfuscate that truth. These are
entirely voluntary surveys. If someone knocks at your door and says,
"Could you fill out a survey?" you don't have to undertake that. They
might say, "Who is this for?" and you say it is for cancer research.
Q774 Ian C. Lucas: I am not sure that people understand what their data is
being used for.
Alexander Nix: I think I would disagree with that. Most people
understand that data is being gathered. You are not filling in a survey
simply for your own entertainment. I think they do understand that the
companies that are collecting this data must be using it for something.
Q775 Ian C. Lucas: When you obtain that data for a particular client, do you
retain that data as Cambridge Analytica and then use that as a resource
for other clients? Is that what you do?
Alexander Nix: As I have already mentioned to your colleague, that
depends entirely on the relationship that we have with the client and also
the territory that we are operating in and what the legislation is. It is
case by case.
Q776 Ian C. Lucas: Let's talk about the UK. If you collect data for an individual
client as Cambridge Analytica, do you use that data for another client?
Alexander Nix: The client data that we collect for clients in the UK
belongs to the clients, who are ultimately the data controllers. We are
just processing or, in this case, collecting, which is part of the processing
function for these clients.
Q777 Ian C. Lucas: In those circumstances, you never collect data yourselves
as Cambridge Analytica? You only collect it for clients?
Alexander Nix: In those circumstances, yes.
Q778 Ian C. Lucas: But you do collect data for yourselves sometimes?
Alexander Nix: For instance, Mr Collins filled out a survey on our
website, or nearly filled it out. That would be data we are collecting for
Q779 Ian C. Lucas: That is very clear, but within the UK you would need the
consent of the individual who is supplying you with the information in
order to transfer it to another client.
Alexander Nix: That is correct.
Q780 Ian C. Lucas: But that would not apply in the United States.
Alexander Nix: Again, it might apply. Some customers might say
specifically, "Our data is our data. Do not share it" and that comes down
to the discussion you have with them at the time you engage with those
clients about how they want their data treated.
Q781 Chair: Mr Nix, the purpose of your surveys is to support psychological
profiling of people, isn't it?
Alexander Nix: The purpose of our psychological surveys is to do that
but, as I just mentioned, we were undertaking up to 400,000 surveys a
month for five months. These were not psychological surveys at all.
These were just political surveys trying to understand what issues were
most relevant to which audiences and to help us understand our resource
allocation, our targeting, our messaging and so forth. They were nothing
to do with that.
Q782 Chair: I appreciate the questions are being framed in that way, but the
way the OCEAN process works is to analyse people's answers to different
sorts of questions and, from that, to develop a profile of the sort of
person they are, the world view they have, what their motivations are,
what makes them happy and what makes them sad.
Alexander Nix: The OCEAN methodology was simply one of many
methodologies that came out of experimental psychology to help
understand behaviours. Your Government have an—I am going to get
this wrong—institute of behavioural science that they use to help
understand how to increase people's tax payments or to encourage
people to reduce smoking and so forth. All they are doing is taking
academic literature in order to understand audiences in order to increase
compliance, often for very importance issues.
Q783 Chair: But if from your surveys you are saying that we know someone is
frightened of crime and they have concerns about gun use and about
whether Hillary Clinton is weak on crime, what you are doing is building
up a psychological profile of someone and you are using that data to
target them with a message. That is the purpose of these surveys.
Alexander Nix: If you can identify that an audience group is frightened
of crime and that is really important to them, you can then share with
these audiences your candidate's policies on how they intend to tackle
crime and how they intend to address a major fear of those constituents,
and I think that is really healthy. These people have an identified
problem—they are frightened of crime—and you are able to say to them,
"Don't be frightened of crime because look what our candidate is going to
do. We have set out our policy. This is our position on crime," and you
make sure that that information gets to the people for whom it is a
worry. That has to be good.
Q784 Chair: You could send a message saying, "You are right to be frightened
of crime because the other candidate is weak on crime and if they win
you and your family is in danger". I think anyone would recognise that as
a kind of psychological profiling. You have other layers on it as well, such
as, "Are you frightened about immigration? Is immigration the cause of
crime?" We saw a lot of messaging like that around the referendum
campaign here as well. That is not just data analytics and people answers
to individual questions; that is using that data to build up a psychological
profile of individuals and then target them, isn't it?
Alexander Nix: I can't speak to the UK referendum but—
Chair: It applies to any election or any campaign, and it certainly applies
to the American campaign.
Alexander Nix: I think I have made my position clear, which is that we
are trying to make sure that we can use data to understand what people
care about and we can seek to address those concerns. If those are fears,
we can allay those fears by telling audiences how we are going to solve
those problems and that has to be good.
Q785 Simon Hart: If you can identify a section of the audience that is
expressing some concerns, perhaps about immigration or gun crime or
whatever it is, are you arguing that what you do is help to allay those
fears or is it the accusation that has been made that all you do is
oxygenate those fears in order to suit the guy who is paying you the big
fee? Which of those accusations is correct? You are making it sound like
you are doing a public service.
Alexander Nix: We are doing a service to our client. Our job as a
campaign consultancy is to make sure that we provide the best
communication technologies and methodologies in order to allow our
clients to get their messages across.
Q786 Simon Hart: I am sorry, I did not put it very clearly earlier on. Is that
the same thing as when you identify an area where fear may be a factor?
Are you saying that you do not contribute to exacerbating that fear, you
do not then develop messages that make people perhaps more fearful
than they previously were, rather than less fearful?
Alexander Nix: I think you need to look at campaigning over the last
100 years. Negative campaigning is a part of every campaign regardless
of the technologies that are being embraced at any given time. The
ability for one candidate to stand up and say, "You know what, under this
particular candidate or political party the country is going to be worse off.
You are going to have less money in your pocket, you are going to have
more crime," is just an integral part of the political process.
Q787 Simon Hart: That is true, but they are the candidates, they are the
name on the ballot paper.
Alexander Nix: And their campaign teams are doing exactly the same
and you well know that, as does everyone in this room. Extolling the
virtues of your candidate and the weaknesses of your opposition is a fair
practice in political campaigns globally.
Q788 Simon Hart: In countries that have similar electoral rules as this
country, and they vary significantly, how do people account for your fee
when they are making their declarations over election or referendum
expenditure? Into what detail do they necessarily go? I accept that you
may not be able to answer that.
Alexander Nix: If it is their responsibility to report on their fees in their
own country, the onus is on them to do so.
Q789 Simon Hart: Are you ever asked to explain or to provide some kind of a
brief description of what the service actually is or do you simply—
Alexander Nix: At its broadest level we are providing campaign
consultancy and communication services, but if anyone wanted more
specific details, we are more than happy. The contracts that we engage in
are based on a statement of work that is mapped out with the client and
these are well documented. It would be very easy for us to point to these
agreements that detail exactly what areas we are working on.
Q790 Simon Hart: You mentioned that you have 4,000 or 5,000 data points on
every adult in the United States—the entire voting population. Does every
adult in the United States know that you have 4,000 or 5,000 data points
Alexander Nix: I can't speculate on what every adult in America knows.
That would be absurd.
Q791 Simon Hart: It is very closely related to the earlier question about the
extent to which people—you say you are not a data miner. How do you
acquire that vast quantity of data without a certain amount of mining? Is
it not a responsibility of yours to be able to ensure that the population is
aware that vast quantities of their data, personal and otherwise, is held
by you and used for electoral purposes?
Alexander Nix: We have made no secret of this fact, as you all know,
because you have referenced at least two occasions that I have stood in a
public forum and talked about the methodology that we use. That
includes the data that we underpin it with. I think we have been pretty
Q792 Simon Hart: But you said you were not a data miner?
Alexander Nix: Well, we are not because there are companies out there
whose singular purpose is to go out and collect and aggregate data, as Mr
Collins said. With his background, he is very familiar with the experience
in the Axioms and Infogroups and other very large companies who have
hundreds, or indeed thousands, of employees whose singular job is to sit
on the phone and speak to companies and acquire their data, match it
together, hygiene it, put it into a database and record, such that they can
then license these data to companies like ours. All we have done is gone
to all the vendors, large, medium and small, and taken these data and
put them into one database and record.
Q793 Simon Hart: That then excuses you from the accusation that you are a
miner? The fact that you are just mining what other people have mined
does not contradict in any sense what you said earlier on?
Alexander Nix: I don't like the word "accusation" because that implies
that we are doing something wrong. This is an established business in the
United States, which is selling data, and we, like many or most brands
and many or most agencies, are able to go out and license these data for
Q794 Simon Hart: At no stage was I suggesting there was anything illegal
about it. I was simply saying that you make a virtue out of the fact that
you possess probably more data on the entire voting population of the
United States, thereby making a political point, than anybody else in the
market. My question was simply to the extent to which that is known
about—I know we know about it, we are talking about it here—and the
extent to which the individual voters know exactly how and would have
access to that information should they require it. If they come to you,
you would be able to disclose the 4,000 to 5,000 data points that you
possess. If I was an American citizen, would you provide me with those
4,000 to 5,000 data points were I ask you for them?
Alexander Nix: Let me address your first question. We are incredibly
proud of the fact that we walked into one of the most competitive political
markets, if not the most competitive political market, in the world as a
small British tech company and were able to develop the sort of
technologies and methodologies and bring them to market so effectively.
This is going to help the communications landscape way beyond politics.
It is going to help in advertising and marketing and make it more
relevant and much more economical. In terms of what you have asked
about American people undertaking what we call in Europe a subject
access request, the legislation is not currently in place in America for
them to do that, but were it there, we would be able to provide exactly
the same service that we provide for companies in the UK and across
Europe. Following GDPR, we will be providing that for very many
businesses to help them manage their own data as they seek to be
compliant with the new legislation that is being implemented.
Q795 Chair: Mr Nix, when Leave.EU applied for designation to the Electoral
Commission to be the official leave campaign, it named Cambridge
Analytica in its designation document. Why was that?
Alexander Nix: I was not aware of that but I can only assume, as I
mentioned before, that they felt that associating themselves and aligning
themselves with Cambridge Analytica would give them extra credibility
and leverage in trying to compete in a bidding process where they were
clearly the underdogs to be the designated campaign.
Q796 Chair: Has the Electoral Commission raised this with you as part of its
Alexander Nix: No, it hasn't.
Q797 Chair: That is slightly strange. It is doing an investigation looking at
Leave.EU's activities in the referendum. Your company is cited in their
designation document as being someone they are working with and it has
not asked you about that.
Alexander Nix: Again, let me circle back to you, but I have not been
asked that question. I can certainly find out for you whether my data
compliance team or colleagues have been asked it. I would like to think
that this inquiry has been going on for some time and we are delighted to
help because we really want to make it clear that we did no work, as I
have been trying to do today in this Committee. I am hopeful that the
Electoral Commission and the ICO have taken on board the evidence that
we have presented to them and that they are going to arrive at the same
conclusion as I hope you will, which is that we were not involved,
therefore we can't speak to these things.
Q798 Chair: That itself is a matter for the Electoral Commission. It is not
something we are investigating but it is just another point of information
that is out there in the public domain that links the two of you.
I have a few follow-up questions and then I think we will be done. If you
were conducting 300,000 to 400,000 surveys a month for the Trump
campaign over a five-month period, so let's say nearly 2 million surveys,
do you or your associates hold the data that was gathered from that
Alexander Nix: These data belong to the Trump campaign.
Q799 Chair: Okay, so you don't have any ongoing access to that?
Alexander Nix: Again, I will circle back to you on those specific pieces of
research, because I don't know what the data-sharing agreement was
with the campaign on that specific piece of research. Generally these
would belong to the campaign, but if they have permissioned us to retain
a copy of them, we would have a copy of them.
Q800 Chair: Thank you. Presumably with the Cruz campaign, you did have
permission to have a copy of the data, because you said some of the data
from the Cruz campaign could have been used in the Trump campaign.
Alexander Nix: That is correct.
Q801 Chair: From a layman's point of view, how is this data held? The data
must exist in a form that means it can be used for one campaign and
then repurposed for another campaign. Do you have data storage centres
where you keep this data or how does it work?
Alexander Nix: That is right. Some data is stored in secure facilities and
some is stored in the cloud, depending on how we need to access it.
Q802 Christian Matheson: You took on Sophie Schmidt as an intern. Why did
she want to come and work for you?
Alexander Nix: You are going to have to speak to Ms Schmidt about
that. I can't speculate.
Q803 Christian Matheson: When you were interviewing her, did you not say,
"Why do you want to come and work for us?"
Alexander Nix: I would like to think that we were a company that she
found interesting and exciting to work for.
Q804 Christian Matheson: When she went back to America, is it likely she
then introduced you to some of the senior players or the better known
players in the tech world, such as her father, who is the boss of Google,
and Peter Thiel, who is obviously very well known in the area as well? Did
she introduce you to Peter Thiel?
Alexander Nix: No, she did not introduce me to her father and she did
not introduce me to Peter Thiel. That is not correct.
Q805 Christian Matheson: She has now gone to work for Uber. Have you
shared any data from Google? Has Google given you any data?
Alexander Nix: As far as I am aware, Google, like Facebook, is a walled
garden and does not share its data. It certainly has not shared any data
Q806 Christian Matheson: That is you, as in Cambridge Analytica and SCL.
Alexander Nix: That is us in the broader sense of the word.
Q807 Christian Matheson: What about Uber? Have they provided you with
Alexander Nix: We don't work with Uber at the moment.
Q808 Christian Matheson: Did you previously?
Alexander Nix: No, we have never worked with Uber.
Q809 Christian Matheson: Okay, so there has not been any sharing of any
data from Uber to any of your companies?
Alexander Nix: That is correct.
Q810 Christian Matheson: Thank you. Can I talk about the process here? You
must deal with huge amounts of data, and the Chairman was asking
about the way that you hold it and manage it. When you bring it all
together, do you aggregate it yourself or do you use a company? Do you
buy in aggregation services?
Alexander Nix: No, we do it ourselves. That is part of the service that
we offer to our clients, and we do that both manually and have products
that we have developed to automate some of that functionality.
Q811 Christian Matheson: Have you ever used a third party in the past to do
Alexander Nix: Never.
Q812 Christian Matheson: What is your relationship with Aggregate IQ?
Alexander Nix: We have no relationship with Aggregate IQ. We have
historically used Aggregate IQ to develop some software for us. It was a
standalone project that lasted about six months, possibly, in 2014. My
understanding—I will have to check this—is that we have not had any
communication with them since early 2015.
Q813 Christian Matheson: There was a licensing agreement in September
2014. Is that what that was?
Alexander Nix: That is right, yes. They built a small piece of software
for us, as a software development company.
Q814 Christian Matheson: I have a diagram of the structure of the group,
and it seems very complicated, with a bit of ownership here and a bit of
ownership there and some shareholdings here and there. In section 10 of
the latest accounts of SCL Elections Ltd there was a £24.2 million
payment. Does this relate to the 19% of Cambridge Analytica in America
LLC that is owned by SCL Elections or is it payment for work that would
have been undertaken by SCL Elections?
Alexander Nix: You are going to have to help me understand the
relevance of the question of ownership of a private company to this
inquiry into fake news, data and communications.
Q815 Christian Matheson: Fake news obviously is the content that is being
put out there, but it is also a question of the delivery method and how
fake news is propagated. The structure of your companies is such that it
is not quite clear not only who is owning them, but who is propelling that
means of delivery. I am just quite curious.
Alexander Nix: We have never published a structure of our company, so
I don't know what you are looking at or where that has come from, but
as a private company we don't speak about our structure, our investors
or our board members. If you would like more information on this, we
might be able to take this out of the public forum in the interests of
helping your Committee, but I don't think that is something I want to
Christian Matheson: Okay. I will leave it at then. Thank you.
Q816 Paul Farrelly: You said that you had worked across the world in political
campaigns. Could you tell us a little bit more about where else in the
Alexander Nix: Again, in the interests of our clients, as a rule we do not
speak about client contracts unless we have the specific permission of
those clients, and that includes commercial, Government and political
contracts. What I can say is that we undertake eight or nine elections
every year, and we are not limited by geography, so this really could be
from the Caribbean to Asia to Africa to Europe or everywhere. Some of
these are very large, very important national elections and some are
smaller, more local mayoral or state elections. It is really anywhere that
you can think of.
Q817 Paul Farrelly: For the Ukraine?
Alexander Nix: Potentially.
Q818 Paul Farrelly: Have you?
Alexander Nix: Well, as I said, we do not talk specifically about clients
but there are elections coming up in the Ukraine in the future. If there is
a good commercial opportunity there, we might look at it. I would have
to speak to my elections team.
Q819 Paul Farrelly: Would you work for anyone?
Alexander Nix: I think I have already addressed this. We only work for
Paul Farrelly: One person's despot might be one person's hero, but
generally there are certain people who are unsavoury.
Alexander Nix: We work for mainstream political parties. We try to work
only in free and fair democracies, and we also have to be mindful of our
other divisions. As I have already told this Committee, we do an awful lot
of work for the British Government, the US Government and other allied
Governments. If there is any question whatsoever about a client that we
might take on in the political sphere, or even in the commercial sphere,
we always discuss this with the relevant parties in the US and in the UK—
so that would be the Foreign Office or the State Department—saying, "We
have had an inquiry to work in this country. Do you have any objections
to this?" It is in the interest of us to make sure that we are not building a
business over here that could damage our part of our business over here.
Q820 Paul Farrelly: With candidates and parties, would you work for
campaign organisations such as super PACs?
Alexander Nix: Yes, we have done a number of campaigns on behalf of
Q821 Paul Farrelly: All on the Republican side?
Alexander Nix: That is correct.
Q822 Paul Farrelly: What about campaigning organisations like the American
Enterprise Institute? Would you work with that sort of organisation?
Alexander Nix: Actually, I am not familiar with them, but we do work
for organisations and lobbying and advocacy groups across America. I
can't speak to them, but I can speak to other ones.
Q823 Paul Farrelly: We have only the briefest biography of you that starts
with Manchester University and ends with you joining SCL in 2003. How
did you get into all this? How did you get into this line of business?
Alexander Nix: A huge interest, I think. Fundamentally, I was working
in corporate finance, which I did not find particularly fulfilling.
Q824 Paul Farrelly: I was the same, actually. Where were you in corporate
Alexander Nix: At a small UK merchant bank up the road from here,
and I was looking for an opportunity to work in a slightly more relevant
and fulfilling occupation.
Q825 Paul Farrelly: You are not a data processing PhD yourself?
Alexander Nix: I am not a data PhD myself.
Q826 Paul Farrelly: What qualifications do you have?
Alexander Nix: In terms of this company, I lead a management team so
I don't need to be qualified as a data scientist. I need to be qualified to
run a business.
Q827 Paul Farrelly: Lots of people in corporate finance are accountants or
lawyers or, like me, none of them.
Alexander Nix: As the CEO of a company, you know perfectly well that
there is so much you can learn in school and then there is a lot you can
learn in life. We have been doing this for many, many hours a week for
many weeks and for many years now and hopefully the fruits of that
labour are beginning to come together.
Q828 Paul Farrelly: Did you start off as a lawyer or an accountant or just a
Alexander Nix: No, neither of those, just a generalist.
Q829 Ian C. Lucas: Do you exercise any editorial control over the messages
that your clients send to, for example, electoral and political campaigns?
Alexander Nix: All the messages that we propose to campaigns are
signed off by the campaigns themselves and go through campaign legal.
There is an internal compliance structure to make sure that we are not
infringing any legislation of the FEC or any other body that might be
governing the work that we do.
Q830 Ian C. Lucas: You propose messages as part of your role. In other
words, the initiative comes from you and then it goes to the campaign.
Alexander Nix: That is correct. We interpret and draw insights from the
data, and we use those insights to devise the messaging strategy and the
messaging content. We then share these messages and content with the
campaign, and we discuss our strategy with them. Often there is a
dialogue about that and some tweaks are made. We then push that
through legal and compliance. They will give us their feedback and
ultimately the messages are disseminated.
Q831 Ian C. Lucas: Presumably the campaign makes proposals to you and a
similar process continues. In other words, they would have an idea for a
message that they wanted to deliver to the elector and they would
present that to you. Would that be a situation that arises?
Alexander Nix: It might well do and then we might go out and roll out
one of those surveys that I talked about. We can go and test that
message, or A/B test it, digitally for instance, and we could give them
empirical feedback about which message was likely to perform better,
how it should be run and who should be targeted.
Q832 Ian C. Lucas: Have you ever rejected a message from a campaign on
Alexander Nix: We run possibly thousands. I think in the last election in
the US we ran 4,000 different advertising campaigns—about 1.4 billion
impressions. We served for five months. I cannot speak to that.
Q833 Ian C. Lucas: Can I just say why I am asking you the question? You
have been very keen to emphasise the benevolence of the role that your
company is performing, fulfilling the public good of informing electors
about particular candidates and making campaigns relevant. But we are
all grown-ups, here. We are all politicians on this side of the table. We
know that there is negative campaigning and there is positive
campaigning, and one of the dangers of mass communication in this
format is reinforcing, for example, very negative stereotypes. Do you
agree with that?
Alexander Nix: As politicians who understand campaigning, I think you
will also understand that winning elections is not about reinforcing
prejudice on either side of the political spectrum. There is no point in
telling hardcore Republicans how bad Clinton is and how good Trump is,
or vice versa. It is about correctly identifying the people that sit in the
middle—the persuadable or swing voters—and presenting to them very
well articulated facts on the particular policies and issues that they care
about most, so they can begin to make their opinions.
Q834 Ian C. Lucas: I agree with all of that, but it is also about finding those
same people and maybe reinforcing fears that they have, or perhaps it is
about emphasising bad aspects that they may possibly believe and that
could be reinforced. That is another way of persuading people in different
Alexander Nix: In the case of negative campaigning, you would be
right, and in the case of positive campaigning, it would be about
emphasising the hopes they have—aka Obama 2008.
Ian C. Lucas: Absolutely.
Alexander Nix: It works both ways, and we are no strangers to positive
and negative campaigns.
Q835 Ian C. Lucas: I am talking about one particular way and, as you know,
this is about fake news. The reason I am asking you has there ever been
an ethical reason why you have refused to run a particular ad is because
I am aware of examples of campaign ads that I would not use in my
campaign, and I am a politician. Are you aware of any example of that
kind, where Cambridge Analytica has said, "We are not going to do that"?
This is very relevant to the question of fake news.
Alexander Nix: I would like to say to you, sir, that I am sure there are
dozens of examples, but—
Ian C. Lucas: Can you go away and bring us some?
Alexander Nix: To go through the 1.4 billion impressions that we served
in last year's US elections, look at each one that we served, look at all
the ones that we rejected and come back to you would not be a
reasonable request to put on us.
Q836 Ian C. Lucas: What level of control is there over the ads that you project
Alexander Nix: I just discussed this with your colleague. The adverts
that we propose to the campaign are shared with the campaign and the
campaign signs them off. That goes to legal, and legal and compliance
need to sign them off. There is a process. There are checks and balances.
Q837 Ian C. Lucas: Can you please give one? I am not asking for 1.2 million.
Alexander Nix: Billion.
Ian C. Lucas: Okay, billion. I am asking for one example. Are they all
different from each other?
Alexander Nix: Yes.
Q838 Ian C. Lucas: Every single one is different?
Alexander Nix: No. There were some 4,000 campaigns that were
Q839 Ian C. Lucas: All I would like to see is one example of an ad where
Cambridge Analytica said, "We're not going to allow this to go out. We
Alexander Nix: That is certainly something that we will look into.
Ian C. Lucas: Thank you very much.
Q840 Chair: Thank you, Mr Nix. Just a couple of final questions. We have
spoken quite a lot about the way in which you store, manage and gather
data, and how you seek to comply with the Data Protection Act in the UK.
Rather than going through any further detail today, perhaps you would
be able to write to the Committee to set out what your policies are on
how you source data, manage it, share it with third parties, and how you
ensure that you do so in compliance with the Data Protection Act?
Alexander Nix: Certainly. We will get back to you on that.
Q841 Chair: You know that Julian Assange claimed that Cambridge Analytica
approached WikiLeaks to work with them and that they rejected the offer.
That is a statement that he has put out. I know we touched on this
earlier, but can you confirm whether any approach has ever been made
to Julian Assange by Cambridge Analytica?
Alexander Nix: Yes, certainly. I would be happy to speak to that as I
did, I think, in front of a very large audience in Lisbon last year. This was
at the time, as you will remember, when the newspapers and the news
channels were reporting that Julian Assange had access to a large
quantity of information that could be incredibly relevant to the outcome
of the US election. We read about these claims. We had no idea, as no
one did, whether this was true or not so we simply reached out to a
speaking agency that represents him—that was the only way we could
find to get hold of him—and said, "Would you pass him a message
commenting on this and asking whether he would like to meet to discuss
this?" and we received a message back through this third party, the
intermediary, saying no, they would not. That was it. We, like probably
every other journalist in this room, were very keen to find out what was
in these data and whether they would have an impact. We were all
Q842 Chair: You said earlier on that you gathered this large amount of data for
the Trump campaign as part of the survey work that was done and that
you will write to us to say whether that was data that you held or the
campaign held. If that was being gathered by you, obviously on behalf of
the campaign, would other people in the campaign have access to that
data and the ability to share that with third parties without your
Alexander Nix: Quite unlikely. Hypothetically, anyone could possibly
have taken advantage of that but they would have had to have been
someone on the inside who had taken the data illegally.
Q843 Chair: Yes, but while you do not know that is the case, it would be
technically possible, even if hypothetical?
Alexander Nix: It is technically possible—and I am certainly not
suggesting this—that an employee may have illegally taken the data and
passed it elsewhere, yes.
Q844 Chair: Thank you. Again it is relevant to our question. I know that it is a
case relating to America, but to have some written evidence from you
about the protocols you have for data management and how you can
make sure that you keep people's data secure, and do so without being in
breach of the Data Protection Act, would be very helpful.
Finally, on other countries where you have worked, have you ever
worked in Russia or on behalf of Russian companies or organisations?
Alexander Nix: We have never worked in Russia. As far as I am aware,
we have never worked for a Russian company. We have never worked
with a Russian organisation in Russia or any other country. We do not
have any relationship with Russia or Russian individuals.
Q845 Chair: If a Russian company came along and said, "Would you work for
us?" would you do it or would you reject that?
Alexander Nix: There are many companies in Russia that are stand-up
companies doing normal and fair business, so we would have to evaluate
that but, given the current climate, I do not think that would be
necessarily our first client of choice.
Q846 Chair: Would you ever work on political campaigns in a third country on
behalf of someone else?
Alexander Nix: We have worked on advocacy campaigns and we have
worked on communication campaigns that have been for the benefit of
other countries who wish to target other audiences. For instance, if the
UK wants to drive tourism in America, we might do a campaign for the
Government here. As a general rule, however, we look at these things
Q847 Chair: A totally hypothetical example—not one that I have been given
but a hypothetical—is the referendum in Catalonia last year. If, say, a
commercial entity said, "I have a big interest in the outcome of that
referendum because that region is commercially important to me. I want
to run a campaign within that region that would encourage people to stay
as part of Spain," would you take on a project like that as a UK-based or
American-based business, working maybe for a company in another
country but targeting voters in yet another location?
Alexander Nix: In that hypothetical instance, we could not necessarily
engage with a company, but if the company had an arrangement with a
political party, and that was between them and they wanted to help that
political party and the party brought us in, again those sorts of
discussions would be outside of our remit and they would never need to
involve us. We would be engaged by a political party. If they had a
relationship with a business or another third party regarding financing,
dependent on the regulations in their particular country, that might be
Q848 Chair: Obviously it depends on the legislation because you could be in
breach of electoral law to be receiving funding from a third country to
conduct political campaigns in another.
Alexander Nix: Absolutely. That is why, again, we look at these things
very carefully. We have an in-house legal team. We have external legal
teams. We choose our clients very carefully. We would never want to put
ourselves in a position like that. Of course not.
Chair: Thank you. I think that concludes our questions this morning.
Thank you very much for your time.