In 2014, GSR licensed data on 30m individuals to SCL Elections. We received this data in good faith, trusting that it complied with Facebook’s terms. We didn’t break any laws.
In 2014, SCL Elections contracted Dr. Aleksandr Kogan of Cambridge University to undertake a large-scale research project in the United States. His company, Global Science Research (GSR), was contractually committed to only obtain data in accordance with the UK Data Protection Act and to get the informed consent of each respondent. GSR obtained Facebook data via an API provided by Facebook, which permitted the survey to collect information on both respondents and their friends. We did not “hack” Facebook. The contract with GSR was for U.S. users only. We have not broken FEC regulations. We also have not illegally or inappropriately collected or shared data with third parties.
Our work with the data was disappointing, and statistically only slightly more accurate than random. We decided to replace it with our own research.
We used the GSR data to train a model to predict personality using commercially available demographic, consumer and lifestyle data. It's important to note that the GSR data was already a model built upon Facebook information, so we were trying to build a model on top of already modeled data. In out-of-sample testing, it was found that this model-on-a-model was statistically only slightly more accurate than random. The approach was then abandoned. We went on to collect our own personality data using a research panel – under an appropriate statement of consent – and built more accurate personality models on this data. Of course, marketing psychology is just one of the tools that we use to better communicate with audiences.
We deleted the raw data from our file server, and searched and removed any of its derivatives in our system. We’ve certified this to Facebook.
Our contract with GSR stated that all data must be obtained legally, and this contract is now a matter of public record. When Facebook contacted us to let us know the data had been improperly obtained, we took action against GSR and immediately began deleting the data. When Facebook sought further assurances a year ago, we carried out an internal audit to make sure that all the data, all its derivatives, including the models described above, and all backups had been deleted. We gave Facebook a certificate to this effect.
We’re sorry that we didn’t scrutinize the GSR data once we received it. But we’re committed to proving that it’s been deleted.
A world-leading data forensics firm will be carrying out an independent data audit on our company. We’re cooperating with the UK’s Information Commissioner’s Office (ICO) in their investigations into data analytics for political purposes, and we’re confident they’ll agree that we no longer hold the GSR data. People are understandably concerned about the use of such personal data, particularly for political messaging, however we want to make it clear that it was not used for the 2016 US presidential election and that no traces of it exist in our work.
We used voter files from the RNC, polling data, data from the campaign, and data available from our commercial brokers.
We used the data to identify “persuadable” voters, how likely they were to vote, the issues they cared about, and who was most likely to donate. We also built a polling tracker for every key state and provided dashboards for the campaign, including the group that planned the candidate’s travel agenda. Our analysis was also used for targeted advertising. In truth, we used the same kind of political preference models used by the Obama and Clinton campaigns; however, we started five months out from election day and did it with far fewer resources and less data.
We held preliminary meetings with Leave.EU to explore working together, but we delivered no work and the relationship ended.
This is a matter of public record. Attempts to link us with Aggregate IQ (AIQ) have been fuelled largely by conspiracy and hysteria. For clients completely unrelated to Leave.EU, we subcontracted some digital marketing and software development to AIQ. The suggestion that we were somehow involved in any work done by Aggregate IQ for Vote Leave in the 2016 EU referendum is entirely false.
His presentation in the media has been distinctly uncritical, despite the fact that he pitched competing services to identical prospects.
Mr Wylie has repeatedly claimed to be a founder of Cambridge Analytica. In reality, he was a contractor for SCL Elections and left in July 2014. He has no recent knowledge of our business or its practices, and has admitted as much in his testimony to parliament in Britain. He set up a rival company called Eunoia Technologies, which pitched the Trump campaign and later tried to work for Vote Leave. Our lawyers took action against Mr Wylie to prevent his misuse of the company's intellectual property while attempting to set up his own rival firm. Much of his account is based on pure conjecture and guesswork, while his own motivations in this saga have remained unexplored.
While the company started by working for Republican clients in the US, globally we work across the mainstream political spectrum.
Recent media coverage has sought to portray the company as somehow politically or ideologically-driven. As anyone who is familiar with our staff and work can testify, this couldn’t be further from the truth.