Presidency Lost: No Cambridge Analytica

Considering the outcome of the 2020 presidential election and the upcoming inauguration, it is a good time to ponder what would have happened if Cambridge Analytica (CA) had not folded in 2018 and had worked on Trump’s re-election campaign. Would the outcome have been different?

In the 2016 presidential election, CA was running 1,500 surveys a week in every battleground state. These sample sizes are larger than what are fielded for most polls and as a result, CA had a better handle on what was happening in the electorate than did the pollsters or the U.S. public.

CA knew which groups of people would definitively vote for Hillary, definitively vote for Trump, which were the most persuadable (undecided), who would go out and vote and who needed a nudge. Moreover, CA’s models were constantly being updated as people made up their minds and the campaign ran its course.



For any political campaign, it’s that persuadable group that garners the most attention.

Well-run campaigns know where all of those undecided voters stand on the big issues, the best way to reach them and what they need to say to convince them that their candidate is the right one. This isn’t a Republican or a Democratic practice -- it is a survival mechanism for any candidate competing for a major office in U.S. politics.

To be fair, there were some key former CA employees who worked on Trump’s 2020 reelection campaign as part of another political consultancy called Data Propria. Trump wasn’t completely without CA and still lost, although some of CA’s most experienced and talented people were working elsewhere.

With regard to CA, it is important to note that the United Kingdom’s data-privacy watchdog, the Information Commissioner’s Office or ICO, cleared Cambridge Analytica of any illegal activity as it relates to Brexit. This came after a two-plus-year-long investigation that included the analysis of 42 computers, 700 terabytes of data, 31 servers, and more than 300,000 documents.

In a letter to MPs, Elizabeth Denham, the information commissioner, said there was no evidence of law breaches. According to The Guardian, the report also said the investigation concluded that “based on what we found it appears that this may have been an exaggeration” and much of the company’s activities followed “well recognized processes using commonly available technology.”

Despite the trouble with Facebook and the resulting negative press, CA was for the most part well run and strict about legalities. The company’s COO was a man by the name of Julian Wheatland, one of the former CA people who appeared in the documentary "The Great Hack." He ran a pretty tight ship.

Unlike most political consultancies that handle one candidate at a time within a given race, CA would manage multiple campaigns for candidates in the same race as well as for some of the super PACs. 

The people working on the candidates’ campaigns were by law not allowed to collaborate on the super PAC campaigns. The reason for this has to do with campaign versus super PAC contributions. The former is limited, and the latter is not. The company was strict about keeping these separate.

For instance, if an employee was working in a firewall for the Donald J. Trump for President (DJTfP) campaign, they were not permitted to discuss the campaign with anyone working on the Make America Number One (MAN1) super PAC campaign.

It is interesting to note that CA’s MAN1 campaign was very successful. In fact, a lot of people might be surprised to know that it was awarded The Advertising Research Foundation’s (ARF’s) David Ogilvy award in the Big Data category (2017).

One almost laughable aspect of the CA/Facebook scandal is that the reviled personality segments that have become the subject of documentaries and whistleblower folklore never needed Facebook data to be created in the first place. They can be modeled from seed segments using regular old online surveys. The use of this data is legal, and it may be used for the benefit of any advertiser or political party.

In fact, today these audience segments are available for purchase on a cost-per-thousand basis in the syndicated digital marketplace. They are one tactic among many that may be employed by marketers who are all doing their best to achieve the highest returns on their media investments.

There is a lot that people don’t know about CA. For instance, the company did not set out in advance to work for Donald Trump. In 2015, the company was working for Ben Carson and Ted Cruz. As everyone knows, these candidates didn’t make it through the primaries.

CA was a business and businesses need clients, so when Carson and Cruz lost to Trump, CA cold-called the Trump campaign and miraculously landed the business.

If you are an NFL football team you want to play in the Super Bowl, and if you’re a political consultancy you want to work in the presidential election.

The fact that CA was owned by a billionaire known to contribute money to conservative candidates and causes likely played a role in the win. Regardless of how people feel about CA or Trump as a candidate, this was a huge accomplishment for a new start-up company that had only recently entered the U.S. market.

It is also not widely known that CA was in the process of moving away from politics into the commercial vertical.

The plan was to make a name for itself in the 2016 presidential election and use its newfound fame to win business with commercial clients.  It had staffed up to work in a variety of categories including CPG, healthcare, TV, insurance, publishing and yes, for faith-based, clients. The then CEO, Alexander Nix, was trying to move the company out of politics.

CA was having a lot of success shortly after the 2016 presidential election until the negative press started and the Facebook scandal made it almost impossible to close business.  

An interesting question to consider is: What if CA had worked for a different, less controversial candidate in the 2016 presidential election? Would the company still have had all the lawsuits and folded? Would Facebook have been fined $5 billion dollars for a data breach? Would Channel 4 in the UK have launched an undercover sting operation to discredit the company? If either Ben Carson or Ted Cruz had come out on top in the primaries and was elected president in 2016, the perception of and story around CA would have likely been very different than what it is today.

The truth is that no one can say with any certainty that if CA had worked on Trump’s 2020 campaign he would have won.

Causality for political campaigns is difficult to parse out. Strategy will also play a role. Brad Parscale, Trump’s former head of digital, is on record for saying that Trump would have won in 2020 if he had been more empathetic with his response as it relates to Americans who are struggling as a result of the COVID-19 pandemic. 
 post-game strategy advice may be accurate; however, without any empirical evidence it is merely a hypothesis. 

Parscales may have data to support it. However, it does not appear to have been shared. An interesting question for the media to ask Parscales would be: Do you have any data that you could provide to support your Trump strategy claim?

Hard attribution data isn’t generally available for political campaigns, and there are a wide variety of marketing channels -- social, digital, TV, direct mail, door-to-door, public relations and earned media -- that all contribute to the final outcome. Some of these are outside the political consultancy’s control. 

The candidate’s likability, track record, debate performance and record while in office are contributing factors as well. 

Data Propria had enough CA-expertise to do in 2020 what CA did in 2016 and it probably used the same National Republican Committee (NRC) data. With this in mind, the outcome would have likely been the same whether or not CA were still in operation and working on Trump’s campaign.

Data, data science, marketing expertise and budget are not the only deciding factors in a political election. The idea that data is breaking democracy didn’t play out in 2020.

It would seem that data is only perceived as a problem when the outcome of an election is unpopular. No one is complaining that data caused Biden to win. In 2020, instead of data, the excuse du jour is voter fraud.

Trump won fair and square in 2016 and lost fair and square in 2020. Just as it was a lie that data broke democracy, it is a lie that Trump lost the election due to election malfeasance. 

The rioting on Capitol Hill over the certification of Joe Biden’s presidency was a sad day for U.S. democracy. People died and were injured because the pursuit of power was a higher priority than the safety and well-being of the U.S. populace.

3 comments about "Presidency Lost: No Cambridge Analytica".
Check to receive email when comments are posted.
  1. Michael Abramson from Dominant Trait, January 21, 2021 at 11:02 a.m.

    Are you actually saying Facebook segmentation data which scrapes data from their own subriber base (plus instagram) is as good as seed pool data + some probablistic modeling (which is a joke when used at scale)?  The quality of that data is day and night.  The reason CA/Facebook data was so valuable was that they were able to tell you specially who the Jew haters were, who the Muslim haters were, who the black haters were.  2016 was NOT a case of finding the 'uncertains' but instead to agitate many of the less educated to be certain that they not only voted but also everyone they knew to vote as well.  Facebook data allowed for dozens of echo chamber messages to exist and the user to create messages, often of hate and lies, specific to each of those chambers.  You could NEVER get to that level just using seed data and look-a-like modeling - NOT EVEN CLOSE.  

  2. David Carroll from Parsons School of Design, January 22, 2021 at 12:38 p.m.

    There are some concerning factual errors in this piece. It should be reviewed, corrected or amended.

    For example, the claim that Cambridge Analytica maintained a clean firewall between campaign and PAC clients is disupted with forensic proof released into the public domain. See Associated Press, October 16, 2020 "Watchdog org: Trump '16 Trump '16 campaign, PAC illegally coordinated."

    The complaint also alleges that Cambridge Analytica — which improperly acquired and used 87 million Facebook users’ profiles to predict their behavior — had a shared project calendar for both entities, among other evidence.

    “The idea that this spending was at all independent is farcical and these emails underscore that,” said Brendan Fischer, an attorney for the government oversight group, whose new filing supplements a complaint filed four years ago. “Cambridge Analytica not only misused people’s personal data, but it was a conduit for the wealthy family that owned it to unlawfully support the Trump campaign in 2016.”

  3. PJ Lehrer from NYU, January 28, 2021 at 2:53 p.m.

    That would be the Mercer family.

Next story loading loading..