Donald Trump, as a candidate for President last year, expressed great skepticism about the use of analytics
in an election campaign. Hillary Clinton made a big deal about her campaign’s use of analytics. Before that, President Obama’s campaigns received great credit for their analytics.
If you compare these experiences, you can begin to understand what separates good from bad in campaign analytics.
Let’s start with the Clinton campaign, whose use of analytics was breathlessly reported, including this Politico story about “Hillary’s Nerd Squad” eighteen months before the election.
However, a newly released book, titled Shattered, provides a kind of autopsy of the campaign and its major weaknesses. A CBS News review of the book highlighted this
weakness in particular:
“Campaign manager Robby Mook put a lot of faith in the campaign’s computer algorithm, Ada, which was supposed to give them a leg up in turning out likely voters. But the Clinton campaign’s use of the highly complex algorithm focused on ensuring voter turnout, rather than attracting voters from across party lines.
“According to the book, Mook was insistent that the software would be revered as the campaign’s secret weapon once Clinton won the White House. With his commitment to Ada and the provided data analytics, Mook often butted heads with Democratic Party officials, who were concerned about the lack of attention in persuading undecided voters in Clinton’s favor. Those Democratic officials, as it turned out, had a point.”
Of course, this had become part of the conventional wisdom since the day after the election. For example, on November 9, 2016, the Washington Post had a story “Clinton’s data-driven campaign relied heavily on an algorithm named Ada. What didn’t she see?”:
“Ada is a complex computer algorithm that the campaign was prepared to publicly unveil after the election as its invisible guiding hand … the algorithm was said to play a role in virtually every strategic decision Clinton aides made, including where and when to deploy the candidate and her battalion of surrogates and where to air television ads … The campaign’s deployment of other resources — including county-level campaign offices and the staging of high-profile concerts with stars like Jay Z and Beyoncé — was largely dependent on Ada’s work, as well.”
But the story had another point about Ada:
“Like the candidate herself, she had a penchant for secrecy and a private server … the particulars of Ada’s work were kept under tight wraps, according to aides. The algorithm operated on a separate computer server than the rest of the Clinton operation as a security precaution, and only a few senior aides were able to access it.”
While the algorithm clearly wasn’t the only or perhaps even the most important reason for the failure of the campaign, that last piece illustrates why the Clinton use of analytics wasn’t more successful. It had in common with many other failed analytics initiatives an atmosphere of secretiveness and arrogance – “we’re the smartest guys around here” so let us do our thing.
The successful uses of analytics in campaigns or elsewhere try to use (and then test) the best insights of the people with long experience in a field. They will even help the analyst look at the right questions –
in the case of the Clinton campaign, converting undecided voters
The best analytics efforts are a two-way conversation that helps the “experts” to understand better which of their beliefs are still correct and helps the analytics staff to understand where they should be looking for predictive factors.
Again, analytics wasn’t the only factor that led to President Obama’s winning elections in 2008 and 2012, but the Obama campaign’s use of analytics felt different than Clinton’s. One article went “Inside the Obama Campaign’s Big Data Analytics Culture” and described “an archetypical story of an analytics-driven organization that aligned people, business processes and technologies around a clear mission” instead of focusing on the secret sauce and a top-down, often strife-filled, environment.
InfoWorld’s story about the 2012 campaign described a widely dispersed use of analytics –
“Of the 100 analytics staffers, 50 worked in a dedicated analytics department, 20 analysts were spread throughout the campaign’s various headquarters, and another 30 were in the field interpreting the data.” So, there was plenty of opportunity for analytics staffers to learn from others in the campaign.
And the organizational culture was molded to make this successful as well –
“barriers between disparate data sets – as well as between analysts – were lowered, so everyone could work together effectively. In a nutshell, the campaign sought a friction-free analytic environment.”
Obama’s successful use of analytics was a wake-up call to many politicians, Hillary Clinton included. But did they learn all the lessons of his success? Apparently not.
Coming back to the 2016 election, there is then the Trump campaign. Despite the candidate’s statements, his campaign also used analytics, employing Cambridge Analytica, the British firm that helped the Brexit forces to win in the UK. Thus, 2016 wasn’t as much of a test of analytics vs. no analytics as has sometimes been reported.
But, if an article, “The great British Brexit robbery: how our democracy was hijacked”, published two weeks ago in the British newspaper, the Guardian, is even close to the mark, there is a different question about the good and bad uses of analytics in both the Trump and Brexit campaigns. In part scary and perhaps in others too jaundiced, this story raises questions for the future – as analytic tools get better, will the people using those tools realize they face not only technical challenges.
The good and bad use of analytics will not just be a question as to whether the results are being executed well or poorly – whether the necessary changes and learning among all members of an organization take place. But it will also be a question whether analytics tools are being used in ways that are good or bad in an ethical sense.
© 2017 Norman Jacknis, All Rights Reserved. @NormanJacknis