Acquisition: Predictive Analytics And The Presidency

November 6, 2012      Admin

History’s most expensive acquisition effort – the $2 billion campaign for the US Presidency – comes to an end today.

Hopefully by late tonight, or at least in the next few days, we’ll know which voter acquisition campaign was the most successful.

Apart from our personal interests in today’s outcomes, we fundraisers should pay particular attention to the 2012 Presidential Campaign. Why? Because many of the approaches used by the campaigns in search of votes — and the media, in search of accurate predictions on the outcome — bear watching in terms of our own acquisition efforts.

Of course the result of the ‘too close to call’ Presidential race has been known for months. Of the 270 Electoral College votes needed to win, as of today (I’m writing this at day’s end on Monday, 5 Nov) President Obama has an 86% chance of receiving 307.2 electoral votes to Governor Romney 230.8, thus sending the President back to the White House for a second term.

This forecast is the product of a predictive model created from an analysis of a range of swing state polls by Nate Silver for his 538 Blog at the New York Times. The math-based work of this young statistician is driving the on-air cable channel pundits and the news magazine and newspaper columnists and prognosticators bonkers.

Despite the fact that Silver called 49 out of 50 states correctly in the 2008 and despite the fact that he has over and over detailed what’s in this model, he is being attacked by the pundits who prize their own ‘gut feelings’ that the race is a ‘toss-up’. His crime? The use of math.

This knee-jerk and negative reaction to the value of statistical modeling on the part of the pundits isn’t limited to politics. The same holds in fundraising, where too many consultants, brokers and fundraisers themselves are far too quick to dismiss math and rely on their own gut feelings and experience.

When I ask why, they respond with a variety of dismissive excuses: “We tried it and it didn’t work.” … ”I know best what my organization/client needs.” … ”I’m not turning my years of experience over to some statistical geek.”

Alas, when it comes to tapping into the power of modeling and predictive analytics to improve our acquisition results, all this ‘business-as-usual’ chatter adds up to a big lump of stupidity.

In the fast-changing, fast-improving world of statistical modeling, the fact is that we rely on them for many decisions every day, including medicine, business, the weather, and pretty much any complex system in which there’s an element of uncertainty as to the outcome. (These are the same methods that enabled scientists to forecast many days in advance where and when Hurricane Sandy was about to hit the US.)

In business, predictive analytics – and the term encompasses a wide variety of statistical techniques from modeling to data mining and machine learning — exploit patterns found in historical and transactional data to identify risks and opportunities.

And so, sooner or later, fundraisers too will come to learn that carefully executed modeling has developed to the point where it represents a viable alternative or addition to using the same-old-same-old response lists for finding new donors.

We’ll deal with the current predictive analytic options available to nonprofits in future posts. We hope that any of you who’ve explored this fast-emerging alternative will share your experiences with The Agitator.

Meanwhile, Agitators in the US who have not yet voted, Get Ye to the polls.

Roger

P.S. For a round up of how predictive analytics is being used by both the Democratic and Republican parties in this year’s campaigns, take a moment to read the terrific cover story The Direct Marketing Election in yesterday’s Target Marketing.

P.P.S.  HURRICANE SANDY UPDATE.  A number of readers have asked how to find out more about US Postal Service Mail Delivery.

You’ll find answers here on the Production Solutions website, where they’re updating the page daily with Sandy-related closures and impacts.

It includes links to USPS spreadsheets of impacted zip codes and opinions from Production Solutions on which ZIPS they would suppress and which they would mail. The updates are done toward the end of each day.

3 responses to “Acquisition: Predictive Analytics And The Presidency”

  1. Robert SanGeorge says:

    Roger,
    Great post! Look forward to hearing about predictive analytics being applied to cause campaigning and fundraising.

    You very accurately describe the tension between quants and pundits, even within their own news organizations. It’s come to a head in this campaign because pundits at high profile places like Wash Post, Politico, NPR stubbornly stick to the “too close to call” lazy reporting mantra. I suspect the pundits see their jobs — and to some degree, their profession — being in danger, since most of them don’t know regression analysis from a hole in the ground.

  2. Such a great post! I recently had a heated argument with a communications professional who insisted her gut feeling was more accurate than any statistical analysis being generated by their direct mail team. I love a good gut feeling as much as the next person, but when there are so many delicious statistics to dive into, it seems foolish to rely entirely on feelings for important fundraising and communications decisions.

  3. Jenny Floria says:

    This is one of the reasons why I love direct mail fundraising — the numbers don’t lie! Or, you can perhaps refer to one of my favorite quotes from Mark Twain: “There are three kinds of lies: lies, damned lies and statistics.”