Creatalytics by Matthew Burgess

I saw an eHarmony commercial on TV, where they seemed to describe a method of matching singles based on obvious similarities discovered in some sort of interest inventory questionnaire.

I thought to myself, admittedly a bit smugly, “how clueless.” Compatibility is way more complicated than that, and either they don’t get that and their “statistical methods” are shoddy, or this marketing is just dishonest.

This isn’t the exact commercial that I saw, but close enough. Take a look:

eHarmony Ad (will launch new window: they don’t allow embedding)

You get the idea: their marketing approach is to describe a higher level of sophistication on matching. You can almost see the white lab coats, clipboards and pocket protectors in the other room, behind the cool West Elm-decorated walls.

“We prescreen each and every member for you, to determine your matches based on compatibility. So when you get to that first date, you know you’ll have something in common. eHarmony does the matching for you, based on who you are, at the deepest level.”

My personal belief is this: true compatibility rarely lines up with obvious similarities that can be addressed in a questionnaire. And I was skeptical that they were doing anything that diligently addressed the true complexities of a magical match.

It sounded like they were taking an interest inventory and match up people with similar interests: artists with artists, sports fans with sports fans, Christians with Christians. Keep the crazy fun-loving-youthful-spirit types away from the serious, sober folks. Hide the Ph.D.s from the GED’s.

In my opinion, and from my experiences, this was an irresponsible course for a couple reasons:

1. commonalities in interests, background, hobbies, etc. often have very little relation to true compatibility, and
2. people are notoriously inaccurate when describing themselves and what they’re looking for.  It’s sort of an unintentional dishonesty that leads us down the path toward describing how we want like to be perceived.

So, it seemed to me like they are doing people a disservice with false and irresponsible analytics, by applying conventional wisdom and intuition to a system to match people who were as similar as possible. Idiots!

But, then I read Super Crunchers, which explained why I was dead wrong (not about what makes people compatible, but about the eHarmony folks being idiots.)

The book talks about how the regression—a statistical procedure that takes historical data and determines how various causal factors influence a single variable of interest—helps us make decisions by improving predictions.

Author Ian Ayres’ first demonstrative case study: eHarmony’s predictive statistical model.

As it turns out, eHarmony’s founder, Neil Clark Warren, surveyed over 5,000 married couples. He developed a predictive model that includes 29 different variables that relate to emotional temperament, social style, cognitive mode, and other factors. He used a regression to predict how these factors influence a variable of particular interest to him: compatibility of a couple.

Super Crunchers (emphasis mine):

“There is a new wave of prediction that utilizes the wisdom of crowds in a way that goes beyond conscious preferences…Unlike traditional dating services that solicit and match people based on their conscious and articulated preferences, eHarmony tries to find out what kind of person you are then matches you with others who the data say are most compatible”

The regression is about working with data and finding the “levers of causation that are hidden to casual and even expert observation,” as Ayres puts it.

So with this approach, eHarmony tosses conventional wisdom or user stated preferences out the window. The data tells them what “works.” On some personality dimensions, the data will say that we should be similar to our match; on others, its best that we’re dissimilar.

The regression drives this kind of data-based decision making by improving predictions. If you’re trying to decide where to take your marketing next, the answer may be right there in your database.

This is one of the tools that allows you to do more with more. As I’ve said, forget “KISS”: embrace complexity by letting data analytics converge with and inform your creative.

Filed under:
Data Driven Marketing by Matthew Burgess