Red State, Blue State? Dig Deeper

July 31, 2024      Kevin Schulman, Founder, DonorVoice and DVCanvass

These two profiles seem like stereotypes or memes for urban/rural, or red state/blue state.  This is 3rd party data.  It’s big data, it gives us mass market reach as I can know these sorts of things for most people.  It’s also entirely correlational, not causal.  

Having said that, the data are certainly suggestive of two different people.  You likely have an image conjured up in your head for each and an immediate preference on who you’d rather hang out with.  That’s because we are a certain way and therefore, show up a certain way.

The problem is we mistakenly think we should market to the profile, that it tells us something useful about why the person would donate to you.  It’s describing people, not explaining their why to your organization. But because some of those descriptors seem on point – e.g., green products – we think it’s giving us more insight than it is.  If you marketed to the profile you’d waste time and money and experience first hand the “correlation not causation” line going from throw-away snark to the reason for your poor outcome.

Enter “small data”, where we have insights, theory and cause meets effect.   In this world you’d learn that the starting point for this profile wasn’t the profiling, it was segmenting donors by personality trait.  That person on the left?  High in Openness.  On the right?  High in Conscientiousness.   You could be a very progressive group and have success with the red person, same for a conservative cause and the blue person…but not until you know what’s causing both profiles.

The Open tailoring that makes your fundraising work better would have nothing in the copy or image choice having to do with the profile.  Ditto for Conscientious.  The profile exists because people with differing personalities make different consumer and lifestyle choices.  It’s irrelevant except that we use the profile data as proxy to find those high in a given trait.

The reason all this works and is doable is because we started with theory>small data>big data.  The order matters.  Mess that up and the profile is nothing more than an alluring waste of time and money.

Kevin