Grandma Craver always advised:  “Never wake a sleeping snake to kill it.”

And generally, I try to follow her advice.

But, over the weekend The New York Times ran a blockbuster of a piece How Trump Consultants Exploited the Facebook Data of Millions.

Behind the story is the tale of a voter profiling firm named Cambridge Analytica that claimed to have developed a “so-called psychographic modeling techniques” – that would be the secret sauce to be slathered on voters to bring out the Trump vote in 2016 U.S. presidential election.

My point in citing this article has nothing to do with the Trump Campaign, Cambridge Analytica or Facebook. (Read the story and judge for yourself).

Rather it serves as a poignant reminder of the eternal quest to find a magic segmentation elixir – for consumer marketing, political campaigning, and yes, fundraising.

Thus,  we’ve decided to stir the “sleeping snake” — the issue of segmentation –again for this week’s series of posts.

Background in Brief

In early February we ran a series on segmentation (here, here, here, and here.

A month later,  in Are Your Segmentation Skills a Joke? , I returned to the subject pointing out that segmentation by RFM alone is not good enough, but focused most of my concern (ire) on the use of stereotypes or “personas” that base segmentation on similar demographic or transactional behavior. I noted, “Resorting to segmentation based on demographics or personas can create fatally flawed stereotypes that don’t get at your donors’ reason for supporting you.

Déjà Vu All Over Again

While reading the Times piece I recalled attempts at voter and donor profiling back in the ‘70s and ‘80s when the methodology of “laddering” was all the rage.

This “new”, “exciting”, “never-before” technique was conceived by clinical psychologists in the 60’s and used by practitioners to try and understand the deeper, values consumers attach to tangible products and features.

In the realm of voter profiling the firm Wirthlin Worldwide pedaled this heavily to Republicans on Capitol Hill with claims of knowing which voter values (i.e. their deep, motivations and needs and aspirations) were attached to core issues – e.g. lower taxes, more defense spending.

(Oddly and invariably every issue seemed to be linked to ‘family values’ and ‘personal responsibility’.   If every issue is driven by these two values and the party/candidates are already beating these two drums steadily, what value is this research?  Well, as it turned out, about $7500 per brief for Wirthlin Worldwide.)

TODAY – 40+ years later– the snake (or at least the snake oil ) has awaken.  It seems to be making its way back in vogue in consumer marketing, voter profiling and yes, fundraising.

And so I’m adding  “Laddering” ( also called “means-end chain theory”) to my list of iffy, dubious, caveat-emptor segmentation techniques offered to nonprofit fundraisers.

Here’s how laddering is (supposed) to work as a method for understanding not just donor motivation but also  for donor segmentation.

First the Positive

This methodology, while nearly value-less, for donor segmentation does make for kick-ass presentation charts.  So look for it soon at a conference near you soon,  or appearing breathlessly in some consultant’s RFP presentation.

Now the Negative.

Using the chart above here’s how the methodology works.  Look at the group of consumers labeled and focused on “Cleanliness” (basic, tangible attribute) These folks should be marketed to to feel comfortable and help promote values of personal happiness (purple “chain”).

Of course, the “Cleanliness” folks should be marketed to differently than the “Equipment variety/variety of exercise/healthy wellbeing segment.

These are lovely diagrams and even intuitively interesting.  Unfortunately, lovely and intuitive are not the same as valid and legitimate.  There are three problems with this approach — any one of which spells death for having this work in-market and lifting performance.

Here, according to researchers, is why this “laddering” or “Means-End” Chain Method isn’t reliable:

• Although this methodology and approach does elicit parts of the mental, cognitive structure of memory and associations, what it uncovers is highly dependent on artifacts of the research process (e.g. where, when, who does it).  In short, it is not reliable.

• Setting aside the issue of “reliability”, it also isn’t valid if the aim is to uncover what motivates behavior. Turns out what this uncovers is a part of memory but not a part of motivation – i.e. it doesn’t work to apply this and expect change in behavior.

• It turns out that most folks can ‘map’ onto most or these chains.  There goes segmentation value – most people fall across most ‘chains’, not just one. And of course, as The Agitator has pointed out before, if your segment is made up of different donors who want different things, it is not a segment.

So much for the “latest” fad in Fundraising Fantasyland. Flashy charts.  Unreliable results.

So, what’s a Fundraiser to Do?

For the rest of this week we’ll deal with what’s real and what’s not when it comes to issues of segmentation and how to best approach various segments.  Tomorrow Kevin will outline how to dramatically reduce the risk of determining what approaches work best for your organization.  Then Nick will pick up from there with posts on how to use donor segmentation identities for Facebook, for controlling costs of acquisition and why donors choose you.

MEANWHILE….  To separate theory that has been validated from “interesting ideas” (snake oil and fantasy) I do recommend the free webinar about identity-based segmentation: “WHY SEGMENTATION AND LACK OF REAL DONOR JOURNEYS COST YOU DONATIONS” to be held on April 4^th and hosted by my colleagues Nick and Kevin.    You can sign up free right here

Have a good week.

Roger