This post is sourced from an article I wrote for the Direct Marketing Association of Washington’s bi-monthly magazine, Marketing AdVents.  I encourage checking out the DMAW membership offer.  They produce a lot of good content with conferences, webinars and this publication.

The only reason to group donors is because you believe you’ll be more financially successful by treating them differently.

Here are five, must-have criteria for a useful and successful segmentation.

• People in the segment are very similar to one another on factors that matter to their giving.
• The people in Segment A are distinct from those in Segment B, C…on factors that matter to their giving.
• The segments are relatively stable over time.
• The segments are identifiable and reachable in the marketplace
• The number of segments is manageable.

Let’s focus on these first two rules and the question begged. What traits or variables do I use to create my segments?

If there were ever a time to measure twice and cut once, this is it. First, consider the types of data you can use. Broadly speaking there is behavioral, demographic and psychological data.

Non-profits use behavior data a lot. It’s usually to decide who’s in or out for a given communication. This isn’t segmentation, it’s selection. Demographics like age are also used but again, mostly for selection, who gets the planned giving communication, as an example.

Demographics rarely explain the “why” of giving and behavior data never does. The former describes “who” and the latter tells us where, when, and how much.

Don’t confuse selection with segmentation. 

This leaves softer, fuzzier, psychological data to crack the code on “why”. I’ll address this in more detail but first, our caveat emptor.

One of the disturbing trends in fundraising is charities being sold a black box answer to the question of what data to use. This is typically packaged as cluster analysis.

The TLDR recommendation: run, don’t walk to the nearest exit if someone proposes using cluster analysis for your direct marketing segmentation.

Most of these cluster analyses are not making choices on what data to use. Instead, they throw dozens of mismatched behavioral, psychological, and demographic variables into a statistical stew and press a button in the point-and-click stats software. This cluster f*$% is hidden from view by trotting out PowerPoint slides with different segments, often called Personas.

These presentations have slide after slide of highly descriptive, cleverly named segments and all manner of rich data on channel, demographics, motivation, lifestyle. It all seems perfectly intuitive and applicable. There are lots of technical and logical reasons why these won’t work, but suffice to say, if your segmentation mostly lives in PowerPoint or was done using cluster analysis then there is a 99.9% chance it’s violating 3, 4 or all 5 of the must-have rules.

What you want is a segmentation that adheres to the five rules and influences every step of your direct marketing strategy and execution.

Psychological data is the starting point. You’ll know you’re on the right track if the data you used to create your segments doesn’t require your charity to exist. This is the old trope about meeting your supporters where they are and “where they are” is not intrinsically caring about your charity nor your beneficiaries.

We should instead rely on well-established academic theory and concepts that have also survived real-world testing among diverse populations.

A couple to consider are Identity and Personality.  Each is a Google search away from a master’s degree worth of content but here’s a primer.

Identity is a personal narrative, a part of who we are.   Think of Identity as what might connect someone to your mission.  For example, there people who identify as Conservationists and because of this, they have potential to care about the work of conservation charities.  Importantly, they don’t need to give to a Conservation charity to live out their Conservationist values and goals…but they could.

Personality is our disposition, our orientation.  There is a Big Five of personality traits and we all have parts of the five but we also skew towards a dominant trait or two.  That skew determines a lot of what we find of interest and value.  It determines what we pay attention to and ignore.   There is specific messaging to match each Personality trait.

You can use Personality as a basis for your segmentation or you can use it to further segment within a given Identity because, returning to our example, not all Conservationists are the same.  The Conservationist who skews toward Openness will be receptive to a story that includes details about habitat restoration. The Conservationist who skews toward Agreeableness just wants a story about providing care to animals.

You can use 3^rd party data to tag individual records in your CRM or acquisition mail merge list as likely belonging to the Openness Conservationist group or the Agreeable Conservationist group.

The messaging and imagery will be very different to match the differing Personality types.

In this scenario, your aim is two controls, one for each segment. The path to growth is traditional, incremental testing with each segment to beat the control but also further segmenting using Identity and Personality. For example, there are Conservationists who are more focused on certain places – beaches, mountains or those focused on flora or fauna.

You don’t have one audience and you almost certainly have more than two. But going from one to two is a big deal and it’ll put you ahead of most of your peers. Take the think time required.

The siren on the cliff of quick and easy—albeit expensive—Persona development is going to leave you disappointed and now you know why. It likely violates many or all of the must-have criteria for good segmentation and it’s mostly or completely detached from your direct marketing execution.

Kevin