The Science of Personalized Matching

August 31, 2022      Di Domenico, Director of Personalized Matching and Donor Experience, DonorVoice

At DonorVoice we use the term personalized matching to describe the process of creating persuasive appeals that align with the phychological characteristics of the recipient donors.

If you’re a regular reader of the Agitator, you already know that personalized matching works. Personalized matching applies the old adage “know your audience” to large-volume marketing.

But you may be wondering: How can you possibly know the psychological characteristics of donors if you’ve never directly measured their personality traits, identities, beliefs, or attitudes?

Let’s start with a few empirical examples from published research.

Researchers at the University of Texas (Gosling, Ko, Mannarelli, and Morris, 2002) tested whether people could accurately judge the personality traits of strangers by simply observing their workplace offices.

Test subjects (“naïve raters”) visited the offices when the occupants weren’t present. The only information the raters had to make their inferences was the physical environments inhabited by the occupants. After the office visit, the personality ratings made by the naïve raters were compared with the personality trait self- and peer-reports of the occupants. The results? The test subjects not only agreed with each other about the personality traits of the occupants, their “naïve ratings” also converged with the self- and peer-reports of the occupants.

How was this possible? The test subjects formed their impressions using cues in the environment that were linked to the personality traits of the occupants. For example, when judging the trait “Conscientiousness”, raters relied on physical cues like how neat versus messy the offices were. And guess what? People who score high on Conscientiousness tend to have neater offices.

When I was still a PhD student at the University of Toronto, I was curious about this phenomenon. Actually, “skeptical” may be a more accurate description.

Is it really possible to accurately infer the personality traits of complete strangers?

I decided to conduct my own study. I invited one group of students into the lab, where they completed a personality assessment and had their photographs taken. Then, I asked another group of students to rate the previous participants’ personality traits. (And yes, in case you’re wondering, I made sure that the students assigned to different groups didn’t know each other.)

To my surprise, I found that the naïve student ratings correlated with the student self-ratings on Neuroticism, Extraversion, and Conscientiousness.

But how? Follow-up analyses revealed that extraverted students were more likely to flash spontaneous smiles when having their pictures taken, conscientious students stood up with their backs straight, and neurotic students looked a little sad. Raters intuitively used these cues to validly infer the personality traits of the photographed students. What’s more, the naïve ratings of Conscientiousness predicted the grade point averages of the photographed students.

There are now many dozens of studies showing that psychological traits can be accurately inferred from all kinds of different cues. Gladstone and colleagues (2019) found that traits can be inferred from objective records of people’s spending habits. Azucar, Marengo, and Settanni (2018) reported meta-analyses of 28 individual studies showing that people’s Big Five personality traits can be accurately inferred from their digital footprints in various social media like Facebook and Twitter.

And why exactly should you care about all this?

Because you can increase donor engagement and response rates by tailoring message to your donors’ personality traits.

Experienced fundraisers are well-versed in third-party data appends. The fundraising industry uses subsets of large commercially available data sets with behavioral and demographic data.

Over the years, DonorVoice amassed a very large database of personality ratings to which a very long list of third-party variables were appended. We developed statistical models that use particular constellations of third-party variables to validly infer the traits of strangers.

Some variables in our models are pretty straightforward: People high in Openness to Experience (creative, artistic types) tend to frequent art-house movies. No surprise there. People high in Agreeableness tend to support the Democratic party. You probably already knew that too. But other variables? Not so obvious. And while a single variable is insufficient to render an accurate judgment on something as complicated as a person’s personality, a combination of, say, 30 variables that are assigned unique weights may surprise you… I know it continues to surprise me!

Here’s how it works.

You give us your donor file. After a quick append, we’ll run it through our model and give your donor file back to you with your donors’ personality traits scored. If you like, we can score identities too. Those pesky environmentalists aren’t always easy to find. But we know where and, more importantly, how to look.

And to make this information useful to you, we can give you strategic guidance on how to systematically develop appeals that are resonant with the psychological traits of your donors. Appeals that are more persuasive because they speak to donors on a deeper level.

We can collaborate with your agency, or if you’d prefer us to do the creative work, nothing would make us more delighted. We’re always happy to write copy and choose effective imagery for your donors.

We hope you’ll explore the enormous potential of personalized matching to increase donor engagement and retention.  Feel free to drop me note with any questions or experiences of your own.  sdidomenico@thedonorvoice.com

Stefano

P.S.  If you want to dig into some of the research buttressing this work here are some references you’ll find helpful.

Azucar, D., Marengo, D., & Settanni, M. (2018). Predicting the Big 5 personality traits from digital footprints on social media: A meta-analysis. Personality and Individual Differences, 124, 150-159.

Di Domenico, S.I., Quitasol, M.N. & Fournier, M.A. (2015). Ratings of conscientiousness from physical appearance predict undergraduate academic performance. Journal of Nonverbal Behavior, 39, 339–353.

Gladstone JJ, Matz SC, Lemaire A. Can psychological traits be inferred from spending? Evidence from transaction data. (2019). Psychological Science, 30, 1087-1096.

Gosling, S. D., Ko, S. J., Mannarelli, T., & Morris, M. E. (2002). A room with a cue: Personality judgments based on offices and bedrooms. Journal of Personality and Social Psychology, 82, 379–398.

2 responses to “The Science of Personalized Matching”

  1. Tom Ahern says:

    Sort of sold ~ Need (minimum) good addresses for Indiana, re: GOTV in 2022 & then 2024 ~ Can you storm those specific ramparts?

  2. Stefano says:

    Hi Tom, Thank you for the reply. Absolutely. To continue with your metaphor, I’d say personalized matching makes the ramparts porous to your message. Did you have any specific questions about personalized matching?