Testing Your Donor Identities
Previously on donor identity:
- It’s good to segment by identity
- But they must be the right identity/ies
- There must be meaningful differences among different identities
- And you must be able to get value by messaging them differently
So, how do you know if you get value by messaging differently? You must test.
Sorry. I wish we could skip right to the knowing part too. But my magic eight-ball is stuck on “Outlook Hazy” and my experiments with chicken entrails yielded messy data.
I want to give you some ways to do this yourself, but will also alert you an alternative provided by DonorVoice, the company for which I work. This is unavoidable since we provide solutions that will discover identities and their values. But if you will be sullied by solutions that require filthy lucre, stop reading when this says “Pre-Test Tool.”
The central challenge of testing identities is that they are often self-reported. Thus, you must collect them from your donors before you know if they are valuable. Ugh. And you must collect them in quantity for any sort of significance. Double ugh.
So if you want to start testing identities on the cheap, start with pre-existing identity data. From the not-at-all-comprehensive list of ideas from yesterday, there are some you may already know like:
- Who has received services from you
- Who has volunteered with you
- Interactions with the organization like pet adoption
- Content consumption patterns
- If they’ve been to your museum/opera/library/monument/park
To that, you can add proxies for identity:
- People who downloaded your brochure “So you have spoy in your fleep” are likely spoy-fleep sufferers and can be coded as such.
- Repeated action alert senders are likely advocate identities
- While more difficult, you may be able to determine preferences like cat versus dog by click pattern in emails
None are as good as self-reported data, but they are OK as long as you bake in some skepticism with your results.
So now, we test? You sure can. Just split your identity files in half, sending half identity version A and half identity version B. For example, if you have cat and dog people, send half of the cat people cat messaging and half dog messaging. Same for the dog people. If cat people like cat messaging and dog people like dog messaging, you are in. (If both prefer one messaging or the other, maybe you have a new control!)
But let’s say you have a largely offline donor file – you won’t be able to test without incurring significant cost.
Instead, you can backtest. That is, take your identities, match them to the last few years of data and compare lifetime values of the donors by identity (ideally controlling for year of acquisition – if you just started acquiring advocates, you don’t want to compare them to 20+-year donors). This will tell you if you have an identity that is worth significantly more than another.
But some identities, like the cat versus dog example above, are likely at the same value. Rather, the identity value lies in matching the message to the donor.
For those, you can start backwards – with the communications. Simply go through all your communications to a select set of donors and code them as to what identities you’d think these communications appeal to.
Ha ha ha ha ha ha! Yes, I was kidding about “simply.”
This is a painful, laborious process. I’ve done it. But capturing key information about the communications related to identity (e.g., is there a petition involved? is the story about someone helped or about the volunteer doing the helping?) will allow you look at people’s previous activity.
This is messy data, because we don’t always know if the communication is the direct reason for the gift or an incidental one. But I once did this with people who had donated 10 or more times in five years. Of those donors, about seven percent of them donated to 80% or more mail pieces with petitions attached (which accounted for about a third of pieces). The odds of this happening by chance were very small; these donors became the starting point of an advocate identity.
This process works for information you have. But what if you don’t have data for an identity you want to test? Here I’ll advocate for something I rarely advocate for: telemarketing. Using a telemarketing script, you can ask an identity question up front (ideally, only one) then branch off customized scripts based on their answer. You’ll get your answer as to who has what identity and whether it matters in the same phone call.
But these are relatively blunt instruments. That’s why I advocate for a Pre-Test Tool or Commitment Study from DonorVoice.
A Pre-Test Tool is A/B testing on steroids. They break your messaging down into concrete segments (e.g. voice, who you help, what your donation does, identity statement, type of gift – whatever is most important for you to learn about), then create five different versions of each segment. Next, donors receive a survey where they go through about eight different questions that work something like your optometrist’s eye test: do you prefer A or B? Since A and B are both made of different versions of each attribute, you can quickly understand not only what version people prefer more, but what attributes matter most.
Additionally, you can place questions about identity at the end of the survey and, from that, determine if different identities prefer different messaging. All without peeling one stamp. You can get more information on this sophisticated and highly effective process from The Agitator Toolbox here.
Commitment studies go into depth with your donors about all their touchpoints with you: brand positioning, donor relations, fundraising asks, cultivation, channels, messaging, etc. It also asks about their identities and their commitment to your organization. By marrying their answers with your transactional data, you can see what touchpoints cause commitment and those that don’t. This allows you to scale up touchpoints that matter, fix the broken ones, and drop those that do not increase retention and donor value.
Moreover, by asking identity questions, DonorVoice can tell you which of your identities are different in value and commitment, plus what makes your donors tick. It’s an essential step in creating different journeys for different types of donor. You can learn more here.
And of course, if you have questions, feel free to email me at email@example.com.
This article was posted in: Breaking Out of the Status Quo, Communications, Donor Identity, Donor retention / loyalty / commitment, DonorTrends / DonorVoice, Fundraising analytics / data, Research, Segmentation, Uncategorized.
You can follow any responses to this entry through the RSS 2.0 feed.
You can leave a response, or trackback from your own site.
Search posts by category
Are you looking to recruit senior staff for your organisation or are you looking for a new professional challenge? Then check out Global Charity Jobs TODAY
Position: Director of Business Development
Location: Washington, DC (flexible)
Closing date: 4 March 2018
More info and how to apply: http://bit.ly/2Dz6vX2
Position: Head, Program Performance
Location: Montpellier, France
Closing date: 11 March 2018
More info and how to apply: http://bit.ly/2Dz6vX2