Letting Go of Sacred Cows: The Danger of Being Data-Driven

So you want to become more data driven, right? Everyone does. Everyone is sold on the power of analytics to enable smart, data-driven decision making to improve organizational results. I don’t think it’s entirely unfair to say that terms like “big data,” “analytics,” and “data-driven” are so deeply ingrained by now as to approach buzzword status (not a good, thing, of course!).

But are you really ready to be data-driven, to go wherever the data leads you? It’s often not as easy as you might think it should be for people and organizations to allow data to guide their decisions, tactics, and strategy, especially when the data contradict something they just “know” to be true. Aren’t some things simply given? Simply and obviously true a priori? We all hold these kinds of “convictions” about something, and when we encounter data that contradict what we “innately” know we naturally resist and find ways to rationalize the data away.

There was a perception at an independent school for which I worked (in advancement) that the school’s culture and programs were not adequately “girl-friendly,” which was driving increasing attrition among female students, particularly at the key transition point between grades 8 and 9 (it was a K-12 institution). There was considerable speculation, particularly at the Board level, about what needed to be done to fix the problem. Not atypically, there were some loud voices on the board seeking to push various pet programs, which their advocates intuitively “knew” would improve results.

It happened that we were working on an institutional research project for admissions, and we had enough data to take a look at gender and attrition. We needed to quantify the problem of accelerating female attrition before we could look for correlations. What we found was that there was no particular female attrition problem. In fact, the retention rate for female students was slightly higher than that for our male counterparts. What was really happening was that way back at the start of the pipeline, in grades K, 1, and 2, the balance of students was 60% male, and that balance largely held all the way through. Not exactly something that should have been hard to notice or uncover, but the perceptions were established and this was overlooked.

Did this “revelation” put an end to the calls to add a cheerleading program to the upper school (one board member’s solution for making the school a better place for girls)? The title of this piece provides a hint. The certainty about the female attrition problem survived unscathed. We would like to think we are fundamentally rational beings, but it is hard to let go of closely-held beliefs, no matter how information points in another direction.

What does this kind of resistance to what the data are telling us look like in advancement? It takes many forms. Here are a few I’ve seen recently:

  1. Continuing to mail the same number of solicitations annually to a large pool of never-givers, despite models that suggest de-emphasizing a population and re-directing resources to more likely constituents.
  2. Spending significant gift officer time on statistically low likelihood prospects who have extremely high wealth ratings (yes, it’s fine to make a discovery call if the prospect will take an appointment; but at some point it’s time to move on).
  3. Allocating significant personnel time to maintaining some sort of “extremely important” constituent list, despite not knowing whether there is any correlation with giving (and sometimes, even knowing that no such correlation has been found).

Some people have very little difficulty accepting data-driven conclusions that run counter to accepted wisdom; in fact, some people enjoy the power of data to undermine unexamined assumptions. I know I find it quite exciting. But I’m sure if some Sabermetrician showed me advanced statistics indicating that up to this point in his career Novak Djokovic has been a better player than Roger Federer during a similar interval, my first reaction would be something along the lines of “that’s not what I’ve been watching.” But, as a data-driven fundraising practitioner and consultant, I would have to shrug that off quickly and listen to what the data have to say. In the end, our biases can be overcome, but it does take a degree of conscious effort and repetition. If you’ve made a commitment to becoming a more data-driven fundraising organization, you’ve taken the first step. Remind yourself to hold opinions lightly and to be open to the messages in your data.

Brandon Ferris
Senior Director of Strategic Services and Fundraising Counsel
Zuri Group


“Small Data:” The ONE Piece of Analytics Every Fundraiser Should Know

The phrase “big data” is everywhere now. Most of us are so well aware of the amazing things companies like Google and Amazon are doing with the oceans of data they collect about us that we’ve come to expect this level of sophistication to be the norm. Most nonprofit organizations aren’t able to collect the amount or kind of data the giants can, but they can still use analytics very effectively to achieve better fundraising results.

Because big data and everything associated with it has become such a part of popular consciousness, I believe that our industry has come to:

  • Expect too much from our data. We simply do not have the ability the collect the amount or type of data that Google does.
  • And yet, in opposition to the bullet above, overlook some very fundamental and useful data we have at our fingertips.

The latter bullet is the one to which I refer in my title. To find out whether you know the ONE thing I’m talking about, try to answer this simple question right now:

What percentage of the constituents in your database have NEVER made a gift to your organization?

That’s it! Simple enough.

You’re forgiven if you have to run a query to get the answer this time, but I would suggest this should be top-of-mind knowledge from this moment forward. Why? Because knowing this critical piece of information and using it to inform fundraising business decisions is the first step in becoming a data-driven fundraising operation.

Here are four things you can do with that simple piece of information that will save money, improve fundraising results, or simplify your operations:

  1. Reduce the number of times your long-term never givers are solicited each year.
  2. Rank long-time high capacity never givers lower than other high capacity prospects when making assignments. This is especially important when you wealth screen thousands or tens of thousands of records.
  3. Purge the oldest records of never givers from your database if you can (not alumni, for example).
  4. When you do send an occasional mass appeal to the oldest never givers, craft a message that emphasizes participation over high ask ladders.

You’ve probably already come up with more ideas. These may sound like obvious decisions to make, and yet I still commonly see a proclivity for “shotgun” tactics aimed at everyone in the database. If you cannot start with simple tactics such as these, it will be that much harder to achieve higher levels of sophistication and take full advantage of all the power your data has to offer.

Analytics-based fundraising, data-driven fundraising, call it whatever you like, is fundamentally about prioritizing your resources and efforts – targeting your dollars, time, and people toward the prospects most likely to make the biggest impact. Start with small data and build your way toward becoming a data-driven fundraising team.