Predicting the Future—It’s in Your Future

I took a graduate statistics course last semester called Advanced Multivariate Methods in Psychology. As the professor was introducing the class, he explained that after taking the course, we would know more than the senior professors in the department—because when they were trained, these multivariate techniques weren’t available. This was as recently as 15 years ago. Now, the majority of research that is published in psychology features these new statistical approaches.

Something similar has happened in the last 15 years in the nonprofit world.

If you are in the peer-to-peer space, you’ve likely been benchmarked to death. We, at Turnkey, probably helped. But benchmarking provides a two-dimensional picture of the world—it isn’t enough. Decision-making based solely on benchmarking puts you behind other segments of nonprofit fundraising.

If you manage a direct response program, you probably have used predictive analytics to avoid (as much as possible) sending an expensive mail piece to anyone who is unlikely to respond. If you are in major giving, you’re using wealth screenings inside predictive analytics tools to allocate expensive human time on the best possible prospects. If you are in peer-to-peer, you’re benchmarking. And that is hurting you.

Benchmarking is an examination of the past. Benchmarking is descriptive and diagnostic, and it has value. But it is not predictive. If “benchmarking” were moving a couch, “benchmarking” would move the couch 20 times and evaluate the effect of each placement on the satisfaction of the home’s occupant. If “predictive analytics” were moving a couch, “predictive analytics” would find twenty people who had already moved their couches, interview them and apply those learnings to where to put the couch. It’s highly likely the couch would land in the right place the first time if “predictive analytics” were moving it.

Benchmarking can tell us what happened and maybe why it happened. Predictive analytics can tell us what is probably going to happen next. It’s ability to predict what is going to happen is dependent on the quality and amount of the data used.

What if you were using predictive analytics in peer-to-peer? What would you be studying? What would you be looking to impact? Two things:

• The people.
• The environment.

If you are studying people in terms of their wealth, age, gender, ethnicity, prior giving, etc., you could predict who, on a new list, is likely to register, activate to fundraise, raise a lot, become a team captain and so forth.  And that is very helpful.

If you are studying the behavior of people within an event construct, with enough data, you could predict their fundraising performance based on those elements of your event that you are able to measure, to turn into quantifiable data.

You could, for example, model what will happen to fundraising if you switched from an afternoon event to an overnight event. You could predict what would happen if you dropped your registration fee. You could predict what would happen if email delivery begins to fail. You could predict what will happen if you change venues. Wouldn’t all that make budgeting and planning a lot easier, and a lot less risky?

This is where analytics is going. Benchmarking, while less expensive to produce, simply does not answer our questions sufficiently. Peer-to-peer fundraising must begin using predictive analytics to plan, design and change. If not, we’ll be moving a lot of couches.