Charity CRMs: The Record-Band Conundrum

As many CRMs used in the non-profit sector have moved to software-as-a-service models, they have attempted to price things in a way that makes the most sense. One of these is using a "record band" - effectively a number of records that incur a particular price.

It is easy to understand why they would go this way. From the nonprofit side, it provides a pretty stable method of determining budget. We have under 10,000 records, so we pay this price. As that number increases, theoretically so does our income.

From the vendor side, as the number increases, so does that processing power required to do things like query your database. Furthermore, storage space increases, which also increases the price for the vendor. For reference, many of these vendors are using things like Azure (Microsoft) or AWS (Amazon) to run their hosting, which charge them based on processing (input/output operations per second) and storage space.

So record bands seem reasonable when you look at it through these lenses.

The Amazing Shrinking Database

One of the externalities from this record-band policy is an aversion to crossing the line. If you look at it from an incremental perspective, adding one record to a 9,000-constituent database in a 10,000-constituent record band is zero. You've already paid for it. But adding the 10,001st constituent costs significantly more.

So charities, constantly at odds with the public's view of overhead, find ways to avoid crossing the threshold when possible.

I can tell you, from numerous conversations, that people are deleting records. En masse. There are many ways that this takes place, and there are many creative solutions (I've even seen *other* record types be used instead, effectively turning the database on its head, because one type is banded and the other is not).

Aversion to Data

As the realm of data science, big data, and the ability to train machines to evaluate datasets and come to conclusions has reached the sector, we seem to have policies that do everything they can to prevent data from going into the system.

All we have is email and a first name? Nope. We have too many of those. We are using another system for a particular type of fundraising method? Can't bring them over and align the databases - too expensive.

We should be wanting as much data in these systems as we can reasonably enter. We want to be able to match records on as many criteria as possible, to gather every person who may possibly want to engage with us, and then analyse using modern statistical tools. Instead, we purge, and avoid. It's perhaps a short-sighted policy, but with that ever-present overhead shadow, it's difficult to avoid.

Pricing the Future

The question is, what will that data be worth? Truth is, we don't always know. In a generation, when we have data going back decades, will we find that having family trees and historical donation data will indicate potential future donors from a familial aspect? Will we discover movement patterns of individuals based on addresses that can indicate when people have landed higher-paying jobs and might have more capacity?

In some cases, those who are purging are keeping the data somewhere, and may be able to tie it back in as machine learning tools make their way into the smallest of charities. But I fear that this policy by the vendors who influence our industry has already had devastating effects, some that we won't fully understand until we learn how donor data truly can be used.

The Crux

In the end, this is a call for my fellow non-profit people to find ways to keep data as much as possible. If you need to archive, archive, but keep and gather whatever you can (within reason), with hopes that the pricing models will change.

This is a call for donors to realize that we work with for-profit vendors (some of which are publicly traded and beholden to shareholders), and that what they do can affect how your dollars are used, and what actions a nonprofit may take to keep costs low.

This is, hopefully, the first of many calls for vendors to rethink their pricing models. Charity data is getting worse, not better, with pricing models like this, and it will be the long term that's affected.