Understanding regular giving upgrades is essential to designing an effective donor engagement plan. However, due to the perceived risk of alienating regular givers by over-engaging and the difficulty in identifying when a donor is likely to upgrade, some charities only run upgrade programs in a sporadic or ad hoc manner.
In Dataro’s view, this approach can result not only in missed upgrade opportunities, but also poorer donor relationships.
The importance of data collection
Timing is critical in a successful upgrade program. And the key to good timing is having strong data collection practices. Organisations that do not make full use of their data run a much greater risk of contacting the wrong donors for upgrades, or contacting the right donors at the wrong time.
Charities with strong data collection practices, on the other hand, know how engaged an individual is, the stage of their donor journey, as well as their individual characteristics. Dataro has found that with this information, machine learning tools can predict with a high level of precision which donors are likely to upgrade. This allows organisations to design rolling upgrade programs resulting in much better returns and a more engaged regular giving donorbase.
How accurate are upgrade predictions?
Of course, the accuracy of machine learning predictions will depend on the data available. However, Dataro typically sees a very strong correlation between donors with the highest upgrade scores and actual upgraders. For example, the chart below shows upgrade predictions based on Dataro’s prediction bands, with a score of 10 being the most likely to upgrade according to our modelling. With this information, our partners can easily identify the best donors to contact for upgrades at any point in time.
Generating upgrade predictions
The factors relevant to identifying likely upgraders varies amongst charities and in our modelling Dataro takes hundreds of different factors into account. For charities just starting their journey to better upgrades, however, some factors that may be important are set out below. These graphs are generated using sample data only, but serve to illustrate what a charity may see when the examine their upgrade program to identify some of the important factors.
Age is often, although not always an important factor in identifying likely regular giver upgraders. In this example, more than 7% of younger donors upgraded their gift at some point, compared to only 3% for other age bands.
Intuition may tell fundraisers that donors with lower gift amounts are more likely to upgrade, but this is not always the case. Charities should look carefully at average gift amounts to help determine the ‘sweet spot’ for upgrade engagements. In this example, moderate gift sizes perform better that low or high gift sizes.
Most fundraisers will be familiar with ‘frequency’ as an important factor in determining single giving. Frequency can also be a key element of a successful upgrades, but it is not as simple as looking at who has made the most frequent gifts. In this example, people with 4 gifts in the past 180 days had higher upgrade propensity than those with 5 or more gifts.
Number of Previous Gifts
Often the number of previous gifts will be a strong factor identifying propensity to give, although this will vary between charities. In this example, new donors with fewer gifts were much more likely to upgrade than long-standing donors with more gifts over time.
Number of Previous Upgrades
Should upgrade programs be targeted at people who have never upgraded before, or at people with a history of upgrading. Often, organisations find that first-time upgraders make up the bulk of their successful conversions. But people who have upgraded before might also be willing to upgrade again.
Are newer donors more likely to upgrade than long term givers? The results will vary, but in this example people with tenure of less than 1 year showed a much stronger propensity to upgrade than people with tenure of more than 3 years.
Total sum donated
The total amount an individual has given, both in terms of regular gifts and single gifts, can be an important factor in determining when to contact an individual for an upgrade.
Combining the factors
Understanding the behaviour of upgraders is important, but the reality is that no single factor on its own is capable of identifying when an individual is likely to upgrade. For example, in the above charts people with fewer gifts (i.e new regular givers) were much more likely to upgrade. But if you asked all new regular givers to upgrade you would run a very real risk of alienating new regular givers who feel they are not in the financial position to upgrade or who for some other reason do not wish to be contacted.
And that is where machine learning comes in. Using a machine learning approach like Dataro’s lets an organisation combine all of these factors, and many more including communications and interactions with each individual donor, to generate much more nuanced predictions. The result of this is that the charity can contact only those people who are the most likely to upgrade and therefore the least likely to be alienated by an upgrade call. The result is better donor engagement and a higher return on investment from an upgrade program. This also allows charities to schedule in regular (e.g. quarterly) programs that will always target the best potential upgraders, regardless of where they are at in their donor journey.