Accurate budgeting and forecasting is becoming a must do for nonprofits, especially as their individual giving programs grow say Andy Tidy and Clarke Vincent…
Some of the most commonly asked budgeting questions in fundraising are, “If I do no acquisition, where will my income and donor base be in one, two or five years?” or “What do I need to invest in order to reach an income or donor volume of ‘x’ amount?”
Alongside these questions, accurate budgeting and forecasting is becoming more important as charities build large individual giving programs. Why? Because if a forecast is out by a relatively small percentage it can have a big impact on service delivery.
The key to accurate predictions is knowing the probability that a donor will or won’t do something. It’s nearly impossible to accurately predict what a single donor will do, but across a large enough group, it is possible to predict with relative accuracy.
Regular giving forecasting
For existing regular giving donors this is fairly straightforward. The main thing we need to know is the probability that the donor will not stop their direct debit (either intentionally or by failed payment). There is also a need to predict the propensity to upgrade and by how much, as well as the likelihood of skipping payments, but these have a relatively small influence compared to overall retention.
For regular giving programs with an acquisition strategy, where early donor attrition can often be as high as 50% over the first year of a donor giving, we typically also want to know expected recruitment volumes (including non-starter rates and first payment phasing), along with their subsequent retention, upgrade propensity and decline probability. These variables are essential for us to confidently manage the large budgets invested in regular giving acquisition and donor development.
It is imperative therefore to identify and understand the factors/variables that influence or correlate with retention of both new and existing donors. It is then possible to split the donors into segments defined by those variables and go on to forecast expected income based on the retention characteristics of each segment.
Typically, depending on the size of the file, existing donors will be split into 50 to 100 segments, with acquisition channel (for example, face-to-face, online lead conversion, DRTV etc) and tenure (months of giving) being the most significant. For new donors a smaller number of segments is used, with channel and age being the main drivers.
Each segment’s likely future behaviour can then be calculated by extrapolating the behaviours of previously observed similar segments. Applying an accurate prediction can also rely on an ability (somewhat an art) to forecast changes in the behaviour of a segment caused by anticipated future changes in external influencing factors such as market dynamics, charity fundraising and strategy development.
The segments are then ‘rolled’ back up into a whole group and their overall predicted behaviour can be used for accurate forecasting. For easy visual trend insight, they are often charted too. The two regular giving forecast charts show a no acquisition scenario where a charity is expected to see income shrink and an active acquisition scenario where the charity is recruiting enough donors to grow.
Regular giving forecast: five year (No acquisition)
Regular giving forecast: five year (Active acquisition)
Cash donor forecasting
Forecasting cash donor (single, one-time gift to an appeal) behaviour and fundraised income is inherently more difficult than for regular givers as cash donors have to make the decision to give each time. And depending on the outcome of each of these decisions, the expectation of their future behaviour changes.
It doesn’t mean a cash income forecast can’t be done, it is just that the forecast will not typically be able to be predicted with as much confidence at a granular level. The principles are the same as for regular giving – we need to identify the factors/variables that predict an appeal response and how much they will give – typically these will be based on recency, frequency and value.
Each time a donor makes a donation it potentially changes these characteristics, so the process becomes an iterative one, therefore each year of the forecast is dependent on what we expected the donors to do the previous year.
Forecast cash direct mail income by year (No acquisition)
Forecast cash direct mail income by year (Active acquisition)
These charts show the difference in forecast income for a direct mail program from doing no acquisition vs acquiring donors consistently over the next five years. It becomes a fairly simple process to manipulate the donor recruitment volumes to achieve steady income or hit growth targets.
An income forecast model can then be enhanced with inputs such as budgets and costings to help model different scenarios for the long-term return of your integrated fundraising program.
Predicting the future can be fun, enlightening and very useful but all forecasts should carry caveats about their limitations and accuracy. The key to improved accuracy is to compare predictions and actuals over time, try to understand the key factors and causes of variations, and make corrections.
Most experienced stakeholders will understand that forecasts should be treated as a useful indication rather than a promise of results. If stakeholders wish to lock-in or guarantee minimum results (for example, to be able to make commitments about future service provision) then it is wise to provide different scenarios that represent different combinations of variables (for example, based on last year’s retention rates, based on retention rates 1% lower than the previous year etc).
What about unknown unknowns?
Some fundraising programs become harder to forecast as they mature, or ‘max’ out, as it becomes harder to find and engage new supporters. While it is possible to more accurately predict the income to be received from retained/existing donors, predictions about the long-term future behaviour of donors yet to be acquired is fraught with more variables and risk. There is, however, a considered approach that can be applied to building longer term models of multi-year donor acquisition programs and I will outline an approach to do this successfully in a future article.
Clarke loves talking about data and has made a career out of doing so for more than fifteen years in the direct marketing industry. He has been our Head of Marketing and Business Development since February 2009. Andy started his career in the UK in 1993 working as an analyst for two large charities. In 2003 Andy began working for Pareto Fundraising where he has been the driving force behind Pareto Fundraising’s international reputation for data-driven fundraising excellence.
This article was first written by Clarke and Andy and was first published in the October/ November edition of F&P Magazine.