Environmental charity

The charity selected Adroit to perform an analysis of donor lifetime value across its donor database. The main objective was to better understand the dynamics underlying LTV — specifically, to identify the drivers that influence it, such as the age of the donor at the time of recruitment, gender, and the value of the initial fee.

At a donor level, LTV is inherently polarised: relatively few donors are truly “average”, even though the average figure remains useful for effectiveness calculations at a portfolio level. The charity needed a way to see past that averaging effect — to understand, cohort by cohort and channel by channel, which factors were pulling LTV up or down, and by how much.

The organisation also wanted the analysis to look forward rather than only backward: not simply a report on how past cohorts had performed, but a robust way of projecting how more recent donor cohorts were likely to behave over their full five-year lifetime, and a clear view of whether sign-up values and retention rates were improving or declining over time.

Adroit proposed an extrapolative approach: using historical behavioural data, survival rates and cross-selling rates to extrapolate future LTV curves from the charities more recent donor cohorts.

The analysis assessed how performance was changing over time and what was driving those changes — for example, whether sign-up values and/or retention rates had shifted.

LTV calculations were built from each donor’s initial giving value, with donors recruited via Face-to-Face used as the core cohort: Gross 5-Year LTV = Total Expected Income within 5 years of 1st Gift ÷ Total Number of Recruits with a financial transaction. To add further value, the charities average cost per acquisition — by year and by channel — was folded into the calculation, turning a Gross LTV view into a Net LTV and ROI view for every acquisition channel.

The Prophesee forecasting engine

Adroit’s Prophesee Income Forecasting platform used past donor observations to project 5-year Gross and Net LTV, running 10 Monte Carlo simulations against the supplied data to produce robust, repeatable results rather than a single point estimate.

Costs folded into the calculation

The charity contributed its average cost per acquisition (CPA) for each year and channel. This let the analysis move beyond Gross LTV to determine ROI by acquisition channel, and to examine in detail how CPA, initial value and retention interact to drive Net LTV.

1. Specification (agreeing requirements and clarifying how specific data features needed to be handled);
2. Data Supply & Load (data loaded into Adroit’s analytics platform);
3. Data Audit (load counts confirmed back t);
4. LTV Analysis (data engineering, modelling and documentation of findings)
5. Feedback Presentation to walk the results back to the charities senior team and agree actionable insights.

Trend and channel analysis

A detailed trend analysis of acquisition volumes, cost of acquisition, initial giving value, retention rates and LTV over time, cross-cut by acquisition channel (volume, cost, retention, LTV and ROI per channel), and — within Face-to-Face specifically — by external agency, recruitment type, location and campaign cause.

The analysis translated directly into a set of insights and recommendations the charity could act on across channel mix, age-based targeting, campaign cause and localisation strategy:

1  Channel mix: headroom in digital

Digital channels showed stronger financial performance than Face-to-Face, with strong annualised ROI. Although F2F volumes and CPA were both moving in the wrong direction, there appeared to be headroom to shift investment toward digital — while recognising F2F’s continued strength in generating volume that digital may struggle to match. F2F itself proved more volatile, with some recruitment years generating a negative ROI, or an ROI implying that other uses of the money — even a high-interest savings account — would have performed better; other years, by contrast, generated strong returns.

2  Age profile: differentiated charging for Face-to-Face

Within F2F, ROI proved sensitive to swings in CPA, with cost jumps in certain years tipping the channel into unprofitability. Since younger recruits broadly generate a loss or are considerably less valuable, Adroit recommended agencies (who typically charge per recruit) adopt a charging structure that charges less for younger sign-ups and links cost per sign-up to initial monthly value. For in-house teams, where volume matters more to the overall economics, the recommendation was to target older donors for a higher value while still allowing drop-asks to secure a sign-up — and to keep recruiting younger donors regardless, since they carry no hard cost, still generate revenue, and help bring down overall CPA.

3  Campaign cause: not predictive of retention

Cause did not appear to be a strong predictor of Face-to-Face performance once looking at post-sign-up behaviour — retention rates were very similar across causes. Its influence on converting street contacts into F2F givers in the first place could not be assessed, since that conversion data wasn’t available to the analysis.

4  Localisation and agencies: promising but inconclusive

Sign-up location ranked highly in feature-importance analysis when left at a granular level, but with many categories and small volumes per location it was difficult to say confidently that one location outperformed another, or to rule out noise. Adroit recommended to group individual locations into categories — by type of location, income level of the area, and similar attributes — and supply more detailed, agency-level CPA data to support a deeper follow-up analysis.

The clearest overall conclusion is that Gross LTV performance is determined above all by the value of the initial sign-up: the higher that initial value, the better a donor’s LTV goes on to be. With cost now built into the same model, the charity can weigh that relationship against CPA, channel and agency economics — and is positioned to sharpen location and agency-level targeting further as more granular data becomes available.