Modelling methods

Most large organisations have:

• A regular giving programme or

• A membership subscription programme or

• A recurring lottery programme

Many have two or even three of these programmes.

For many large organisations they are the dominant income driver, so trustworthy modelling of recurring transactions is critical.

I have expertise in modelling regular giving, lottery and membership programmes, including acquisition, attrition, renewal, reactivation, upgrade and up-selling. 

The month-by-month dynamics of income and costs are complex. Without specific modelling skills, extrapolating over several years can result in an unwieldy model; hard to interpret, error prone and impossible to maintain.

I don't make assumptions about your recurring programme. I work to understand its details, and select the most suitable modelling techniques, to give the projections that you need with the least possible complexity.

Most organisations promote different methods of support such as:

• seasonal appeals

• merchandise sales

• visitor ticket sales

• virtual gifts

• emergency appeals

These lend themselves to simple iterated-RFM (recency, frequency and monetary value) methods which calculate how the supporter file could evolve over months and years, under varying assumptions.

However most organisations also cross-sell between support methods. Understanding how cross-sales drive income over time requires more nuance. There are typically dozens of cross-sell pathways and - without good modelling discipline - you risk creating a tangle of logic that seeds confusion rather than decision confidence.

So I work with you to identify the paths that matter most, and include these at the right level of detail.

Acquisition costs (CPAs) are always critical inputs for modelling an individual giving programme.

But incorporating CPAs into an overall model needs careful thought, as there is usually no single fixed CPA number per channel.

Considerations for your organisation might include:

• Different fixed and variable cost structure per acquisition channel

• Supplier contracts with volume discounts and tiered cancellation credits

• Variable returns from wear-in, wear-out and saturation effects

• Halo, carryover and brand-building effects

• Product innovation and test budgets

• Uncertainty in historic CPA calculations caused by imperfect media attribution

• Multi-step acquisition programmes e.g. from prospect to donor to regular giver

• Earned impressions from celebrities, artists and brand ambassadors

We'll discuss which of these are relevant, and how to build them into the model. The goal is always to elucidate your big investment decisions, without over-complicating things.

Investment in legacy marketing yields mainly long term returns. So unless you are working on a very long planning horizon (more than 10 years), it is immaterial to your risk profile.

This means that legacy income is often overlooked for short and medium term planning, but this is a mistake. 

For many organisations, legacies form a large fraction of gross income and, and so are a crucial buffer when other income is volatile.

It therefore pays to understand legacy income dynamics:

• How many latent pledgers do we have? How was this influenced by past marketing efforts?

• How will notification volumes change over the 'boomer mortality peak'?

• How could asset price growth - past and future - influence legacy values?

These are important questions for short and medium term financial risk, but often overlooked in the planning cycle.

Fundraising innovation often involves commercial contracts, and requires new ways to evaluate risk, return and growth potential. 

Each situation is unique, so where necessary we will discuss:

• Market sizing

• Cost and revenue sharing

• Minimum revenue and downside risk protection

• Pricing, margin and elasticity

• Contractual break clauses

• Time lags between expenditure and income realisation

• Tax treatment

• Fund restriction

With my expertise across both commercial and charity worlds, I can quickly interpret the commercial dynamics of your situation. 

You will gain clarity on how innovation risk fits into your total fundraising portfolio.

Fundraising directors frequently encounter binary risks, for example:

• A large grant application that will or won't be successful

• A privacy risk assessment that might render part of the supporter base unmarketable

• An innovation pilot that will or won't reach the threshold ROI (return on investment) that justifies full rollout

• A pending tax decision that could drive a permanent step change in costs

• A potential adverse publicity event that could drive a temporary spike in cancellations

An integrated financial model enables you to see these risks in context of the whole fundraising programme, and how they could play out in the short, medium and long term.

Fundraising programmes - especially recurring transaction programmes such as regular giving and lottery - are exposed to inflation

Income does not automatically rise with inflation, whereas costs unfortunately do rise.

Moreover, staff costs tend to rise faster than general inflation.

So all else being equal, net income will come under pressure over time.

The nature of that pressure will depend on the structure of your fundraising programme, and you need to understand these effects.

When presenting medium and long term plans at board level, you undermine credibility if you haven't considered inflation at all, or haven't given careful thought as to how it should be factored in. 

Discussing your cost structure and inflation exposure is a core part of developing an integrated scenario model.

Sometimes a model contains many unknown variables which materially affect the income trajectory. These unknowns might be CPA (cost per acquisition), attrition curves, repeat transaction rates and binary risks (see examples above).

For budget purposes we usually need a lower income figure that we are confident in reaching.

For other planning purposes we often need a central estimate with upper and lower confidence limits. For example "Year 3 net income will be around  £55m, and we can be 90% certain that it will fall within £49m and £58m". 

Monte Carlo methods can help us reach conclusions like this systematically; in place of unknown variables, we feed random inputs into the model, and record the outputs (e.g. net income and reserves). We do this repeatedly, say 1,000 times, and summarise the results including the average and extreme values.

This is definitely a more advanced technique and is - by no means - always appropriate. It is mainly used when there are several binary risks. In other situations, simply inputting model assumptions for 'best, middle and worst case' scenarios will give a robust answer.