Beyond good intentions: Using propensity modelling to predict behaviour

Why good intentions don’t always translate to action

Whether we’re testing a new campaign, exploring a fundraising proposition or understanding supporter behaviour, we spend a lot of time asking people what they think, want and do. Understanding people’s attitudes, motivations and behaviours can tell us a lot about an audience. But there is often a gap between what people say they will do and what they actually go on to do.

This say-do gap can make it difficult to predict which ideas, campaigns or events will actually succeed once they’re launched. Someone might tell us that they will donate to a cause, volunteer their time or change their giving behaviour. Yet when the moment comes to act, those intentions do not always translate into action.

The say-do gap is especially prominent in the charity sector, where social desirability can heavily influence responses. People naturally see themselves in the best light possible, as generous, compassionate and socially responsible individuals. As a result, we often see a lot of overclaim when asking about willingness to support a cause. After all, is someone really going to tell us that they would never donate to a charity helping sick and dying children? Probably not…

The challenge is that what people tell us often reflects who they aspire to be, rather than how they behave in reality. Understanding this gap helps us get closer to predicting real world behaviour and ultimately, understanding how to turn good intentions into action.

Looking beyond what people say

This is where propensity modelling can add real value. Rather than relying solely on what people say they will do, it uses data on existing behaviours, values, attitudes and characteristics to better predict how likely someone is to take action or display a particular behaviour.

Propensity modelling can help overcome some of the limitations that come with self-reported data such as overclaim, imperfect recall and social desirability. The result is a more realistic view of likely behaviour, giving organisations greater confidence that the decisions they make are grounded in how people are likely to act, not just what they say.

Choosing between good ideas

Imagine a charity is exploring new fundraising ideas to diversify their portfolio. The survey results come back, and the good news is that audiences like all three ideas. The bad news? They like all three ideas almost equally!

Trying to choose between ideas which look equally promising on paper is no easy task, especially when launching a proposition requires a significant investment of time, money and resources. 

We’ve encountered this challenge at Humankind when helping charities evaluate new ideas and propositions. By combining responses to the concepts with data on existing behaviours, attitudes and values, we can understand more about the people attracted to each idea. If the audience interested in a particular idea shares characteristics with people who have engaged before, we can deduce that they are more likely to turn their interest into action.

We can take this a step further and understand the characteristics of the most engaged audience to understand who the concept is really resonating with. The audience with the highest propensity to engage can often differ by idea. This is where propensity modelling becomes more than just a tool to select the winning concept but one which can help shape wider audience strategy. For example, one idea performs better amongst existing supporters, helping the charity deepen engagement with people who already know and love the cause. Whilst another might attract people who have never interacted with the charity before making it a better option to grow and reach new audiences. Both offer significant opportunities, the right choice depends on what the organisation is trying to achieve.

Turning insight into action

Ultimately propensity modelling isn’t about replacing traditional research methods, it’s about strengthening them. In a world where budgets are tight and decisions carry real risk, having a clearer picture of what people are actually likely to do can make a real difference. Whether it’s about testing new ideas, identifying the target audience or shaping long term audience strategy, propensity modelling helps organisations make decisions with greater confidence and can help turn insights into real action.

We’d love to chat! Get in touch to find out how we could use propensity modelling to help your organisation make more confident decisions – sacha@humankind-research.com

Photo creds: Kate O’Keefe (https://www.linkedin.com/posts/kateokeeffe_cleanshot-2025-03-21-at-1749512x-share-7309798340257333252-5qDx/)