Nick Mason deconstructs the stereotypes that traditional forms of data segmentation can paint
Ever since the promise of a revolution in fundraising by Ken Burnett’s epiphany of donor-centric approaches, segmentation and personalisation have held a great lure to fundraisers as to being the key to improving profit.
The logic is irrefutable: if you treat individual donors as individuals – or at least give that impression – then you are more likely to raise money. Of course, only major donors typically are lucky, or rather valuable, enough to get true individual treatment. But, as the box below shows, there are countless ways of segmenting an audience and so tailoring a message,
‘Big data’
Despite the obvious logic, there remain several reasons why so many segmentation projects sit on dusty shelves, while fundraising teams continue with one-size-fits-all communications. Not all of these are easily resolved, but the advent of ‘big data’, which is getting innovators across the globe increasingly excited, might be able to unlock some of them.
First of all, there is an unresolved tension between the different approaches to availability and applicability of data. On the one hand, if an organisation wants to know about potential audiences – often called a ‘market sizing’ segmentation – it will find there is very little actual behavioural data, so the segmentation is likely to be based on geodemographic information, possibly including attitudinal and claimed giving habits. On the other hand, if you are looking at your existing donor base, then you probably have more information about their giving, and so are more likely to use a behavioural segmentation, overlaying geodemographic data to provide more insight.
Critically, most charities have not been able to combine these two approaches effectively. If you have two differing versions of the truth, all you know is that one is wrong. And so it is with segmentations.
Secondly, there is an economic argument – often it is easier and/or cheaper to use the same materials, mailed to everyone. This can come down to the nuanced difference between profit maximising (targeting net income) and revenue seeking (targeting gross income).
Pen portraits
A set of ‘pen portraits’ is one of the key products of a segmentation, but also one of the most insidious. When you collect together the typical characteristics of individuals in a particular segment, you end up with a stereotype. But they can often hide huge variances, or give false positives. For instance, a pen portrait is likely to feature just one individual, who will be either male or female, despite the fact that the split of the segment might well be close to 50/50. Indeed in some cases a pen portrait might legitimately show the gender of the minority group, because they over-index.
Similarly, averages can be very misleading: an ‘average age of 40’ can hide all sorts of nuance – it might be a tightly defined group of 40 year-olds, or a random sample of the whole UK population, which also has an average age of 40. It could even be two equal-sized groups of 20 year-olds and 60 year-olds, with very few 40 year-olds at all!
The long tail
Marketeers are only interested in a small part of the segment. As the old saw goes, “we’re wasting half of our marketing spend, but I just don’t know which”. Segmentation can help narrow things down, but it’s never perfect – and the smaller the group, the more likely that it’s actually very different from the segment itself.
Indexing problems
A further pitfall lies in the use of indexing, where the incidence of a factor in a segment is benchmarked against the average. Where the difference is significant, this will often be highlighted, and reflected in the pen portrait. The risk is that this ignores the baseline. For instance, newspaper readership is often of interest, both for buying media and as a way of approximating the sorts of attitudes that the individuals have. A segment might over-index Guardian readership, and under-index the Daily Mail. So while media buying might be better targeted via the Guardian, because its readership is much less (about a quarter), it is probable that more people in the segment are Daily Mail readers. It’s these sorts of statistical and mathematical nuances that make the proper use of segmentations tricky.
The impact of digital
With all these technical and financial challenges, it’s no surprise that segmentation and personalisation have not brought about the promised revolution. However all this might be about to change, with the increased access to data and the flexibility that digital activity brings, just as the commercial world has been turned on its head by the ability to target individuals.
The more agile charities are starting to use more complex web advertising strategies, and analytical tools to track donors’ online behaviour. These are making it easier to combine attitudinal and behavioural segments, enabling personalisation and enhanced supporter retention. As print becomes more flexible and digital comms more easily tailored, so will the financial obstacles diminish. Of course, there is a high chance that everything will change again as more disruptive technology arrives on the scene, but the direction of travel is clear.
A segment of one
90 per cent of data in the world today has been created in the last two years. If PCs define the 90s, and the internet the noughties, then data is, to quote McKinsey, “the next frontier for innovation, competition, and productivity”. As the tools for analysing and accessing this data are developed, then the pitfalls and difficulties experienced to date will fade away. The accuracy and value of data will grow, finally leading to a personalised experience, reflecting all the aspects of an individual – personal, behavioural and attitudinal – that make them unique. A segment of one.
Segmentation at the RNIB
At the RNIB, we have a market segmentation which splits the UK population into eight segments. We are particularly interested in three of them, each about 10 million strong. The first we call ‘upstanding altruists’; they are typically well-educated, busy retirees who have had professional jobs, but who aren’t necessarily immediately engaged with the sight loss cause. The second group we refer to as the ‘family first (small “c”) conservatives’, who are older, more traditional charity supporters, and quite likely to have experienced vision problems. Our third group we call ‘youthful engagers’. This is a great example of the problem of stereotyping – although the majority are under 50, many of them are not necessarily young! But as people who are close to someone with sight loss, and like to be ‘charitable’, we see them as future donors.
At the same time, we have segmented our donor base this time into 10 groups – for example – ‘big-hearted donors’ (who give a lot, but not just to us), ‘raffle loyalists’ (who respond well, but behave as if they appreciate something in return), and of course the idiosyncratic group that are face-to-face recruits.
Overlaying these two approaches is an ongoing exercise, but one that helps us get closer to understanding what is the right ask, to the right person, at the right time.
Nick Mason is head of fundraising strategy at RNIB
This article first appeared in The Fundraiser magazine, Issue 20, August 2012