Which type of Segmentation do you need? What is it for?
Segmentation is one of those customer insight and marketing terms that divides opinion.
Leaders have their favourite approaches. Boards can be ardent fans of the need for segmentation, or complete unbelievers in what is perceived as marketing “spin“.
One of the reasons for this appears to be, the mixed fortunes of implementing segmentations. Some companies extol real benefits and the focus that has come as a result, whilst others bemoan wasted spend with consultants and agencies.
The case for more than one segmentation
My own experience is that an appropriate segmentation can add real value and enable a clearer understanding. It can help a business to focus on appropriate target audiences. But a few misconceptions need to be addressed.
The chief misconception I would cite is, the belief that any company or market only needs one segmentation. One of the guiding factors for selecting the most appropriate segmentation approach is the purpose for which that model will be used.
A segmentation to guide market participation strategy, is a very different challenge, to one for new proposition development, or to target different customer treatments. For this reason, it can be beneficial for a company to have more than one way of segmenting it’s customers (even if one is considered primary when seeking to embed into the culture of the organisation). One analogy for this is the benefit of having a Rubik’s cube set of segmentations for decision making.
Which type of Segmentation do you need? What is it for? It's a misconception that a company or market only needs one segmentation. Click To TweetWhich segmentation approach is suitable, or in favour?
Once the challenge of identifying the purpose of the segmentation is overcome, using incisive questioning, a CI leader needs to select the most appropriate tool for the job.
Here there does appear to be a degree of fashion influencing choices over the years. Many years ago, simple demographic segmentations were popular and can still perform a useful function. At the height of influence from market research teams, attitudinal segmentations were favoured and are still more viable than many believe. Since the success of Dunn Humby and others, behavioural segmentations took centre stage.
Directors, particularly finance directors can favour value-based segmentation and operations directors can favour a simpler life-stage or needs-based segmentation.
The potential of hybrid segmentations
All these segmentations have had their day, and most still have their advocates. So, it’s not surprising to find more organisations these days with hybrid segmentations (combinations of the above options).
Popular combinations for hybrids appear to be value+life-stage or behavioural/trigger+value-based segmentations. Having once achieved developing a rich attitudinal segmentation, from substantial quant research and then producing predictive models to overlay this onto a data warehouse for targeting – I regret how often attitudinal segmentations are dismissed nowadays. If you have the behavioural data to model such an overlay, it can be a rich & insightful approach.
However, my guidance to customer insight leaders is, to be aware of as many potential approaches as possible. Then focus your efforts on being clear as to the purpose of any potential segmentation. At the end of the day, it is not a ‘universal truth‘ about customers, it is just a modelling approach to achieve a specific purpose.
Which segmentation approach is working for you?
Since originally writing the above blog post back in 2014, we have published a number of other posts on the topic of segmentation. So, you might also want to read:
- Customer Segmentation in Cognitive Computing age
- Marketing Personas, fact or marketing fiction (part 1)
- Marketing Personas, fact or marketing fiction (part 2)
- Stakeholder segmentation & video personas
- The dangers of Public Segmentations
I have also since read more about the Bain inspired “jobs to be done” approach to consumer segmentation. That is also worth of a blog post for more nuanced consideration. So, expect one in future.
But what about you? If you are a data, analytics or research leader who has successfully implemented a consumer or market segmentation – which approach worked for you? Please share your experience & advice in the comments box below. We’ll see how much this topic divides (segments) opinion!
Too true.
We used to see a new segmentation every 18months…..and a new Marketing Director…yep you guessed it every 18months.
On a recent enquiry for a discussion on a new segmentation we turned up armed with all the previous variants, complete with pro’s and con’s, and asked the simple question…”What do you want to do with the segmentation”. Basically they wanted it to solve every marketing challenge and then resolve world peace. We asked them to go away and come back with a slightly more focussed need based on our experience. We never saw them again. Allegedly they wanted to tick the segmentation box.
Thanks Graeme, sadly I recognise your experience of marketers not being clear what they want a segmentation to do but knowing they must have one. Well done on challenging them, it seems their need was not so great. I’m interested, does that mean you continue to operate without any customer segmentation for your business?
One of the hybrid approaches I’ve seen work well, and which partially addresses the challenge of balancing the simplicity of a single dimension model vs a more insightful multidimensional one, is the attitude and life stage approach. Whilst attitude is a powerful insight tool for strategy formulation, adding a life stage dimension brings a higher degree of ‘targetability’. This can be derived by adding a life stage element during the clustering process. Whilst this may diminish the likelihood of finding pan-generational attitudinal segments, it allows for easier execution. And of course it allows for more clearly defined pen portraits, which are always a sought after piece of collateral following a segmentation exercise.
Agreed, Alan, that approach can be powerful and is at the least a useful first step to getting an organisation to engage with and understand a customer segmentation. I suspect the addition of the life stage variable also makes predictive modelling of the clusters for overlay easier or were you thinking of simpler operational proxies?
Either way, I agree with your point about the importance of ‘pen portraits’ as well. These can be very evocative (I still miss some life-size ones I had cut out of aluminium), but there is also the risk of a rich segment description being reduced down to a demographic stereotype, so caution is needed.
Thanks for your advice.
Yes, I wholeheartedly agree that “appropriate segmentations can add real value and enable a clearer understanding and focus on appropriate target audiences”. I have seen for myself, just how well they can work.
Their likely longevity is a particular issue though that should be a primary consideration before investing in one. It depends on the sector of course: Funeral directors are probably lucky in being able to rely on just one segmentation remaining valid for a life-time (if you will forgive the pun). Social Networks and Mobile Operators are probably faced with a need to re-fresh on a far more frequent basis.
Unless refreshed at the ‘right’ time, there is a danger that segmentations may lose their value very quickly: Refresh too soon and it’s a costly replication. Refresh too late, and there is the danger of misreading the market and failing to respond (quickly enough) to change.
Crucially, even the most excellent segmentation work does not extend its longevity. It is change in the market, of course, (and the impact that change has on attitudes and behaviour), that determines that.
And so: should we move away from segmentations altogether and towards a more fluid and fleet-of-foot analysis of ‘big data’ – accompanied of course by insightful and expert qualitative diagnosis (done relatively cheaply, regularly and at great speed)?
Hi Paul. Distinctions, I find, are often needed between the practice of segmenting a customer database and segmenting the market/category (customers and non customers). The latter is more strategic. The former more operational and tactical. Linking between the two can be useful and you need to know what your purpose is first. There are arguments made by the Outcome Driven Innovation (ODI) school that you also need an separate segmentation for innovating.
To make segmentation actionable you need to be able to capture some business value metric so that you can use this to prioritise and allocate resources in line with business objectives.
When you involve international global segmentations it is often best to develop an overall set of Global segments, then break down to country level. Going the other way seems much harder, unless the marketing teams are very separately managed.
Thanks for the advice, Martin. Sounds like quite some practical experience there.which approach has worked best for you in the past? Any surprising success stories?