The Shifting Sands of Demand for Segmentation

Data, technology, and cultural trends come and go – but when done right, segmentation remains as powerful as ever.

As researchers, we all know there are no two segmentations the same, regardless of industries, sectors, and markets, where a meta-analysis reveals some discernible trends that have emerged over the last 15 years. These trends, when combined with observations about contemporary trends in client demand for segmentation work, are important since they provide a looking glass into the possible future of this type of work.

There are clearly a range of different types of segmentation, a very useful distinction being between those conducted to aid in business strategy, be they in services marketing, retail or FMCG contexts, versus those used for tactical purposes, where the segmentation itself is a means to an end (e.g. personalization).  For purposes of this article, we’ll be focusing on the former – segmentation for strategic purposes, and trends and dynamics we have witnessed in the Australian market.

If we go back to the early 2000s it’s fair to say that most segmentation work for companies such as banks, telcos and insurers likely fell into one of two camps.  They were either extremely basic, developed as a simple representation of demographic variables, or overly complex and esoteric, typically leveraging a bespoke set of attitudinal constructs.  These attitudinal segmentations were correctly criticized for being interesting but not actionable and were typically shelved quickly after they were developed.  Little or no work was being done in Australia at the time which adequately struck the balance between these stakeholder needs, or put differently, which delivered on both.

That changed in mid-2000s when Telstra quite famously developed both B2C and B2B segmentations which were not only strategically insightful but also appended across their databases.  This was fundamental to Telstra’s transformation toward being a more marketing-led and customer-centric organization, which was of course fueled at least in part by the fact the organization re-structured along segment lines.

Accounts vary as to the effectiveness of this strategy, but one thing that can’t be disputed is that this effort set the high-water mark for how to develop and activate segmentation in the Australian marketplace.  Many other services organizations subsequently began to follow suit, all seeking to deliver a strategically actionable and forward-looking market segmentation, while hitting the silver bullet database attribution rate of 70% that was widely reputed to have been achieved by Telstra in appending its consumer segments.

Courtesy of Nature

Those experienced in the art of segmentation – which genuinely operates at the intersection of science and creativity – will appreciate that segmentation solutions don’t drop off the back of a truck.  They’re driven and influenced by stakeholder engagement, framing, survey design, and nuanced trade-offs at the analytics stage as to the trade-off between segment meaningfulness and richness on the one hand, and database appending accuracy on the other.

Getting this right is a balancing act, and there is in fact no hard and fast rule to say that 70% is the right number – it depends entirely on one’s objectives, and where one is willing to compromise – is it in strategic richness or predictive accuracy.  It’s rarely possible to achieve both at high levels, unless one is fortunate to be working in a category in which there are high levels of association between readily observable criteria and category needs, attitudes and traits.  These are rare!  But for many years in the noughties and the teens, the market had high expectations in this respect.

A notable feature of many of the segmentations many client organizations operated across several industry verticals or playing fields, segmentation frameworks themselves were often sought at the macro-organizational level.  This meant there was a trade-off being made for single customer-centric insight and simplicity, often at the expense of category or vertical-specific specificity or depth.  Notable examples of case studies in this space are GE Money (2012) and Simplot (2009), which while for very different types of organization, have in common the fact that the segmentation was framed to be at a very macro organizational and consumer behavioral level.

Courtesy of IStock

Over the last five or so years, and consistent with the rise of client organizations’ investment in their customer and other data, alongside their own internal data science capabilities, it is evidenced again a divergence in the style and type of segmentation work being commissioned in the services sector.  On the one hand, have arisen a new breed of hyper-complex, data-led segmentations, reflective of client appetite to leverage internal data assets as much as possible, before looking to external ones in building a market segmentation framework or view.

These approaches are referred as ‘inside-out’ ones, given they start by leveraging resource ‘inside’ the organization but typically end with the leveraging of those ‘outside’ it to gap fill and enrich.  The ‘inside-out’ approach makes moot the ambition of seeking 70%+ attribution accuracy sought in ‘outside-in’ strategic segmentations but present the inverse problem of landing on a sufficiently rich and powerful strategic market lens for those clients seeking that.

In parallel with this emergent trend of hyper-complex data-driven segmentations is that of client appetite for hyper-simple frameworks, which land and operate at a high level of generality rendering actionability severely challenging.  This approach to segmentation is particularly prominent – and of utility – in categories in which there is a strong inherent relationship between factors such as demographics and behavior in the category in question.

Good examples are superannuation and health insurance.  In these and others, demographics form useful lead indicators of behavior or needs, or partitions of the market – which are then overlaid by key dimensions or factors that drive actionable targets within the demographic partitions.  This type of segmentation also has the distinct advantage of being easy to grasp and understand by stakeholder groups, making them perfect for clients seeking to engage in cultural transformation and greater customer centricity, without over-complicating stakeholders.

There’s no doubt that the abundance of customer-level data is changing the segmentation game, but at the end of the day, clients need be careful not to get caught up in this fact and its availability when considering the type of segmentation best suited to their organization.  Careful consideration needs to be given to the ultimate use case(s) of any investment in segmentation before going down any path, which inevitably involves significant senior stakeholder engagement. Segmentations rarely play multiple competing roles well, so setting a course with a clear focus on outcome and prioritizing accordingly is essential in maximizing ROI.