How many times have you come across a situation where traditional qualitative research, let's call them what they are, surveys, just don’t deliver and, in fact, create more problems than you started with?
I have been in the data game pretty much all of my career, so that’s a long time, and I still see this fractured line existing between those that do behavioural data analysis and those that do research.
One of our clients, an enormous conglomerate in the B2B space, had invested heavily, I mean millions of dollars, in developing a whole suite of customer segmentations to identify their customers and what their needs were.
This was implemented and duly disseminated across the organisation through a significant programme of training and internal communications.
Posters of these customer types were plastered everywhere, including as table tent cards in the office canteens.
On the surface, it looked like an incredible initiative and one that, if you read the annual report, you’d be forgiven for believing that the business had got the full customer-centricity bug and you could expect huge changes in the way it was going to operate and serve its customers in existing and a raft of new product lines.
However, we were called in because they were experiencing some real difficulties in how they applied the right customer segment and profile to each of their customers.
This was a pretty important issue as the business was also on a path of huge transformation in the form of digitising everything it did.
The impact was that all of the personalised customer journeys, product targeting, communications, etc., that were being enabled through these new digital platforms required knowing which customer sat in which segment to direct the systems on how to treat each customer.
As they were releasing more and more new functionality and services, it was becoming apparent that many, well, a large majority of customers were getting the wrong experience – sometimes really inappropriate.
Imagine a FTSE100 company being sent down a customer experience designed for a startup.
It didn’t take long to unearth the problem.
The customer segmentations and profiles had been built using the responses from a 2-3 page customer survey asking each customer about them and their needs.
On the surface, this sounds reasonable, but this company had more than 50,000 customers and they got less than 1,000 responses – many of which were quite patchy in their responses.
You might be asking, on what planet did someone think, “well, that will do, we’ll just have to make the best of it”?
Well, someone did and created some very elegant, very believable needs-based segments based on this core research and had the temerity to produce a 2-page Word document justifying the statistical validity of the responses.
I can hear the sound of anyone who sits in the behavioural data camp heads banging gently against the walls of their office cubicles as I write.
This stuff is still happening!
Anyway, as I’m sure you can imagine, this led to all kinds of technical challenges and the need for manual efforts to re-allocate clients to segments accurately.
Consequently, confidence in the segmentation's precision and understanding of client needs was severely undermined.
Our solution was to take all the historical transactions, products, services, and touchpoint interactions with each client and create a behavioural-based segmentation to identify core, distinct behaviours that were significant and relevant to defining the right proposition and experience for each customer.
This isn’t new, in fact, it’s the same solution that has been used by supermarkets across the world to identify and group up different customer types.
The methodology is simple and is essentially using technology to do what we, as humans, do naturally.
You know when you are standing in the queue at the grocery checkout and you cast your eyes over your neighbour's shopping and formulate a pretty instant set of judgements as to who they are and what they like?
It’s exactly that, but using technology to do it at scale and preferably being slightly less judgy or biased than we might be as individuals.
This new solution, because of where the source data was collected, was joined at the hip with the systems, products, and services that made up the client's business and therefore connected in seamlessly to their digital platforms and did the job originally intended.
With this new solution came further benefits. The client was able to use the segments to pre-select research candidates based on behavioural traits, to understand more about the attitudes that drove the behaviours they were seeing.
This introduced the more human angle into play, but in a way where it was no longer disconnected from everything else but equally joined at the hip and therefore could be applied across all customers.
This new research in concert with the behavioural segmentation together drove the new creative execution of the communications and customer experience that finally matched the right customer behaviours and needs.
Lessons Learned
Data Literacy and Trust: Ensuring the business understands and trusts their data is fundamental. In the case above, you can’t help but wonder why they chose a research route in the first place, given the data they were sitting on. Lack of trust, often stemming from data literacy issues, can severely affect the adoption of new segmentation strategies.
Balancing Research and Behavioural Data: While research and survey data are hugely valuable, they must be balanced and used in concert with actual transactional data to bring any long-term, sustainable value and measurability.
Granular Personalisation: A deep, granular understanding of client needs is key to delivering targeted and personalised sales and marketing. Expecting to get these through sample-based surveys is just being too optimistic, as the number of options and permutations available to customers now is so huge a survey could never capture them.
Conclusion
To realise the full potential of segmentation, a company must blend analytical rigour and soft human understanding with practical application.
This company's eventual evolution towards a data-driven behavioural approach exemplifies industry best practices, enhancing customer understanding and enabling more personalised engagement.
Combined with a commitment to a test and learn strategy that prioritises quick wins and builds upon data capabilities, this will allow the company to leverage segmentation even further.