The Segmentation Landscape in China

Always connected, eyes and fingers on a phone: the complete integration of digital into life in China. But then why do so few segmentations show meaningful, useful differences in digital attitudes and behaviors across segments? Making sense of the Chinese consumer landscape can be daunting. Without incorporating meaningful digital behaviors into a highly prized – and priced – piece of segmentation, questions arise: why was so much spent on a project that is neither strategic nor forward thinking? How can this possibly be used here? And why has it happened? In the increasingly connected world, understanding how to capitalize on differences in digital attitudes and behaviors is a necessity and this should be front-of-mind when designing a segmentation study. In China, segments with no digital differentiation are like shades of grey.

When digital is not well-represented by inputs into the segmentation, differentiation across segments in the outputs is much less likely. Why is this so often the case? What hypotheses exist to explain the status quo?

Hypothesis A: is it because agencies have always used the same approach or even questionnaire, regardless of market, that has not taken into consideration the cultural context and the fact that digital has overtaken life in China?

Hypothesis B: Or perhaps it is because when digital inputs have been used, they have been superficial on a platform/channel level, hence contributing very little to driving people apart into segments?

Why Do We Put Forward these Hypotheses?

Why Hypothesis A: As an industry, we are not great about actually acting upon new learnings and updating our old approaches with new thinking. For example, how many times have you heard ‘big data’ at a conference without hearing specifically what was done? How many times have you heard about falling participation rates with online research, without hearing about a solution? These truths are often pushed to one side to continue on with the status quo.

Why Hypothesis B: Maybe researchers and analysts involved with segmentation builds do not have a sufficiently detailed understanding and appreciation of specific contemporary digital attitudes and behaviors, so fall back on using something more generic. And experts in the digital space are usually not involved in segmentation builds: a typical research agency may not have dedicated digital specialists to deliver input on segmentations. Maybe it is a matter of time, or our old foe questionnaire length; and in the rush to get something programmed it is easier to retain basic inputs like purchase frequency or attitudes towards self. Again, we fall back on what has been done before. 

How Do We Achieve ‘Rainbow’ Separation?

It is a fine balance between higher-level and commonly used attitudinal and behavioral statements and the more granular digital inputs: both are needed in order to form a meaningful segmentation for a highly digitally-involved China market. But then again, with China accounting for 20% of the world’s population, a bespoke approach would seem fitting in order to maximize a brand’s potential. 

What does all this mean? The Research team at Labbrand is partnering with our colleagues at MADJOR, Labbrand’s digital transformation agency, to develop a new approach towards segmentation in China for our clients. The process follows four steps:

1. Form a hypothetical framework of digital drivers in China. What are the psychological motivators and drivers of digital behaviors and attitudes?
2. Stress-test that thinking with quantitative research in China. Does it hold true? Is what we thought was important actually important?
3. Construct category-agnostic ‘digital’ only segments through statistical modeling. Determine: what fundamental types exist? What are the specific and detailed differences driving these segments?
4. Use these segments, and learnings on which inputs were most effective at discriminating between segments, to check: do we still see higher-level attitudinal and behavioral richness differentiating? Is this an improvement for clients? 

Segmentation – What Are the Possible Barriers?

Without predicting the future, one hurdle that is foreseeable: since segments will be differentiated across multiple indicators (digital, behavior, attitudes, demographics) it will take a lot more time to arrive at a solution that makes sense: a solution where each segment’s essence feels consistently realistic across these multiple dimensions; a solution where we are able to recognize these segments in our daily lives; a solution in which character holds true across all metrics. It is likely this means more iterations of the model will need to be run, with more time invested to see if we have reached an optimal solution (and one with rainbow separation!). 

Segmentation – What’s the Prize?

There could be trade-offs, but the prize is that we optimize an approach that allows rich multidimensional segmentation of consumers which includes meaningful digital indicators. As the team progresses – we are currently on point 3 –we will release updates on our progress. In the meantime, tell us: what are the very best and very worst segmentations you’ve seen in China?

AUTHOR

• Chris WallbridgeBusiness Director