Guest Contributor Bill Pink
| January 15, 2020
Head of Global Analytic Leads
In a recent post, Nigel shared a buyer classification focused on what I like to call the buy and the why, more specifically consumer behaviour and predisposition. The concept is straightforward but it begs the question of how can we build a system of insights and analytics to bring this growth matrix to life.
For a few years now, we have talked about the value of using connected data intelligence to uncover otherwise hidden opportunities for growth. Connected data intelligence is based upon two main ideas:
- Consumer behaviours and feelings do not exist in a vacuum, and neither should any research or analytics of them.
- We must purposefully design how to bring data together for us to feed the analytics we need to run, because simply slapping data together may lead to less understanding than we started with.
Building a measurement system to connect sales and brand data at a consumer level is a great example of what connected data intelligence is meant to deliver. Below are three guiding principles for designing this system.
First, whenever available we should leverage passively measured observations of consumer behaviours rather than stated memories of purchase. In many categories it is simply too hard for people to accurately remember what they bought, which brands, where, how often and how much they paid…you get my point. And passive data is not bound by survey structures such as competitive sets for a given category. This is key, especially as category boundaries continue to blur and consumers move from options like traditional beers to flavoured seltzers and back again.
Second, history matters. If we only know how a consumer behaves at a moment in time, that is only a window into the behaviour at that moment. Instead, we need repeated measures of the behaviours of the same person over time to know their general behaviours, their anomalies, and what is bought in and out of season. And across consumers we then have a mix of recurring, new, and lapsed buyers in our data to work with.
Third, behavioural data alone is not enough - it captures the buy but not the why. Accordingly, we need to connect behavioural data to other forms of data to ensure we measure predisposition. This includes survey questions, neuro techniques, social data mining, … all the forms of consumer expressions of feelings and associations towards brands. This is where we must be purposeful and ask some tough questions. Can we sample directly off of the behavioural data to match survey and sales data per person? If not, do we have survey and sales data over similar time frames with common variables that we can use to fuse data sets together and impute consumer level observations? Or, if independent data sets are all that are available, will we need to rely on behavioural proxies in survey data that we can calibrate to the behavioural data?
And this is where the vision of implementing a linked buy and why measurement system meets the realities of connected data intelligence. Do you agree? Please share your thoughts.