The secret of capitalizing on customer insights
This great McKinsey piece finally made it to the top of my reading stack last night - Capitalizing on customer insights (subscription required). As I read through the paper's description of the challenges of using information effectively in the context of more customer segments, more distribution channels, more store formats and more products I was struck by the potential for Enterprise Decision Management (EDM) in this context.
The challenges in getting useful insight from a wide variety of data were well articulated.I particularly liked the concept of "cell-level insight" as it seemed to me to focus on the real challenge of very tightly targeted insight. The paper discusses an established Fair Isaac customer, Best Buy, as an example. One of the first things Fair Isaac did was help Best Buy hook up its revenue to individual customers, a key step to providing "cell-level" data. Lots more has gone on since and a couple of articles have been published on how Fair Isaac is working with Best Buy. This one in Viewpoints -Best Buy aims analytics at purchasing triggers - discusses the use of Peacock (an analytic modeling product) as a key part of the Best Buy targeting strategy and this article in Fair Isaac's Marketing Decisions Newsletter - Best Buy Plugs into the Power of Customer Centricity - discusses how the use of sophisticated analytics has helped Best Buy become more and more customer-centric.
However it seems to me that this is necessary but may not be sufficient. Once one has "cell-level" insights or very finely grained ones, turning these insights into actions, especially in high-volume operational systems, is a second key barrier. Indeed the paper discussed the need to embed insights into "key decisions" to get value. I wonder if these "key" decisions are strategic or operational. A company could have strategic decisions that are key (launch this product, scale back this one) but it may also have operational ones that are key (what cross-sell offer to make to a given customer, how to up-sell on the website). Getting insight into these high-volume transactional systems is a whole different challenge. For instance, applying insight to the point of purchase for most retailers means automation thanks to the volume of decisions.Indeed one of the signs that you need to improve listed in the paper is that elaborate segmentation models are developed but not used to set channel strategies. This example of analyzing customers 30 ways while treating them only 3 is a classic "You need EDM" trigger. One of the three characteristics of an insights-driven company is that it can embed insights into front line decisions. This is the promise of EDM.
The paper also explores the idea of vendor networks for improving insight by tying together different data sources, skills, services and analytics One interesting example of this is ScoreNet, Fair Isaac's approach to delivering integrated data, services, analytics and applications over a network.The paper talks a lot about the concept of merging internal company information with external sources of data to create more meaningful insight. ScoreNet can help customers with these "external data augmentation" challenges. However one of the issues in doing this is knowing what data might be relevant - something that requires data expertise in a vertical. Indeed Fair Isaac has done a lot of work "Collaborating with insights partners" by developing the kind of data consortia needed to drive down fraud in an industry for example. We call this being a Decision Service Provider or DSP and it is perhaps the logical end result of the process of capitalizing on customer insights.