Helping link Analytics and Insights to Profit Growth

Irrespective of the data rich industry - banking, financial services, retail and telecommunications - there's quite some disconnect between the marketing hype surrounding all the "nuggets" waiting to be found in the world of data mining, and the actual business benefit to be derived from having actually discovered them.

Herewith one approach, using a banking case as an example, to help resolve this apparent disconnect.

The business challenge and the gap between Business and IT
On the one hand you have the data architects and data scientists who are able to do considerable magic with data, from sourcing it to cleaning it to storing it in a more accessible, central form to finally strategically analysing it. 

On the other hand you have business, typically involving marketing, product management and channel management, all desperately wanting to find an edge to their various strategies in mature markets, all aiming for the ultimate goal of achieving "an unfair advantage" and thereby achieving or exceeding their sales objectives and increasing their profitability. 

Figure 1: Insights from a data mining intervention in a top bank, from a
presentation made by Guy Pearce to the Academy of Marketing Science
in Vancouver, BC

The abyss between the two can typically be summed up in that one group is simply more comfortable speaking about numerical analysis than the other group is.

From figure 1 for example, the data analysts would say that the SMB need for banking products at the per-product and per-segment level clearly varies depending on the amount of time a customer has been banking with us (some increasing, some decreasing), and that many of the results are statistically significant at the 80% level. It would seem obvious to them that product, channel and marketing should leverage this insight in their domain, perhaps in the form of at least a test case to begin with. But alas, this is where moving forward often comes to a halt. To some degree, the break may arise from the implied assumptions both teams make at this point.

The implied assumption from business is perhaps that this is as far as the analysis can go, and have some difficulty asking the next questions because the insights, however valid, cannot cleanly be used as is. On the other hand, the implied assumption from the analysts include that marketing would know what to do with these insights and that they would thus be able to direct the follow-up actions. In reality, little happens, with a probability proportional to the innovation involved in the analysis.

Bridge-building

Figure 2: The business results of actioning data insights, as implemented by
a business and IT matrix team, from a presentation made by Guy Pearce
to a Teradata conference in Las Vegas, NV

As a way of building a bridge between the two groups, I suggest that both teams get used to asking "So what?" of each other, with the analysts answering this question proactively in their presentation to business, and the business teams asking it until a point of improved mutual understanding about the insights is achieved.

Another way for IT to build a bridge is for them to let business know that each individual with sales potential can be individually identified and profiled from the analysis. Simplistically, building on this angle normally sets the audience alight, and then the action begins in earnest, at pace!

Once it is agreed that acting on the findings could result in sound business value being generated, the teams should propose a test case to either prove or disprove the findings in a real world context. The results of the test case will drive subsequent actions, either halting the process, or ultimately resulting in a regional, provincial or national rollout depending on the success and scope of the initial case, to the ultimate benefit of both the banks' customers and the bank's financials.

In closing
By no means would I ever suggest leaving statistically significant findings on the table without action, since this particular type of analysis might turn out to be exactly the edge the business is looking for, such as the application of the insights such as those in Figure 1, resulting in the business benefit outlined in Figure 2!

Ultimately, measuring the worth (the return on investment so to speak) of the analytics team means they regularly need to know whether a type of analysis is useful and practical, and based on this, whether the particular stream of analysis should be extended or halted. The only way sufficient information can be gained to answer this question properly is to regularly test their findings out in the market. Without this feedback, many analysis teams end up in a black hole of perpetual analysis, with little of their work being validated, little of their work being used to generate value for the organisation, and thus often resulting in sub-optimal team self-esteem... Best make a plan to see if those nuggets you're ignoring aren't real gold!