This is the fourth posting in a four-part series on business insights.
As an analyst, one of the worst feelings is to conduct a great analysis, produce a unique and useful finding, and then see it rot away in a PowerPoint deck somewhere and never get used. In my experience, this is worse than the analysis that never yields a potentially useful finding.
The topics discussed so far in this blog series will help you better define and create business insights, but this is only half the battle, or maybe even less than half. In fact, Kevin Hillstrom of MineThatData recently defined four categories of findings on his blog, and only one of the four results in the business actually using the finding. Here are the four categories:
- The results simply aren’t actionable.
- The results are actionable but the decision makers can’t understand it.
- The results are actionable but the decision makers decide not to act upon them.
- The results are actionable and the decision makers act upon them.
As analysts we are often quick to lay the blame on the people for whom we create the findings if they end up in one of the first three categories – they didn’t define the question correctly for us, or they’re too dense to understand our elegant discussion of odds ratios or random forests, or they made a foolish decision to put resources toward a different project even though our finding suggested a much better option. While the reasons our findings fall into one of these four categories are sometime out of our control, there are steps we can take to increase the probability our insights will be used by the business.
The article that inspired this blog series – “Unleashing Hidden Insights” by Marco Vriens and Rogier Verhulst – provides a few practical questions to ask during the insight development and deployment process to encourage the adoption and use of the insights. Here they are:
- Do the decision-makers and influencers know and understand the relevant insights? This seems like a simple question to answer, but just telling a decision-maker about an insight or inserting it into a presentation is often not enough. It is best to make the insights available through multiple channels, including personal communication, and to reinforce these communications over time to ensure the insight stands out through the clutter.
- How credible are the insights? As Vriens & Verhulst state, Credibility is a function of: 1) how well the insights are understood; 2) the methodological rigor used to generate the insight; 3) the degree to which insights are validated more broadly; and 4) the degree to which the insights can be contextualized.
- Are the insights actionable? This one is tough to answer since “actionability” can be affected by many factors. However, one of the best ways to get a “yes” answer is to ensure your analysis is informing a specific decision and ensure the insights are delivered before that decision has to be made. Know the decision you want to influence.
- Are the insights accepted by the right people and enough of them? Insights often fail because we don’t get buy-in from the right person – a senior exec or key decision maker. This could result from not communicating well-enough with that person or because that person is resistant to the message. To counteract the latter situation, make sure to get enough of that person’s peers, superiors, or key influencers on board and make it harder for them to resist.
- Do the decision-makers embed the insights in the right way into their decision-making process? Our insights cannot have an impact until they influence a decision. There is an emerging field called Enterprise Decision Management (EDM) that focuses on embedding data and insights into business processes to inform and impact business decisions. It’s going to be important to the future of business analytics.
For those of us who analyze customer feedback these questions are extremely important. Many of the people we want and need to use our insights are going to be resistant because they “already know what their customers think.” For those of us who are outside consultants or research vendors, it is often even harder to get our clients to use our insights. However, the tips and concepts outlined in this blog series, if put into practice, will go a long way toward increasing the hit rate on our analyses and the impact of our insights.
Troy Powell, Ph.D.
Vice President, Statistical Solutions