When your data doesn’t tell you the whole story: An insurance industry parable

In my crusade for process improvement, I was poking around looking at sales cycle time data for the financial services verticals. In fact, I have a little Six Sigma project that I’m working on to identify factors that influence cycle time in analytic technology buying cycles in financial services. My going-in assumption was that all types of financial service companies (banks, insurers and investment managers) make the same kinds of buying decisions in similar time frames. Analysis of my data proved me wrong.

While sales close rates were consistent across industry verticals, I found that the cycle time for insurance was almost 40 percent longer! I’ve spent my whole career in insurance, so I obviously had no experience with how banks make technology spending decisions; and if I just looked at the data by itself, I could jump to all kinds of conclusions about the effectiveness of one sales team or process over another. Instead, I began to search for context. My first step was to interview several internal insurance account executives. The prevailing theory centered on a key difference in the insurance versus banking business model: Banks act more like retail organizations because there is an “immediacy” in their customer relationships. For example, as a consumer, I may interact with my bank on a daily basis, but chances are, I only interact with my insurance company a couple of times a year. The theory being that banks are more likely to make faster technology buying decisions because they need to keep up with changing customer needs, demands and risks in near real-time.

Again, this is just a theory, and like any crusader for truth, I am seeking to validate. However, this will be a significant challenge because there isn’t any way to quantify this theory – the best I can do is qualify it. My next step was to find external research on the topic.

In May 2011, the research firm Celent published a paper entitled “North American Insurance Software Deal Trends 2011.” They tracked 1,764 deals between insurers are software vendors in the 2009-2010 period. The authors note than in that period, “although the economy created a strain on IT department budgets, deals were getting done, but the decision timeframes were longer than in previous periods…the 2011 analysis reflects the same extended decision times.” Okay, but wasn’t the banking industry even harder hit over this period? Why were they still making faster buying decisions?

I checked the Standard and Poor’s Indexes for Insurance (^IUX) and Banking (^BIX) to see if stock prices had precipitously fallen in insurance compared to banking, increasing the reluctance to make significant capital expenditures. No luck there – both indexes are well correlated.

The consulting firm McKinsey & Company wrote in a Spring 2011 article “An IT Growth Strategy for Insurers” that “some insurers may find crafting a new approach to technology difficult…they often see IT primarily as a cost center prone to overruns and a megaproject mentality” and “the industry is built on high levels of trust in product offerings and often on personal relationships…as a result, insurers fear to experiment with new technologies that could damage these fundamentals.”

While this information doesn’t give me any cold, hard facts, I can now add additional qualitative context and information to my analysis. If I just stuck to the data that I had, I would have not been telling the entire story (or a fair story, from the account executives’ perspective). Two lessons learned: look at the entire ecosystem of information available to you (qualitative and quantitative), don’t jump to conclusions, and if you lack context, seek it out. Now I’m on to the next challenge – how do I make this insight actionable!?

tags: analytics, insurance, six sigma

Post a Comment

Your email is never published nor shared. Required fields are marked *

*
*

You may use these HTML tags and attributes: <a href="" title=""> <abbr title=""> <acronym title=""> <b> <blockquote cite=""> <cite> <code> <del datetime=""> <em> <i> <p> <pre lang="" line="" escaped=""> <q cite=""> <strike> <strong>