The Key to Analytics: Ask the Right Questions
People think analytics is about getting the right answers. In truth, it’s about asking the right questions.
Analysts can find the answer to just about any question. So, the difference between a good analyst and a mediocre one is the questions they choose to ask. The best questions test long-held assumptions about what makes the business tick. The answers to these questions drive concrete changes to processes, resulting in lower costs, higher revenue, or better customer service.
Often, the obvious metrics don’t correlate with sought-after results, so it’s a waste of time focusing on them, says Ken Rudin, general manager of analytics at Zynga and a keynote speaker at TDWI’s upcoming BI Executive Summit in San Diego on August 16-18.
For instance, many companies evaluate the effectiveness of their Web sites by calculating the number of page hits. Although a standard Web metric, total page hits often doesn’t correlate with higher profits, revenues, registrations, or other business objectives. So, it’s important to dig deeper, to challenge assumptions rather than take them at face value. For example, a better Web metric might be the number of hits that come from referral sites (versus search engines) or time spent on the Web site or time spent on specific pages.
TDWI Example. Here’s another example closer to home. TDWI always mails conference brochures 12 weeks before an event. Why? No one really knows; that’s how it’s always been done. Ideally, we should conduct periodic experiments. Before one event, we should send a small set of brochures 11 weeks beforehand and another small set 13 weeks prior. And while we’re at it, we should test the impact of direct mail versus electronic delivery on response rates.
These types of analyses don’t take sophisticated mathematical software and expensive analysts; just time, effort, and a willingness to challenge long-held assumptions. And the results are always worth the effort; they can either validate or radically alter the way we think our business operates. Either way, the information impels us to fine-tune or restructure core business processes that can lead to better bottom-line results.
Analysts are typically bright people with strong statistical skills who are good at crunching numbers. Yet, the real intelligence required for analytics is a strong dose of common sense combined with a fearlessness to challenge long-held assumptions. “The key to analytics is not getting the right answers,” says Rudin. “It’s asking the right questions.”
Posted by Wayne Eckerson on June 10, 2010