Analytics should rarely live alone- they should support (but not replace) a broader ecosystem of decision inputs. But we can empathize with a common challenge- “what do I do when I’m being told two contradictory things?”
Hence the adage, known sometimes as Segal’s Law: “The man with a watch knows what time it is. The man with two watches is never sure.”
Such was the case when we were helping a manufacturer set price on a particular assortment. They sold individual pieces, as well as a bundle that had 4-5 of the most popular eaches with a ~20% discount. There was some existing research that indicated that they could raise price (effectively reducing the bundle discount) for the product because bundle customers were looking for convenience rather than a discount. However, historical behavior patterns indicated that the bundles were very responsive to price changes, and that the impact of lowering the price would be offset by increased sales velocity.
Facing this dilemma, many firms will strive to chase down the underlying root cause of the disparity- which is rarely successful. It’s an awkward situation, whichever way you go, you’re going against one of the recommendations. And it takes courage to turn left when something is telling you to go right. Decision theorists have shown over and over again that regret is a powerful emotion and people will go to great lengths to avoid it. So outside of the rare actual error in the math, more analysis usually just delays the problem. We see incredible amounts of time wasted while approaches are reworked, data sources changed (and sometimes tortured) until the original result falls into submission. And of course sometime despite all this effort the original result stays doggedly fixed.
So what’s the right answer? To be a business leader, one has to be comfortable working with ambiguity. But as “big data” and analytics become a bigger part of research they must also function within a confluence, but not necessarily a convergence, of evidence.