Wednesday, April 22, 2015

Is Analytics Killing Strategy?

In this week’s reading Bringing Science to the Art of Strategy authors Lafley, Martin, Rivkin and Siggelkow discuss a scientific approach to strategy, which is “marked by rigorous analysis.”

Analytics and “big data” collection techniques were some of the fastest growing trends in 2014. More than ever, executives are on-board with leveraging their data to understand their customers and improve all aspects of their businesses.

The authors posit that a “scientific” approach to strategy relies not only on analysis but also on two additional key elements, the “creation of novel hypotheses” and the “custom-tailored tests of those hypotheses.” While this approach may sound simple, I feel the authors should expand upon the analysis piece of the process.

Given the rapidly increasing availability of analytics tools, organizations’ available data is now growing by orders of magnitude – strategists are likely coming into the process already armed with data and analysis. Is this a good practice? Should strategists rigorously analyze all of the available data beforehand? At what point are there diminishing returns to rigorous analysis?

If complicated upfront analysis is performed, strategists may not be able to get to the hypothesis part of the process in a reasonable amount of time. Or worse, hypotheses may not be valid.

Similarly, the authors posit that, “Having recognized that a choice needs to be made, you can now turn to the full range of possibilities you should consider.”  What if using analytics to draw conclusions about customers leads your strategic possibilities astray?

In a world with a glut of data, is it possible for the “logic of what must be true for success” to be completely sound? Sometimes what’s “genuinely new” can’t be quantified easily with data, and, as the authors mention, sometimes analyzing the status quo does not tell the entire story.

Finally, once strategists move to the testing phase, some possibilities may not look as promising as others “by the numbers” and data could preclude teams from testing it well.

In many of these steps, strategists should probably suspend prior analytical findings in order to test possibilities.

No comments:

Post a Comment

Note: Only a member of this blog may post a comment.