The ascent of data science is increasingly both undeniable and self-evident.
To one degree or another, nearly all large companies are engaged in it and see emerging artificial intelligence and other data-related technology tools available to them with great promise.
Joel Shapiro, an associate professor at Northwestern University’s Kellogg School of Management, observes that data science has proven immensely valuable to some companies; Target, for instance, turned to it intensively in the 2010s as many of their performance metrics were underperforming, resulting in improvements in operations and efficiencies at a time when the company was seeking both.
The challenge, however, is that data science works best when it is fully aligned with management and its performance objectives.
Such alignment is rare, however.
More commonly, Shapiro writes in this Harvard Business Review article, companies find a “misalignment between expectations at the top of the organization and the foundation of what data science can realistically deliver.”
What accounts for this misalignment?
Sometimes, Shapiro argues, it is executives’ excessive expectations for what data science can realistically deliver.
Unreasonably inflated expectations are misguided, however, since they “can reflect a lack of appreciation of just how impactful small improvements can be—for example, slight increases in profitability per customer or conversion rates,” Shapiro writes.
But not all blame for such misalignments between executive and data science teams lies with executives’ oversized expectations. One other major contributing factor is the tendency of consulting firms’ sales teams to overstate these expectations in a bid to win a company’s business; this misaligns expectations from the get go.
Shapiro lays out three steps that can assist in resolving these and other challenges, resulting in closer alignment between management and their data science teams—and make the great promise of data science even more promising to companies: #1 Give a dose of reality; #2 Build on past successes and achievements and; #3 Let data scientists do the talking.