Great article — maybe 2017 will be the year Big Data goes through the “trough of disillusionment”? There are signals pointing to the realization many things are wrong with digital marketing, digital analytics, and data science in general. In all cases, the root cause isn’t technology nor is it data, it is “people”: bridging the gap between various teams; difficulties in embracing modern, agile methodologies; and lack of discipline in “eating their own dog food” by measuring the results of what they do.
Honestly, I’m not really surprised by what remains an anecdotal, non-scientific, unverified poll of 150 data scientists at a conference… As you mention, I would have expected a few hands to go up — if only from over confidence or pride!
I work in the field of digital analytics, which is closely related to Big Data. This industry has been at it for at least 10, if not 15 years. Yet, when asked about “the best strategy to define actionable KPIs”, 42 experts came out with 42 different answers. To me, this was quite revealing of the (immature) state of our industry. After all, the question wasn’t about the best KPI in a given context — it was about the methodology to define actionable KPIs.
In 2009, I developed the Digital Analytics Maturity Model, a framework that helps organizations assess their current situation and provide a structured, actionable path towards improving competence at leveraging data and analytics for enterprise-wide business decision-making. Basically, for several years now, I’ve been looking at strengths and weaknesses along five key areas of success. Two years ago I started asking people at conferences or attending my workshops (digital marketing and analytics practitioners, consultants, vendors), a simple question: “just between us, tell me, how is it really going for you (or your clients)?” The answers literally baffled me… (well… maybe not so much… that’s what I suspected!) This led me to write the Radical Analytics Manifesto — height premises to approach analytics in a different way (at least try something!). Those points readily applies to data science and some of them are exactly what you suggest.
Thanks again for writing this — it’s reassuring to see I’m not alone thinking about those challenges.