The first edition of Big Data London, replete with Communist iconography, this week (3-4 November 2016) has been buzzy and busy, and testament to the rude health of the data science scene in and around the capital. The familiar data world vendors are here, and there are data analysts (and, no doubt, scientists) lining up to hear them talk.
Matt Aslett, analyst from the 451 group, gave a promising-looking talk on the “myths” of big data. And, more interestingly still, he told us he was using the word “myth” in the traditional sense of sense-making story telling. Quite right, too. Narrative is, after all, one of the strong imperatives of human culture: vide Fredric Jameson’s The Political Unconscious: narrative as socially symbolic act (sticking with the classical Marxist theme). Matt did not, that is to say, merely deploy “myth” in the sense of “total nonsense” (as in, “that’s just an urban myth”).
Instead the myth he sought to deconstruct (a nod to Jacques Derrida as well as Jameson — one of the most important theorists of postmodernity, after all) was that of big data as a fetish governed by the three Vs of volume, variety and velocity. Instead, he contended that you should always start with a business purpose, then do the data work; not gather the data and then examine it for glimmers of business value.
Jonathan Ellis, CTO and co-founder of Cassandra distributor DataStax made a similar point in a briefing I had with him at the event. Jonathan said: “crossing the chasm from early adopter to early majority means being a technology company is not enough. You need to articulate how you will solve business problems to a degree that the NoSQL [movement] has not done so well to date.
“That’s why I am excited about the customer 360 use cases that graph database technology enables. Businesses are very aware that they need to up their game around customer experience. Now, if I come in and say: ‘graph enables you to do path-based queries’, no one would care. But if I say graph helps you build customer 360 analyses 3 times faster than relational technology, that gets their attention.”
It may be boring, and we have all heard it before, but data work needs to have a business purpose. Unless it’s improving operational efficiency, opening up new revenue streams, or is vital to a wider business transformation programme data is (in and of itself) , as Matt Aslett puts it, the wrong place to start.