Big data is on the cusp of being either a really useful concept – or yet another clumsy marketing buzzword that never quite squares with reality. The first time I heard it, it was obvious that big data was a “retread”. A new word for something that’s been around for a while. Big data, I realised, is just a new word for analytics.
Why the new name?
Concepts do need new names (at least in the media) when they are about to enter a new market and reach new users, or just change their emphasis. Cloud computing, for instance, offers much the same thing “ASPs” offered ten years before, with the difference that this time round it is going to work.
Similarly, analytics has been available for many years, as a high-cost service using high value supercomputers, and operated by white-coated high priests who have come into the field from linguistics, philosophy and computer science.
If you have a big data set, and the money to have it explored, analytics has been there to reveal the secret trends within you information, which might give your business an edge.
In the last few years, cloud services have made more power available using ever-improving hardware. Open source projects such as Hadoop have made it possible to navigate giant datasets without a massive commitment to proprietary software.
In short, analytics, once an elite occupation, is coming to the masses. And in that area it needs a new identity. Hence – big data.
The promise of big data is that huge stores of information can be accessed on the fly, for very little money, producing significant results. As before, the aim is (usually) to provide a business advantage. But the exclusive club has opened its doors to anyone who wants to play.
This shift makes it tough to get the story across to potential customers. To back up their ideas, vendors are still bringing out the same case studies that used to justify giant data warehouses. And the analysts all look a bit shaken, because the barbarians have entered their walled city.
At a recent big data lunch in London, a vendor showed us what it called a taxonomy: a useful classification of different kinds of big data company.
“Oh dear,” said the analyst – the tame analyst the vendor had brought in to lend some academic weight.”That isn’t really a taxonomy at all. I came to this field from linguistics, and I’m afraid a taxonomy is something quite different.” This analyst lapsed into silence soon after, as the discussion ranged onto subjects that were too practical for the kind of discussions that used to hold sway in the old analytics world.
The issue is important in the way big data is marketed. Before Autonomy was bought by HP, it was in the old analytics world. It knew all its big customers, and offered something very costly and special to them. But that market was about to open up, so selling the company to BP with its big sales force made sense.
However, since then, Autonomy hasn’t made much of a splash at HP – and Autonomy CEO Mike Lynch has left the company.
The thing is, even if the underlying concepts are the same, it is being sold to a new market, and that requires people who understand what is under the covers of the technology – but also understand what the new market is going to be asking.
That takes longer than just putting a new name on the box.