Sounds like a conundrum from a corporate cracker, don’t you think?
In fact, this question was put to me by a potential customer just last week. And not for the first time, either. Sometimes ‘Big Data’ is replaced by the latest technological acronym – that week’s panacea for all corporate ills – but the sentiment’s the same. So, what did I reply?
”How can I get extra value from my data by using more and more analytics?”
“What?” I could see they were impressed.
“What’s going to spark a new business culture – a data culture where individuals challenge current thinking and drive competitive advantage through access to more accurate commercial insights?”
“Come again?” Hmm.
“What do you think Big Data is anyway?” I asked, finally. “It’s not like a massive cluster of the latest technology, built, just for the sake of it. Anyway, a big bit-bucket and some fancy analytic technology won’t improve your business a jot.”
The fact is, a technological solution should be based around the business case, not the other way around. I know from personal experience.
An Asian Telecom customer of mine wanted to segment their users based on mobile browsing habits, intending to feed the information into their campaign management systems. To do this, they wanted to build a near-real-time web crawler on a Hadoop cluster, which would download pages that had just been browsed by the user.
On the surface, this looks like a good fit because the original use case for Hadoop was a scalable web crawler. But what sort of Hadoop system could build a web crawler capable of downloading several hundred thousand pages which are, simultaneously, being browsed by 7-8 million subscribers?
A basic knowledge of networks would flag-up the fact that the data is actually flowing through their own pipes while subscribers are browsing the web. And by using deep-packet inspection, they can start storing the content without building a separate web crawler. In the end, they still might need a Hadoop cluster to store the multi-structured data acquired through the deep-packet inspection interface. But the project’s cost would be reduced significantly, while delivering the same value to the enterprise.
Leading analyst firm, Gartner, coined the term Logical Data Warehouse to describe the evolution of information architecture from a single technology, to multiple architectural constructs based on the type of analytics, the cost of storage, and high-performance concurrent access. Clearly, future analytics will exist on several different technologies. So the technological toolset you need depends on what you’d like to do with Big Data (and your business), bearing in mind the pragmatist’s equation: ‘old business process + expensive new technology = expensive old business process’.
The fact is, smart operators don’t employ analytics simply to gain insight. They operationalize that insight to improve the way business is done.
So, where were we? Oh yes, if the answer is Big Data, I reckon the smart-alec question would be “What’s my most valuable commercial asset?”
– This article was originally published on forbes