The big data leap of faith

In a recent blog I talked about eight disruptive technologies in supply chain, and although we’re still a few weeks from closing our 2014 Chief Supply Chain Officer Survey, it looks like we have a winner!

Big data (based on the 660 surveys completed so far) is well ahead with nearly two-thirds of all respondents saying that it is “disruptive and important” against only 6% who deem it irrelevant. Next best is digital supply chain with substantially fewer (48%) agreeing that it is important.

What’s really there?

I had a breakfast meeting recently with a hi-tech CSCO whose company has every reason to trumpet the killer app that is big data analytics. Rather than regale me with tales of whopping ROI from big data initiatives, however, he asked very candidly whether we at SCM World had seen much of anything going on. The conversation was telling – neither of us could string together a meaningful set of examples, despite the overwhelming sense that this tech trend is a winner.

Undeterred, I had one of our interns scour the web for public examples. Apart from Tesco’s use of predictive analytics to tune inventory according to weather, promotions and other demand influences at the store level and a pretty compelling video interview with the CIO of UPS on how its data centre operations support a whole slew of detailed tactical decisions, most were not really what I think of as “big data analytics”.

Dan Gilmore of Supply Chain Digest had a good blog on this topic recently, which says in essence that big data analytics sounds good, but is still in its formative stage. I agree. Having discussed this first hand with everyone from Google to ABB, I find myself taken back to the salad days of e-commerce around 1999. You know it’s real, even if you can’t put your finger on it.

What’s really there is an obvious convergence of capability and opportunity that no sane supply chain strategist can afford to ignore.

Big data analytics for supply chain, broken down:

• Big – Six billion people, most equipped with mobile devices plus cameras, audio recording devices, geo-location technology and, of course, billions of fixed nodes creating, requesting and re-mixing digitised information, is most certainly big. Since this human mass is the ultimate aggregate demand data set, finding ways to mine it makes a tonne of sense. Amazon’s predictive shipping is the tip of the iceberg.
• Data – Internet of things adds another 20-40 billion information generating and consuming entities to the mix. Supply chain people see this as the third most important disruptive technology on our list. The important thing here is that, unlike human data which thanks to free will is inherently fickle, internet of things data is reliable. Deterministic modelling of supply chains based on temperature, distance, weight and so on offers considerably more precision than anything based on consumer data.
• Analytics – Processing power continues its march onward and upward, meaning mega-computing power like that wielded by IBM’s Watson can actually handle simulations and queries that were previously impractical or even impossible. At the same time, accelerated learning among mathematicians, operations research specialists and systems engineers means companies are getting better at framing business questions as algorithms.

The power of simplicity

Big data analytics has all the ingredients to revolutionise supply chain, except one: the right questions to ask.

Perhaps the déjà vu of e-commerce 15 years ago when companies budgeted a few million dollars for exploration offers a guide. Everyone knew it was big, but most were ready to waste money short term because they didn’t know what would stick. The first thing that worked then was the simplest – online auctions as a way to save money.

The next wave was selling online, including B2C basics like books plus B2B baby steps like GE Plastics offering tech support to resin buyers. Again, simple applications begat complex ones and the world was changed forever.

Big data is in the same place. Experimentation is better than sitting on the sidelines waiting for others to take the bleeding edge. Clear ROI is unlikely for a year or two, but abstinence will definitely not pay off.

It’s probably enough to start simple and ask: how do we sell one more widget?