Economist this week on algorithms and business

This week's The Economist (Sept 15 issue) has a good story on algorithms in business, with examples for UPS to Microsoft and very interesting data for example from British supermarket giant Tesco. Tesco loyalty card apparently has 13 million members and tracks their usage patterns across 55,000 product lines. Makes for a massive database ha-ha, and that is where algorithms come in to help find the needles of customer insights amongst the haystack of repetitive near-identical purchasing data. The resulting improvements in Tesco segmentation result in re-arranging which items are paired next to each other on store shelves etc.

The article does mention telecoms data briefly but doesn't point out that by far the biggest amount of customer data - by more than an order of magnitude more than say banking, retail, credit cards etc - is consumer data collected by mobile telecoms operators. So for all the complexity - and the ability to drive super-power segmentation beyond the wildest dreams of the marketing professors in academia - is with mobile operators today.

Yet only a handful of the mobile operators have fully commercialized such tools as measuring social networking dimensions of their customer base. These yield wholly new classes of customers such as the Alpha Users we talk about at length in our book Communities Dominate Brands with Alan Moore - and both of us often discuss in our lectures worldwide. But even those - Alpha Users are the influencers of any community - are only the pale beginning to what truly is a revolution in customer understanding. One of the hottest concepts is Best Next Offer - used by a couple of the leading operators around the world. It uses user data and their contextual data such as social networking scoring - to give customized recommendations.

Just like Amazon rates books based on similar purchases, except that being far more complete in profiling due to vastly more points of contact, very precise targeting can be done with mobile telecoms profiles. Now for example in Finland if you call up the calling center, you get a personalized offer (other customers who behave remarkably similar to you, have liked this new service) where your offer is totally different from the offer made to your mother, your sister, your son or grandchild, etc. Two brothers might get totally different offers simply based on their unique behaviour.

Oh, and let me mention, this is not typical mobile telecoms segmentation where an operator/carrier might manage a dozen or a few dozen segments. No, this is just like Tesco and the airline industry and automobile manufacturers etc, whose segmentation runs 10,000 segments and beyond. THAT level of granularity. No human can possibly comprehend that level of complexity, and the only way to manage it is with technology - and algorithms, ha-ha.

This is the future of marketing. Not estimating what we might want based on surveys and expert consultant views etc. But rather the true behaviour of customers on a perfect profile of what we've been doing in the past (and with whom..)

Note also - there is nothing wrong with this, customers LOVE it. All mobile operators try to sell us something if we happen to step into their stores or call up their calling centres - and so did Finnish operators too, before Next Best Offer. But back then it was the flavour-of-the-week kind of sales, which service did management want us to push this week or what service had the best commission bonus, etc. Now the calling center identifies us while we wait in the queue, and by the time the customer service rep starts the dialogue with us, they have the option of the three most likely services I might really like (and obviously also that I don't already use) - and the service rep who doesn't necessarily even know that given service at all, sees the sales script on the screen, so the service rep can tell me about what that service is and what it would do for me, etc.

But yes, it comes down to algorithms. A great company in this space is our friends Xtract from Finland (Alan's done some big seminars with Xtract and I've done some workshops with them) who started off doing the immensely complex mobile telecoms algorithms. Now they have branched out to "easier" areas as well, such as media, banking, credit cards, retail etc, without forgetting their telecoms roots of course.

Think Amazon recommendations. But that for all walks of life for us. Fascinating.