Freescalin’ at the Gartner BI Summit

This week finds me at the Gartner BI Summit in London on Monday and Tuesday; in Helsinki on Wednesday and Thursday morning – and then back in a forecast cold, grey London on Thursday afternoon and Friday. Memo-to-self: next January, go and see some of Teradata’s Mediterranean customers!

Joking aside, attending the Gartner BI Summit is always a great way to start the New Year; aside from the wit-and-wisdom of the assembled Gartner analysts it represents a great opportunity to network with 800+ delegates from a variety of diverse organizations across the region and to get their feedback and perspective on where the industry is headed.

And Teradata had much to celebrate at the event this year: there was welcome confirmation that we continue to lead the industry with the publication of the new Data Warehouse DBMS magic quadrant; our effervescent CTO, Stephen Brobst, and his co-presenter Bobby Ghoshal (Freescale Semiconductor) presented a brilliant case study; and last-but-not-least there was the selection of a Teradata customer – Komerční Bank – as the winner of the Gartner BI Excellence award, with another Teradata customer – JD Williams – as runner-up.

Phew! Not bad …

Phew! Not bad for two days work… So long as my boss doesn’t expect this level of achievement for the rest of the year…

One of the “hot” topics gaining attention at the moment and that also featured at the Gartner event is Web Analytics and the use of Social Network data for sentiment analysis. Possibly coincidentally, this was also the subject of a special report in this week’s edition of The Economist.

“Traditional” web analytics – counting sessions and click-through rates – has typically been the preserve of the IT department and has often had more to do with capacity planning than with improving customer experience. All of which is changing, and changing fast.

Integrated web analytics is about regarding the web as just another channel – capturing and integrating the data from the organization’s web-site with the rest of the organization’s data. If that sounds obvious, then that’s probably because it should have been that way all along. But it still isn’t routine practice in very many organizations, with the web data locked in “data jail” in weblogs that can be accessed only by a select few.

All of which means that there are an awful lot of e-tailers out there making some really bad decisions. For example, at Teradata’s recent EMEA kick-off meeting, one of the Teradata Account Directors that has been working with an online retailer provided some fascinating insights.

The organization in question had started down this road to reduce the number of baskets abandoned on its online shopping channel. The logic went something like this: if we can understand why customers are abandoning baskets before checkout, well then we can incentivize them not to do so with a promotion or an additional discount.

Except that when the organization started identifying online shoppers uniquely and analyzing this data on Teradata, they realized that the abandonment rate was far lower than they had thought. Many of the “abandoned” baskets weren’t actually abandoned at all – purchase of the very same item was just being deferred until later.

Even more interestingly, when the e-tailer brought this data into the full production data warehouse (running on Teradata, naturally) and integrated it with the existing customer and transaction data, they found that another huge chunk of these “abandoned” baskets were actually being fulfilled via the call centre. If the organization had gone with its original hunch of throwing money at customers in the form of additional discounts as soon as it appeared that they weren’t going to complete a particular online transaction, they would actually have been giving away margin to consumers who were always planning to come back and purchase, quite likely via a channel – the call centre – with a far higher cost-to-serve. Many of these consumers didn’t need a financial incentive to complete their purchases; they just needed help and encouragement to use the organization’s web-site. It’s (yet another) great example of the multiplicative value of bringing data from across the enterprise together in one place.

Use of Internet data and Social Networking data for sentiment analysis is even more exciting. As those of you who have seen Stephen’s Active Customer Management presentation or read this week’s Economist will know, recommendations from friends – positive and negative – are incredibly powerful in influencing our buying decisions, which makes understanding them very, very valuable indeed. Harvesting data from the Internet and the Blogosphere in this way represents the bleeding edge of modern BI; but in our globalized economy, today’s bleeding edge is often tomorrow’s widely-adopted best practice. And as I blogged here last year, Apple, for example, already knows exactly which features customers want added to its products – and how much they are prepared to pay for these enhancements.

Know this, Apple’s Analytical Wizards: I want an iPad – but I don’t know that I need one. And I do know that my wife says I can’t afford one!

Martin Willcox
Director of Product & Solutions Marketing (EMEA)
Teradata Corporation