Skip to main content
Live from Teradata Partners - Tom Davenport

I'm at the Teradata Partners conference, just for today, and to have just listened to Tom Davenport (about whom I have blogged before, most recently in this post on decision automation). Tom presented on competing on analytics. Now at Teradata Partners, one is somewhat preaching to the choir when it comes to leveraging analytics. Tom's view is that the planets are aligned for analytics - lots of technology, lots of data, computer-savvy folks, business need for optimized, differentiated processes. While Tom clearly thinks that the folks with Teradata are doing lots right, he differentiated between reporting and analytics as two aspects of Business Intelligence. He premised that reporting is necessary but not sufficient and that analytics delivers more intelligence and more competitive advantage. He proposes that making analytics and fact-based decisioning key to strategy and competition is where the value comes (see this post on strategic alignment)

Tom used some examples of companies to make his point. Marriott is one example that uses analytics aggressively to optimize pricing for revenue management. Harrahs focuses on loyalty to improve profitability and used analytics to transform the business. Tom also used the Red Sox to show that you have to take action based on analytics at the front line by giving the example of the Red Sox manager who failed to apply analytic insight about Pedro Martinez and lost game 7 to the Yankees. Not all companies transform themselves with analytics - Capital One for instance was built by information-based strategies and the use of analytics.

So, if it is possible to compete on analytics, how to do this? Tom has some great ideas and some key requirements.

Technorati Tags: , , , , , , ,

  1. Get CEO to commit
    True of most strategies after all. Tom pointed out that intangibles requires a culture that cares and a culture of learning and testing (e.g. Harrah's CEO says that not using a control group will get you fired!). The senior level support is essential so you must demonstrate value and get the leadership on board. This can be particularly hard in companies that like the "manage by gut reaction".
  2. Widespread use of modeling and prediction
    This means developing the skills (both skills for developing insight and for using it at front line)
  3. Use analytics to support your distinctive capability
    Have a Major and a Minor like Loyalty with Service. Focus on how to compete distinctively - perhaps by using Decision Yield to assess possibilities
  4. Manage analytics enterprise-wide
    Hence the "E" in EDM. This is particularly a problem when there are data-fiefdoms. Not only must data be a corporate asset (as Tom noted), so much decisions based on that data. The focus on a core analytics group is something you find in long time analytic users like banks.
  5. Have big goals

He identified 5 stages of development in the companies he surveyed:

  • Stage 5: Competing on analytics
  • Stage 4: Clear intent, almost there - not passionate yet
  • Stage 3: Vision but long way to go
  • Stage 2: Local, non strategic activity (previously BI best practice)
  • Stage 1: Wrestling with the basics

The trend over the last few years is a big growth in above-average analytic capability, though not much growth in competing on analytics yet. Tom has some data suggesting that high performing companies correlate really highly with analytically sophisticated companies - analytic sophistication tends to drive improved performance. Tom made some suggestions for serious analytic competitors that included automating decision processes, knowledge (rules) management and new measures by mining of data. He also talked about event-based/real-time decisioning and how effective it can be. He ended by talking about how this is a long-term plan and I would add that you can get started incrementally - don't wait for all the data, focus on delivering insight from the data as you go!

One last thing. Tom used Katrina as an example of rules and analytics - the analytics said that there was higher than usual occupation in Houston and so prices should be raised but people (rules) said don't raise prices for refugees. Now his example was of manual rules applied to automated analytics but it still shows the need for both rules and analytics.

Great talk. Tom is always worth listening to. If you believe in competing with analytics, you should think about applying analytics to your operational systems and that will lead you to think about Enterprise Decision Management.

If you read this while at the conference, drop by the Fair Isaac booth (#317) and say hi to the folks working there. I'm speaking at 1:30pm too...

related posts