Analytic truth and myth

Copyright © 2009 James Taylor. Visit the original article at Analytic truth and myth.

Alison Bolen posted a nice list of analytic truths, or perhaps myths, on the SAS blog today and asked what people thought. I was, of course, unable to resist:

1. To make analytics successful, the CEO has to have a personal interest in it. MYTH
   While it is true that the only companies I see who have made it to what Tom Davenport called “analytic competitor” are those that have CEOs who are involved with the analytics, I do not believe that CEO involvement is central to all analytics success. Line-of-business managers and other executives can successfully drive analytic projects; I just don’t think you are going to get company-wide adoption without the CEO.
2. Analytical organizations have to be positioned in a central high-power position. MYTH
   I think that centralized analytics are a consequence of success not a pre-requisite for it. As you get some localized success you will want to bring it together to drive more success but I don’t believe a central group is needed or event desirable to start.
3. Every company in a competitive environment needs analytics to be successful. TRUTH
   As I have …

Copyright © 2009 James Taylor. Visit the original article at Analytic truth and myth.

Alison Bolen posted a nice list of analytic truths, or perhaps myths, on the SAS blog today and asked what people thought. I was, of course, unable to resist:

1. To make analytics successful, the CEO has to have a personal interest in it. MYTH
   While it is true that the only companies I see who have made it to what Tom Davenport called “analytic competitor” are those that have CEOs who are involved with the analytics, I do not believe that CEO involvement is central to all analytics success. Line-of-business managers and other executives can successfully drive analytic projects; I just don’t think you are going to get company-wide adoption without the CEO.
2. Analytical organizations have to be positioned in a central high-power position. MYTH
   I think that centralized analytics are a consequence of success not a pre-requisite for it. As you get some localized success you will want to bring it together to drive more success but I don’t believe a central group is needed or event desirable to start.
3. Every company in a competitive environment needs analytics to be successful. TRUTH
   As I have said before, your data is your one truly defensible competitive edge, and if you are not using it (which takes analytics) then you are stupid foolish incompetent missing out.
4. Analytical expertise can be out-sourced/in-sourced/off-shored. TRUTH BUT…
   While you can and should bring in outside expertise you need to have a basic understanding of the power of analytics in-house. Someone must grasp the potential for analytics and understand the business, even if they cannot develop the models.
5. Getting data and technology in place is a long and cumbersome process. TRUTH
   It also cannot be rushed and should be done incrementally with each stage developing additional capability that is put to work adding value. Don’t build all the data and technology infrastructure before you start delivering value. And start with the decision in mind – build what you need to improve a specific decision.
6. Without data and technology you cannot do analytics. TRUTH and axiomatic
7. Analytics is a thing mainly insiders and experts understand, and vice versa. TRUTH
   And this is a challenge, see #8
8. Communication of analytics is more important than analytical people think. TRUTH with bells on
   This is so true it is hard to over-emphasize. Analytics people are often terrible at this – talking about statistical measures not business measures, over-explaining the approach and under-explaining the consequences, etc., etc. If you can improve the skills of your analytic team in only one area, this is it.
9. Analytics only should do things which have a measurable impact. TRUTH
   And measurable in business terms, not mathematical or statistical ones. Business people don’t care about lift curves, they care about results. Remember this.
10. Analytics mainly is applicable in retail/standardized environments. MYTHish
    Any business that has large numbers of repeatable decisions – operational decisions – can and should be using analytics to improve them. This implies a repeatable environment and one with lots of participants so retail or B2C environments are more common for sure. But companies can have many thousands of partners or locations, parts and suppliers so decisions about these things can be analytically enhanced also even in B2B environments.

Great list. Thanks Alison.

Link to original post