Monday, January 14, 2008

Complex business problems need diagnosis, not packaged solutions

Dave Snowden, whose posts are always interesting and instructive, says this in a post today:

What I am finding is that the more accurately you can describe the situation, the less you need formal intervention methods. For example if I can show a statistically valid trend, supported by narrative then most people in leadership or management positions can work out what they need to do.

In other words: if you fully understand the problem, you don't need complex managerial methodologies to solve it. Over-relying on "five practices" or "seven habits" or "four steps" amounts to short-cutting the real difficulty of understanding complex situations. There are lots of reasons, which Dave describes well in his post, why best practices and case studies are not good guidance for action in most circumstances.

And it is difficult to know what the problem is in a complex environment. Standard assessments & surveys demonstrate this. Asking a thousand employees, "How innovative are we on a scale of one to five?" produces, at best, a pretty-looking chart that signifies nothing. At worst, it can point you in a completely-wrong direction. But everyone wants a short cut.

Exhibit A: the current romance with performance dashboards. In my experience, at best dashboards point you to a situation that you need to understand more deeply ("Is this drop in region 5 sales an anomaly, or is there something substantial behind it?"). In no circumstances is dashboard information enough to act on.

So the need is more for problem-understanding skills, and less for problem-solving skills--meaning managers will need to get more in tune with narrative sense-making. Here are two examples where intelligent executives looked beyond the data, into the narrative mess behind a problem or dilemma, and used the story that emerged to guide their actions:

1) A.G. Lafley of Procter and Gamble, as described in the book "The Opposable Mind":

While trying to decide to whether to roll out highly-concentrated laundry soap, Lafley faced a dilemma. Merchants loved the idea, but consumers, as measured by P&G's quantitative research, were neutral to negative. In most cases, this would answer the question: don't roll out a product that consumers didn't love. But Lafley didn't accept the data at face value:

...[He] decided to dive into the voluntary comments that some of the consumers added to their quantitative research forms.... Lafley took many evening and weekend hours to pore over more than four hundred handwritten voluntaries. He came to the conclusion that while consumers weren't wildly enthusiastic about compact detergents, few were actively hostile....

And, as a result, Lafley decided that the product could be a success. And, of course, it was. (Stay tuned for a review of "The Opposable Mind" later this week.)

2) A Fortune 500 VP of Human Resources, evaluating a new performance appraisal system, as described in this blog:

Last year we had a pilot of a new performance management system for our employees. The trial group was 4000 people. We had spent a lot of time on the pilot and gathered a lot of data. At the end of the trial, the VP of Human Resources printed out all the comments that had been received on the survey forms. He took them home one night and read every single one. Then he came in the next day and said, "We can't roll this system out." And that was it. The trial was very expensive. We'd gathered lots of data, lots of numbers, but the final determinant was what he read in those comments.

One observation about the above two incidents is that reading narratives and evaluating qualitative data, if done at all, are relegated to nights and weekends.

This is perhaps a sign of some of the difficulties we face in the corporate world--you can spend all day looking at charts and spreadsheets, but read stories on your own time!