22 January 2009

Business Debriefing

Many years ago, I became acquainted with the CEO of a storied software publisher. Bill Goodhew had recently taken over as CEO of Peachtree Software in a management buy-out. While Peachtree had the pedigree of being one of (if not) the first personal computer based accounting software packages, it had never made a profit. Peachtree was sold to Management Science America (MSA), an early mainframe software company, and during their tenure as owners, the division still did not make any money. Bill and his management team took it over, drastically reduced the price of the software, and marketed it predominantly using direct response advertising.

While that seemed very extraordinary at the time, Bill filled me in on his seemingly simple strategy. He selected three different advertising placements, all with coded responses, so he always knew which was producing the results. He continued these for some set period of time and then would pick the one that worked the best and double its frequency. The one that was the worst, he discontinued. And the other, he continued as it was. He also added one more publication and tried that.
Interestingly enough, using this strategy and the much lower price, Peachtree’s owner-management team made money for the first time in the company’s history. Bill made it very clear that he was no marketing genius. In fact, Bill seldom guessed correctly in advance which advertising placement would work. However, after receiving the results, he was in a perfect position to learn and adjust.

The lesson I learned from Bill’s strategy expanded well beyond advertising and marketing. What this taught me was the value of feedback. Learn & Adjust - that idea has lived with me for the several decades since our encounter.

Every business process can benefit from this concept of learn & adjust. None of us is smart enough to predict the actual results of our actions. But assuming we have a clear and available way to listen to the results of our actions, treat them as objective data – free from our personal prejudices, and be willing to act and adjust our behavior based upon this data, we can create learning systems that create impeccable institutional knowledge that can be shared within our organizations.

The US Air Force uses Stealth Debriefing sessions after every sortie. Immediately after a flight, the pilots debrief in an objective and selfless manner. They learn and adjust from these sessions so that they can institutionalize better practices in their very next encounter. As the Air Force puts it, for them it’s a matter of pure survival.

In business we usually survive multiple mistakes. But the best organizations are compulsive about feedback and learning processes in everything they do. I have participated in great sales organizations that use debriefs as a critical part of their process. After each sales call, the sales person and her team, get together to discuss what went right and what went wrong. These sessions enable the team to determine what is working and what is not, enabling them to hone in on their sales pitches in their very next call.

Seems pretty simple? Yet there are all too few organizations that institutionalize this type of behavior. Meetings occur without any summary of the results. Business decisions get made without the benefit of any recorded feedback. Marketing dollars get spent, without any clue as to their effectiveness. In today’s economically stressful times, using this simple (and cheap) process to immediately improve what we do, every time we do it, is long overdue.


  1. Les,

    I really like this approach. One question that comes to mind is how to filter signal from noise. I imagine that when collecting feedback, one can get caught up in listening to too much information and miss the real signal, causing adjustments to be off. It seems like it would make most sense to identify one metric and follow it so as to not get distracted. Thoughts?

  2. This comment has been removed by the author.

  3. Good point Sinclair. Here's what I've used to handle situations where results can be specifically measured:

    1) Create an expected result set
    2) Analyze actual results received vs. expected results
    3) Determine action to be taken
    4) Revise expectations, and
    5) Repeat

    In other situations, simply recording the results (the important parts - not all the details) and distributing them or making them available is also helpful. Most of the time if it isn't written down, it doesn't really exist (and more important people forget).