May 10, 2009

On making forecasts

Making forecasts and predictions are quite fun, and I've been responsible for futurecasting many times on this blog. As he is leaving Merril Lynch, economist David Rosenberg wrote a note with twelve rules for good economic forecasting. For anyone making forecasts, reading all twelve rules is a given as it is a short read, but four of his rules can be used in most organizations and for most decision-making. I've added a short, business-oriented, comment to each rule.

2) Never be a slave to the data - they are no substitute for astute observation of the big picture. When running a business the reality is that few problems are solved by better systems or more data. First you need to have a group of smart people who understand the business and how to find and use data to drive actionable suggestions. Then, but not sooner, it makes sense to spend on better systems. It is amazing how many businesses run on brains, sound judgment and quite simple Excel sheets.

7) Always seek out corroborating evidence. It is very easy to fall in love with your big idea. It is much better to try to disprove it yourself before a plan turns into something that costs a lot of money. So, if one cannot find facts supporting one's idea, it might be a good idea not to move ahead until such facts are found.

8) Have respect for what the markets are telling you. Plans are great, but be ready to adapt if customers (via their behavior, sales people, support etc) start to tell you that reality differs from your plan.

11) Highlight the risks to your forecasts. As no-one knows what will happen in the future, a lot of the usefulness of predictions is understanding what could happen and what that would mean to you. Most of the time, explaining key assumptions and alternative scenarios if assumptions are wrong are more helpful than predicting what specific euro-number sales will hit twelve months out.

Somewhat related, Nothing should become too big to fail

No comments: