The 6 principles of good pharma forecasting by Gary Johnson, CELforPharma faculty member

Well, I don’t think there is one secret, but I think you can probably boil it down to six key principles. And the six key principles of good forecasting that tell you whether you’re producing a good forecast or whether you’re consuming a good forecast are first of all that you have to break your problem down. Generally speaking, there is no magic trick that can take you from zero to an accurate forecast in one giant leap, you have to break your problem down into simpler subproblems, address them individually and then put them back together and that tends to produce better forecasts. The second thing is as forecasters we have to be conservative. Which means something very special in forecasting. By conservative I mean that the bigger the change, the slower it tends to happen. So, think of a relatively trivial change: changing from a branded version of a medicine to a generic version which clinically is a very small change, generics tend on average to get adopted very, very quickly; small change happens very quickly. Think of a much bigger change: a revolutionary new type of therapy, gene therapy or whatever, you’ll find that that is adopted by the market much more slowly. Big change happens more slowly. Keep things simple. There is no evidence in any area of forecasting that making things complex produces more accurate forecasts. In fact, it tends to produce less accurate forecasts. Because one of the keys to good forecasting is the fourth principle which is that you want to look at the problem from different angles. And the more uncertain you are, the more you need to look at the problem in different ways. If you make things complex, you don’t have the time, the energy, the money, the resource to look at things in different ways. The fifth principle would be that we use science scientifically. So, I think some people think that scientific forecasting is about using complicated mathematics but scientific forecasting, well, what is a scientific theory? A scientific theory is a theory that can make predictions and you can test those predictions. And a good forecasting technique has been proven to make good predictions. It has been proven to predict the unknown future. That is what a good forecasting technique does. And lastly we have to use techniques that are proven to work in the real world, you would be amazed how many forecasting techniques have been predicted to work on MBA students and don’t work very well at all in the real world and so, we look at techniques that actually work in the real world. And if you apply those six techniques you will tend, on average, to produce better forecasts and if you don’t, you’ll tend on average to produce worse forecasts. Of course, there are no guarantees, you can follow really bad practice and get lucky or you can follow best practice and you can be unlucky but generally speaking on average you will produce consistently better forecasts if you follow those six principles.