New Product Forecasting in the Pharmaceutical Industry
Foldes, George, The Journal of Business Forecasting
As forecasters, we all make our careers from connecting a series of dots with a line and then project it into the future. We do all of this under the assumption that, if all else remains the same, historical trends are predictive of what is to come. In other words, if nothing changes in the environment, why should the current trends change? Using the latest and greatest statistical packages, whether Excel based or full enterprise solutions, we draw the straight lines, calculate the statistical variances and probabilities, and go through complicated processes to incorporate seasonality and other variables. We do this day in and day out, and are quite comfortable in our abilities. However, when it comes to new product forecasting, our confidence is shaken. We do not have those series of dots to connect, no past history of seasonality and other variances on which to base our projections. This is what intimidates us. We have to realize that all these make preparing a forecast a little bit more difficult, if not impossible. We need to use a whole new series of our competencies.
The ideal new product is a new invention that has never before been seen and that satisfies a need never before fulfilled. As one might expect, this ideal is a rarity. Of course, what we tend to see much more often are new products as line extensions, product improvements, and new uses of an existing product aimed at a new target authence. Whichever format the new product takes, the assumption for this discussion is that the product is a single SKU; it is new to the company / industry, and it meets some form of unfulfilled need.
GATHER PERTINENT INFORMATION
As with any new task, the first step in forecasting a new product requires preparation. This is a key step. Missing this will result in disaster almost every time. Preparation means mobilizing all our resources by bringing together those guardians of information, our subject matter experts. We do this because we have to find out everything we need to know about our new product. Our product development gurus can tell us all the bells and whistles they built into the product. The market research folks can tell us how this product compares to what is currently out there and/or will be out there at the time of launch or shortly thereafter. They should be able to tell us who our customers are, the size of the customer base (epidemiology), what they think of bells and whistles built in by product development, what is the likelihood that they will buy, and what is the current size of the market we are launching in. Our marketing department (the smoke and mirrors team) can tell us their promotional plans, the unique selling proposition that will differentiate our product from the rest, who they are targeting, and how all this compares to what the competition is doing. The logistics people can give you a wealth of information on batch sizes, manufacturing lead times, and distribution channels; the sales team needs to provide us with their sales plans. This includes initial sell in objectives, samples, and other promotional materials they are planning to give away, discounts and deal campaigns they have planned, etc. (By the way this is usually different from plans the marketing folks make. Believe it or not, sales and marketing teams, the company's experts in communication, are usually not very good at communicating with each other). And last but not least, you also need to know when the product has been planned for a launch.
This panel of experts will be very helpful in determining the maximum potential, that is, the peak sales for the new product. (This is where the Consensus method or the Delphi method will be very helpful.) These experts will offer a different perspective and with different biases, so be prepared for a wide range of opinions. However, by assembling different points of view on peak sales potential you will be able to define the upper and lower limits, and make a value judgment on what is realistic. …