The 5 most useful sales forecasting techniques for pharmaceuticals

By Gary Johnson, expert-trainer of The Pharma Forecasting Course

1. Simple Conjoint-type Models

The best way to link the strength of your product's profile to a market share

Conjoint models work by working out how attractive your product is - relative to the competition - on the key areas of product performance. Extensive academic research has shown that the very simplest conjoint models - so simple that you can produce a do-it-yourself version on an Excel spreadsheet - are as accurate as the most complex and expensive conjoints.

💡 Did you know? The inventor of conjoint also invented the most popular form of perceptual mapping!

2. Zipf's Law

The best way to link entry order to market share

Zipf's Law is a simple law developed by an eccentric Harvard linguist in the 1940's.  It can predict many, many disparate and seemingly unconnected things: the relative size of cities in a country, the size of earthquakes, the popularity of pages on the web and ... also pharma market shares. On average, brands that get to market sooner maintain larger market shares than brands that get to the market later. The average relationship between pharma brand market shares and entry order is best predicted by a simple modification of Zipf's Law.

💡 Did you know? The modified form of Zipf's Law was created by Murray Gell-Mann - the discoverer of the quark subatomic particle!

3. Simple Elasticity Model

The best way to link marketing spend to market share

If you change your marketing spend by 1%, how much will your market share change? The answer to that question is known as the marketing spend 'elasticity'. That one simple number is the best way to link your marketing spend to your market share.

💡 Did you know? The weight of marketing behind a brand has a much bigger impact on its sales than the weight of marketing behind a therapy class has on its sales!

4. The Bass Model

The best way to predict the speed of uptake for a brand or therapy class

The Bass Model is the most famous - and the most studied - marketing model of all time.  It is based around the fact that when people are nervous about really new things they tend to imitate what other people are doing. But when they are confident (or desperate) enough they will adopt the product themselves without waiting to see how other people get on first. The model just needs three numbers to predict the entire sales history of a new product. An extension to the Bass Model - the Law of Contest - can predict the entire history of all the products on a market.

💡 Did you know? The Bass Model is actually mathematically identical to a simple epidemiology model that was used to predict the spread of swine flu!

5. Simple Extrapolation

The best way to forecast stable trends

If you have an established product in a stable market (no new competitors about to launch, no major clinical trial results about to be reported) there is no more accurate way to forecast than to just project the current trend into the future. Large forecasting competitions that have pitted the most sophisticated and complex extrapolation techniques against the simplest and most straightforward have consistently shown that the simple techniques are as good as, if not better than the complex ones.

💡 Did you know? One of the simplest and best forms of extrapolation - exponential smoothing - was invented in the 1940s by a US navy analyst who used it for tracking and firing at submarines!