Optimal evaluation methodology: sales schedules using scenario planning
This is the 3d of a serial of 5 articles dedicated to look at valuations with respect of their contribution to company governance. In the process we propose to take the view point of the practitioner: e.g. what works, is close to the field and is easy to implement as opposed to what is mathematically fancy.
“ Risk comes from not knowing what you are doing ”
Warren Buffet
The first article of the serial focused on the use of sensitivity analyses and tornado chart as a major governance information for decision makers. The argument being that a number is a decision by itself: an NPV is either positive or negative, whereas decision makers manage for the future and must identify what are the key drivers of the decision to be made, must be aware of their impact, and must understand how to keep abreast with their evolution, so that future decisions are well documented a priori.
In the second article we have looked at the real option methodology for the estimate of future peak sales, a variable that can most influence the results of an evaluation. We have listed the reasons why it represents a significant progress over NPV models in thinking about uncertainty and incorporating it in development models.
This third article is dedicated to scenario planning methodology as the preferred tool for the estimating of peak sales. We shall provide a background analysis, how-to support, an example and a comparison of results with option theory.
1. overview
“…One of the most difficult things to do in the [pharmaceutical] industry is to predict peak sales of a product. The classic example is Lipitor, for which the original forecasts were for peak annual sales of $800 million, but which proved to be 16X higher.
Recently, the company EvaluatePharma predicted that the world’s biggest-selling drug in 2018 will be Merck’s Januvia for type 2 diabetes with annual sales of almost $10 billion. No one would have ever predicted this when Januvia was launched.” (1)
2. The case for scenario planning
From our experience scenario planning is in our eyes a practical and accurate proxy for estimating probable future outcomes. Scenario planning forces managers to think their development and launch strategies in competitive terms, and to document their hypotheses:
It allows to take into consideration the flexibility of management in deciding how to pursue or abandon the project as it proceeds through development.
It is close to the battle field as it takes into account the competitive profile of the compound in development at the time it will be launched
It takes into account the volatility of estimates pertaining to the future 10 years down the road
It allows for ongoing updates in the hypotheses as they are collected as the project progresses
There is no need to include complexity, rather focus on essential value drivers.
3. Sales drivers
When attempting to develop estimates for an essential parameter, it is essential to corner its own generic drivers. Sales depend on a few generic factors:
What is the size of the potential targeted market
What is the competitive profile of the product to be introduced
Has management identified the threshold of the marketing and sales budget
Is management willing to invest
We would argue that any proposed methodology has to deal with these dimensions and establish documented hypotheses for each of them.
4. Expected competitive profiles
When developing scenario for sales potential and estimates, we recommend to start with estimating the probable outcomes of project development along the following dimensions:
We recommend to use the dimensions and categories that your company is most familiar with. If your management has not yet performed such systematic analysis then it is an opportunity for you to help your boss in structuring the common wisdom.
Early in the development process you might document such a table for each competitive scenario chosen. Even though careful attention is brought to the selection of the compounds entering clinicals (probably the most difficult decision in the R&D department, and the one with greatest impact on the company’s future, let’s cheer to those daring) at that point in time the development staff has had little time to extract the relevant clinical and safety information from the compound.
Once you are done you ideally come up with 3 to 4 scenarios. Say that in the example taken in the previous article, we estimate the following values:
This might be perceived as a heck of an administrative effort. If it is, then your management has missed the point and you might want to revisit how you come up with valid non filtered information, internal sources, external sources?
5. Assessing probabilities
The next step consists in assigning a probability to each of the scenarii: based on what we know, what is the likelihood that scenario 1 materializes (the blockbuster with a superb safety profile in a new fast growing therapeutic area with a huge patient base, that is going to enter the market as #1, and benefit from a humongous launch budget), what is the likelihood for scenario 2 and etc.
This might seem complicated, but it is a rather simple process if one keeps politics out of the picture. Methods exist such as Delphi, round tables, etc to get the best estimates from a goup while avoiding heated debates.
6. Assessing ranges
Next we need to define how confident we are with our estimates. This is where our methodologies borrow from the real option theory, by assessing what is the range for each of the key variable of the scenario.
For example, in the best case scenario the point estimate of targetable market is € 1100. Your next effort is to assume by how much you can be off in estimating this value, and between which limits it could fluctuate. The experience from the contractors’ industry is here of help: a good initial guess is that most point estimates at the beginning of the project are within a +/- 25% range. If you prefer 20% take 20%.
7. Summarizing hypotheses
Finally after a day extra-muros your team comes up with the following estimates:
You already spot that they have a clear idea of what the compound and the development team can deliver.
By processing systematically, recognizing that they might be off, but accounting for that by assessing probabilities and developing ranges, the team has developed an extraordinary body of knowledge on potential outcomes, which will later be of essential value for the board to come up with strategic decisions.
8. Estimates of market share
The same process can be repeated to estimate the potential market share that the company could obtain, based on the competitive attributes of the targeted market, the retained competitive profile of the product at launch time, and the amount of launch budget that the company would be granting.
You should not try to make sense of these numbers based on your own experience: they relate to people, company and opportunities that you are not familiar with. Just acknowledge that a group of wise people have come up with a great strategic information for their management.
Professional and affordable programs exist that will transform these assumptions in distribution curves, as shown in the chart on the left.
9. Impact of selected hypothesis and scenarios
The following table shows the estimated peak sales for each scenario.
The plot of peak sales estimates for each scenario shows that by applying the method carefully, unexpected outcomes might surface: in our example, the highest peak sales are not correlated with the largest targeted market. By systematically looking at each parameter of the sales estimates we have derived a powerful decision making model, closely related to field opportunities.
Another point highlighted by the plot of peak sales estimates for each scenario is that when planning sales estimates it is not wise to use averages of scenarios values: by averaging values one loses the enormous amount of strategic information embedded in each scenario, and the average number, though more aesthetic in a way, does not relate to any reality.
10. Comparing NPV, Real Options and Scenario planning
The graph on the right shows the peak sales estimates generated using the real options formula applied to the distribution of the averaged peak sales.
Although mathematically correct we believe that these predictions of sales estimates have less strategic usefulness as those provided by scenario planning as seen above
11. Next steps
In the next 2 articles we shall deal with the issue of discount rate -which discount rate to use and why?- and with the issue of how to incorporate the various scenarii into a global development model that will be close to the market, and will allow for management to exert its options as information is gathered along the development path.
Note that the methodologies that are delineated here for a bio-pharmaceutical development and launch are easily transferable to other kind of industries and types of projects, including M&As, plant expansion, licensing and the like.
He is lecturer of Entrepreneurship at the Grenoble Management School in France. He is founder of Foro Ventures, a company dedicated to provide assistance and interim management for top-line growth projects and turn-arounds. Jean-Louis has experience with B2B industries ad high technology sectors.
