Just like any other predictions or forecasting – on a broader sense, pharma sales forecasts are also a tough and tedious task. Availability of many sophisticated state of the art digital software tools and techniques, notwithstanding.
In an article, published in the July-August 2007 issue of Harvard Business Review (HBR), Paul Saffo – a forecaster based in Silicon Valley, California – expressed this point succinctly with a nice example. He said: “Prediction is possible only in a world in which events are preordained and no amount of action in the present can influence future outcomes. That world is the stuff of myth and superstition. The one we inhabit is quite different—little is certain, nothing is preordained, and what we do in the present affects how events unfold, often in significant, unexpected ways.”
At this point, I would respectfully prefer to slightly alter the last sentence of the quote as, “….and what we and (others) do in the present affects how events unfold, often in significant, unexpected ways.” This is important to me, as we may have control over what ‘we do’, but may not have much control over what ‘others do’ in the present, which may also greatly affect how events unfold, often in significant, unexpected ways.
However, the author distinctly differentiates predictions from forecasts by clarifying that prediction is concerned with future certainty, whereas forecasting looks at how hidden currents in the present, signal possible changes in direction for companies. Thus, unlike a prediction, a forecast must have a logic to it and the forecasters must be able to articulate and defend that logic.
My own hands-on experience in the domestic, as well as the pharma industry of the western world tells me that the actual sales and profit may seldom be a replica of the respective forecasts for the same. However. a reasonably good forecast is the one that is much closer to reality.
That said, it is important to note in the same context, what the above HBR paper has said, in this regard. The author underscores whether a specific forecast actually turns out to be accurate is only part of the picture. Citing a nice simile it says, even a broken clock is right twice a day. Thus, the forecaster’s one of the key tasks is to map uncertainty where our actions in the present influence the future. Uncertainty is an opportunity, he articulates.
In this article, I shall try to explore the possible reasons why, despite the availability of so many sophisticated digital software tools and techniques, the reality in most cases is much different. In a significant number of occasions, the actual sales is much less than the sales forecasts.
The criticality of forecast accuracy:
As we know, sales forecasts today are generally data pooling or consensus forecasts for better buying-in by the implementer, as there exists a critical need, just not to deliver closer to the forecasts, but to exceed the same, especially for the new products.
One will get the flavor of criticality of sales forecast accuracy from the McKinsey research study titled, “The Secret of Successful Drug Launches”, published in March 2014. It found that two-thirds of the sample group of drug-launches failed meeting pre-launch sales forecasts in their very first year on the market. The sample for this study comprised 210 new drugs launched between 2003 and 2009, for which McKinsey gathered necessary consensus-forecasts data for launch from EvaluatePharma. Three important findings of this EvaluatePharma – McKinsey analysis may be summed up, as follows:
1. Actual sales during the first year of launch as % of sales forecast one year before launch:
- % of launches below forecasts: 66
- % of launches on or near forecasts: 8
- % of launches exceeded the forecast: 26
2. Of launches that exceeded the forecasts in the year 1:
- 65% continued to do so in the year 2
- 53% of those exceeded forecasts in the year 3
3. Of launches that lagged forecasts in the year 1:
- 78%continued to do so in the year 2
- 70% of those lagged forecasts in the year 3
In an eloquent way, this study highlights the benefits of sales forecast accuracy for a sustainable performance excellence, especially with new products.
Wide room for improvement in forecasts:
Although, my focus in this article will be on sales revenue forecasts, there is a wide room for improvement in other related forecasts, as well.
Another interesting article titled, “Outsmart Your Own Biases”, appeared in the May 2015 issue of the Harvard Business Review revealed, when researchers asked hundreds of chief financial officers from a variety of industries to forecast yearly returns for the S&P 500 over a nine-year horizon, their 80% ranges were right only one-third of the time. The authors considered it as a terribly low rate of accuracy for a group of executives with presumably vast knowledge of the economy of the United States.
The study further indicated that projections are even further off the mark when people assess their own plans, partly because their desire to succeed skews their interpretation of the data.
Such a scenario prompts the need of yet greater application of a mix of creative and analytical minds to ferret out the reasons behind general inaccuracy in forecasting, which incidentally does not mean setting out an easy target, and then exceeding it. Right sales forecasting with high accuracy, is expected to make use of every potential future opportunity in the best possible way to achieve continuous excellence in performance.
Analyzing outliers in consensus forecasting:
A recent paper deliberated on this area backed by some relevant case studies to capture the relevance of outliers in consensus-forecasting for the pharma companies.
The 2017 study of EvaluatePharma, titled “The Value of Outliers in Consensus Forecasting” flagged some important points. It also asked, are we questioning the level of agreement or disagreement, while leveraging each estimate for consensus forecasts?
However, in this article, I shall highlight only on the relevance of outliers in pharma sales forecasting, and keep aside the question on the level of agreement or disagreement while leveraging each estimate for consensus forecasts, for another discussion.
As many of us have experienced, there will always be outliers in most of the consensus forecasting process, which are usually removed while arriving at the final numbers. Nonetheless, this article brings on to the table the importance of outliers who, on the contrary, can provide an insightful view, especially in those areas with more upside potential and downside risk.
Just to recapitulate, an outlier is a data point that lies at an abnormal distance from other data points, which in this case is data related to consensus-forecast. This divergence can be either very high or very low. Which is why, outlier removal is a common practice, as it is considered as bad data by many. Nonetheless, before singling out and elimination of outliers, it will be a good idea to analyze and understand the exact reasons behind the same.
The above paper also indicates that combining consensus forecasts with the analysis of outliers will enable the pharma companies:
- To better balancing risk and upside
- Improving accuracy of new product selection
Just as in any business, for pharmaceuticals too, sales forecasting holds a crucial importance, having a far-reaching impact. This is primarily because, many critical decisions are taken based on sales forecasts, such as internal revenue and capital budgeting, financial planning, deployment of sales, marketing and other operational resources, including supply chain, to name a few. All these, individually and collectively, necessitate that sales forecasts, especially for new products, should be of high accuracy.
One of the recent trends in this area, is pooling or consensus forecasts, though, it is not free from some criticism. The recent EvaluatePharma study, as quoted above, clearly demonstrates that this approach helps increase forecast accuracy, especially in situations with a high degree of uncertainty.
The upper and lower bounds of consensus known as outliers, may often identify potential upside or downside events that could significantly affect the outlook of a pharmaceutical company.
With this perspective, it now clearly emerges that in-depth analysis of outliers is of high relevance to improve accuracy of pharma sales forecasts, in a significant way.
By: Tapan J. Ray
Disclaimer: The views/opinions expressed in this article are entirely my own, written in my individual and personal capacity. I do not represent any other person or organization for this opinion.