Portfolio analytics core to pharma portfolio decision making
Portfolio analytics core to pharma portfolio decision making
The pharma industry has experienced productivity decline for decades, measured by many metrics including return on R&D investment and the number of New Molecular Entity approved per billion dollars spent. Maximizing value creation from pharma R&D investment, therefore, has been a focal point for the industry, and recent data explosion as well as advancement in analytical tools holds excellent promise in leveraging pharma analytics to maximize business impact.
A lot is at stake when it comes to pharma portfolio decision making, whether on funding a project to the next phase, taking on a program with multiple indications, seeking to partner for financial or capabilities reasons, or filling a portfolio gap through acquisition or in-licensing. Decisions are often tied with significant investment up to billions of dollars and have both immediate and far-reaching impact within the organization. In the broader context of declining pharma productivity, it is ever more critical to make informed portfolio decisions anchored in facts and analytics to more effectively allocate financial and people resources.
Portfolio analytics is at the heart of pharma analytics, by aggregating the most critical data points across the organization for pharma assets (pipeline projects and inline products), from research to development, commercial, Manufacturing, and other supporting functions (e.g., Forecasting, Competitive Intelligence), to inform on portfolio decision making and maximize value of pharma investment.
Central to portfolio analytics is the “value triangle,” composed of opportunities, risks, and uncertainties for projects, programs, and portfolios. Key inputs across projects (including the probability of success, projected sales forecasts, projected R&D if any, marketing, SG&A, manufacturing costs) are first aggregated, validated, and triangulated in an annual or bi-annual portfolio review process, which could be bottom-up or top-down. Inputs are then fed into a valuation model where various financial metrics, including risk-adjusted net present value, are generated. Finally, critical information and outputs are rolled up to the portfolio level for analysis and decision-making. The more sophisticated analysis could be performed to address specific questions, such as decision analysis to map out options for an asset or portfolio, sensitivity analysis to test select assumptions’ impact on the portfolio, Monte-Carlo analysis, scenario analysis, or business cases for business and development purposes.
A few watch-outs in utilizing portfolio analytics
Most often, conducting analytics is not a challenge. Rather the contrary, it is often too tempting to generate more analytics than what is needed and could be reasonably consumed by pharma executives for decision-making. Translating analytics to tangible business impact requires efforts at many levels. First, it is instrumental in clarifying upfront the strategic and financial context and objectives before scoping and conducting analytics. If one could link portfolio analytics to the most important corporate and strategic portfolio priorities, quickly narrow in on critical questions to address, select a few analytics based on past experiences, or stakeholder interviews, it would already be battle half won before any analytics is conducted.
In addition, any analytics is only as good as the data coming in; therefore, it’s essential to assure quality data and unbiased view. A lot of effort is spent on validating inputs, triangulating assumptions, understanding the sensitivity and uncertainty (ranges) in assumptions and how that affects the output. To give one example, most often, different scenarios (high, base, low) are generated for sales forecasts to inform on the potential range. Similarly, pharma companies frequently calibrate the probability of technical and regulatory success, one of the most critical drivers of project valuation, on an annual or more frequent basis, post extensive expert pressure testing.
In cases of program valuation, interdependencies across projects (in terms of probability of success, commercial potential, costs, etc.) should also be something to be mindful of and may warrant more advanced tools such as decision trees if needed.
Future opportunities in transforming portfolio analytics
As we enter into the digital and significant data era, there are many opportunities to turn pharma portfolio analytics to maximize business impact, and below provides a peek into this opportunity space.
Enhanced analytics visualization: A picture is worth a thousand words. While Powerpoint slides are not yet giving way for its mainstream role in senior leadership communication, we have seen more utilization of data visualization tools such as Tableau in day-to-day portfolio management. These tools are gaining traction thanks to their enhanced visualization capabilities and intuitive use and help foster transparency across the organization around decision input and outputs.
Streamlined analytics workflow: given portfolio analytics integrate many different data types across the organization, combining data collection, valuation modeling, advanced analytics, and analytics visualization into one streamlined process, ideally in one online digital platform, would be a tremendous opportunity. The benefits are multi-fold: it enables more automated and up-to-date portfolio views, provides real-time digital access to stakeholders to enhance the impact of analytics, and increases the agility of decision making as we embrace the inherent risks and uncertainties in drug development and the fast-changing market environment.
Use of AI/Machine Learning: the field has started to see pilot experimentation from both academia and industry around using AI/Machine Learning to better predict key drivers of valuation, such as Probability of Success. Along the value triangle, there are additional drivers where there are plenty of data points to tap into, such as sales or pricing, which could be good use cases for AI applications.