Industry-wide performance themes
The fact that we see a higher proportion of discontinuations occurring earlier than we used to signals that top pharma is being increasingly selective about which drugs it chooses to progress.
The P3i focuses on eight key indictors, shown in the table below. Of these eight, five are worsening. Cost per new drug has more than doubled, from $49M to $108M. The median timeline from Phase I to approval has lengthened from 8.4 to 9.3 years. The probability of technical and regulatory success (PTRS) fell from 9.6% to 8.2%.
Figure 1: The Pipeline-to-Patient Productivity Index scorecard
Norstella data is uniquely comprehensive, and at industry level the numbers are directional. Disclosure timing, coverage samples and consensus forecasts all carry uncertainty, and samples shrink further at therapy area, indication or company level, so findings are best read as signposts. For competitive intelligence, the value lies in the direction of travel and in the divergence between competitors.
The report also features a ranking of the top pharma companies, based on these indicators. In this analysis Eli Lilly ranks first, and the result reaches well beyond GLP-1s. Novo, Vertex, Daiichi Sankyo, UCB and Roche also score well. Lilly returns below as a case study.
One indicator reads counterintuitively at first glance - we consider more failures occurring at Phase 1 to be an improvement. If you consider that most drugs are not going to make it to approval, meaning they are discontinued along the developmental journey, then the cost of failing at Phase 3 is significantly more expensive than the cost of failing at Phase I. The fact that we see a higher proportion of discontinuations occurring earlier than we used to signals that top pharma is being increasingly selective about which drugs it chooses to progress
Pipeline creation
Top 25 pharma disclosed about 3,200 new drugs from 2016 to 2020 and about 2,200 from 2021 to 2025, a 32% drop. Over the same periods, the estimated preclinical R&D budget rose 47%, from $161bn to $237bn. Estimated spend per disclosed drug moved from $49M to $108M, a 2.2x increase.
Figure 2: R&D costs are inflating
The data raises a question: are companies creating fewer drugs, or disclosing fewer? A higher cost per drug is consistent with greater selectivity in both what companies advance and what they announce publicly. For anyone assessing a competitive set, a thinner early pipeline may reflect higher conviction in fewer assets.
For anyone assessing a competitive set, a thinner early pipeline may reflect higher conviction in fewer assets.
Development timelines
The 2025 benchmark, which reflects timelines over the prior 10 years, shows a median of 9.3 years from Phase I to approval, up from 8.4 years in the 2020 benchmark. At this pace, 10 years or more is a real possibility in the near term, and any competitive intelligence timeline projection should allow for it.
Some companies outperform. The best pharma company reaches 7.7 years and the best biotech 5.8 years, which shows speed can become a competitive advantage.
Figure 3: Median phase duration and development timelines (2011-2025)
The pooled figure also hides large differences by therapy area. Metabolic sits at 8.4 years and CNS at 10.8 years. Oncology has the longest Phase I median of any therapy area, 3.0 years against 2.2 overall, and the shortest Phase III median. For oncology, Phase I timelines may carry more weight than Phase III timelines.
Figure 4: Development timeline by therapy area (2016 – 2025)
