Drug development keeps getting harder and more expensive — even as the tools get better. Here's what's really going on, and what it means for your R&D operations.
In 1965, Gordon Moore made an observation that became a prophecy: the number of transistors on a chip would double roughly every two years, steadily compressing the cost and size of computing power. Moore's Law has held, more or less, for six decades. It is one of the most reliable trends in the history of technology.
Then there's the other law. The one that works in reverse.
Eroom's Law: the number of new drugs approved per billion dollars of R&D spending has roughly halved every nine years since the 1950s.
Spell "Moore" backwards and you get the name. That's intentional. Coined by researchers Jack Scannell and colleagues in a landmark 2012 Nature Reviews Drug Discovery paper, Eroom's Law captures a disturbing and persistent trend: despite extraordinary advances in biology, chemistry, genomics, AI, and computing, the productivity of pharmaceutical R&D has been declining for decades.
We are spending more, discovering less.
In our first article in this series, we introduced the concept of the Discovery Tax — the compounding cost of operational friction that quietly drains time, money, and momentum from scientific programs. Eroom's Law is that friction playing out at the industry level. It is the macro-level consequence of what happens when the Discovery Tax goes unaddressed, organization after organization, decade after decade.
Understanding Eroom's Law is not an academic exercise. It's a lens for understanding the structural forces bearing down on every biopharma R&D organization today and why eliminating operational friction is no longer a back-office concern. It's a scientific imperative.
Drug R&D costs have increased 100x since the 1950s — but the real drag on productivity isn't the science. It's the operational layer around it.Let's start with the data. In the early 1950s, a billion dollars of R&D investment (inflation-adjusted) produced roughly 40–50 new approved drugs. By the 2000s, that same billion dollars was yielding fewer than five. By some estimates, drug development costs have increased 100-fold in real terms since 1950, while the number of approvals has stayed relatively flat.
The U.S. pharmaceutical industry now spends over $100 billion annually on R&D. The average cost to bring a single new drug to market (accounting for failure rates) has been estimated at anywhere from $1 billion to over $2.5 billion depending on the therapeutic area and methodology. These aren't just large numbers. They represent a deepening crisis of efficiency.
And it's getting harder, not easier. Complex targets. Regulatory requirements have grown. Clinical trials are more expansive and expensive. And yet the infrastructure supporting the scientists, program managers, and procurement teams who orchestrate all of this has, in many organizations, lagged far behind.
Scannell and colleagues proposed four interconnected forces driving R&D productivity decline. Each one has implications for how organizations manage their operations.
1. The 'better than the Beatles' problem. The first drugs in a category — statins, SSRIs, ACE inhibitors — were extraordinary breakthroughs. New drugs must now compete against them for regulatory approval. Demonstrating efficacy against best-in-class incumbent treatments is far harder than demonstrating efficacy against a placebo.
2. The cautious regulator effect. Post-Thalidomide, post-Vioxx, regulatory agencies have become more risk-averse and have steadily raised the evidence bar for new approvals. More trials, more endpoints, more data. All of this is rational, and all of it costs more.
3. The throw money at it tendency. As organizations have grown larger, there is a documented tendency to respond to productivity decline by increasing investment rather than changing approach. More headcount, more programs, more trials, without fundamentally addressing the efficiency of how science gets done.
4. The basic research-brute force gap. There has been an assumption that scientific advances like genomics, high-throughput screening, and computational tools would translate directly into drug productivity. They haven't, at least not at the rate expected. The bottleneck isn't knowledge generation. It's the ability to act on knowledge efficiently.
The bottleneck isn't knowledge generation. It's the ability to act on knowledge efficiently.
The third and fourth point is where the Discovery Tax lives. And it's these that organizations have the most direct ability to address.
Eroom's Law is what you observe when you zoom out across the industry. The Discovery Tax is what you experience when you zoom in on a single organization. They are the same problem at different levels of magnification.
Consider what actually happens inside a biopharma R&D organization on any given week. A scientist identifies a scientific supplier that could accelerate a key study. The supplier qualification process takes six to twelve weeks. Another team needs a specialized reagent. The procurement workflow involves manual emails, PDF approvals, and a finance review that adds three weeks to the timeline. A program manager tries to understand spend across six external suppliers and produces a spreadsheet that's out of date by the time it's reviewed.
None of these activities generate scientific value. All of them consume time, budget, and human attention. Individually, they seem like minor irritants. Cumulatively, across a portfolio of programs, they compound into something that looks a lot like Eroom's Law.
Five weeks per scientist. Across a team of 50 scientists, that's 250 weeks of research capacity that's currently being consumed by process friction. Across a large pharma organization with hundreds of scientists, it becomes an enormous drag on productivity — one that Eroom's Law is, in part, measuring.
It's tempting to look at the tools available today, like AI-driven molecule design, high-throughput screening platforms, predictive toxicology models, and assume that the productivity problem is being solved. It isn't. Not fully. And understanding why matters.
The scientific tools have become vastly more powerful. But the operational infrastructure around them, the systems by which scientists access external partners, manage suppliers, initiate studies, track spend, and coordinate across stakeholders, has not kept pace. In many organizations, the science happens on next-generation platforms while the operations run on email threads and spreadsheets.
This is the structural gap that the Discovery Tax exploits. It's not the science that's slow. It's the scaffolding around the science. And because that scaffolding is largely invisible, its cost tends to be absorbed and normalized rather than addressed.
It's not the science that's slow. It's the scaffolding around the science.
Five weeks per scientist, per year, lost to process friction. Multiply that across your team and you're looking at an enormous hidden tax on discovery.
The organizations beginning to bend the Eroom's Law curve share a common trait: they are treating R&D operations as a strategic discipline, not a cost center. They are investing in the infrastructure that connects science to execution — the systems that allow a scientist's need to translate immediately into action, without bureaucratic delay, manual re-entry, or process circumvention.
What does this look like in practice?
These are not glamorous interventions. They do not generate headlines. But they are the operational conditions that allow scientific talent to do what it was hired to do: advance programs, not manage process.
One of the most valuable things about naming a concept is that it creates the ability to measure and manage it. Eroom's Law has been a useful lens for understanding the industry's productivity challenge precisely because it gives researchers a concrete, quantifiable trend to study and debate.
The Discovery Tax gives organizations the same lever at the operational level. When you can quantify how much time, spend, and momentum is being lost to process friction, you can make a rigorous case for addressing it. You can compare the cost of the status quo to the cost of change. You can demonstrate ROI in terms the CFO and CSO both care about.
In the next article in this series, we'll explore how leading biopharma organizations are beginning to calculate their own Discovery Tax — and what the numbers reveal about where the biggest opportunities for recovery lie.
For now, the point is this: Eroom's Law is not destiny. It is a description of what happens when the operational layer doesn't keep pace with the scientific one. The organizations that close that gap, that eliminate the hidden friction between scientific intention and scientific action, are the ones building the architecture for faster, more productive discovery.
Eroom's Law is not destiny. It is a description of what happens when the operational layer doesn't keep pace with the scientific one.
The Discovery Tax is the invoice for that gap. And unlike most taxes, this one is optional.