The Lean Startup Is Great—Except When It Isn’t
Principles over tactics: “Startup success depends on identifying your dominant risk. If it’s market risk, Lean works. If it’s technical risk, Lean fails.”
When Eric Ries first published The Lean Startup in 2011, it quickly became a sensation in the entrepreneurial world. When it first came out, I remember reading the entire book as a 21 year old in a single day (with copious amounts of coffee). Newly into the world of tech and startups, this book was like my bible. Its iterative approach to company-building was something I had not considered before. It offered a method that tried to minimise risk by emphasising early market feedback, rapid prototyping, and continuous iteration.
Although Ries’s principles have merit—particularly for software startups operating in fast-moving markets—I soon learnt these ideas are not universal. As I got older and wiser, I realised that as we look at more complex ventures such as biotech, hardware, and even certain creative industries, we start seeing the cracks in a methodology that was forged in the internet age.
In this piece, I want to critique The Lean Startup by taking a close look at how it fails in biotech settings and how that breakdown reveals deeper limitations. From there, I will explore how “lean” lessons might be adapted (or abandoned entirely) for industries where market demand is not the primary risk factor—and indeed, where the low-hanging fruit of easy software solutions has already been picked.
I will then try to propose a more foundational, principles-based approach. It is still not a fully formed concept in my mind. But hey, what better way to improve your thinking than to get the internet to comment on it?
Is this the end of "The Lean Startup"? Unlikely. The problem lies in its overemphasis on tactics rather than the principles that underpin them. When the landscape shifts, tactics become obsolete, but principles stay.
At its core, the key principle is risk management. Yes, boring old risk management.
The Biotech Case: Where Lean Startup Methods Break Down
Ries’s methodology hinges on the idea of “Build, Measure, Learn.” A team conceives a minimum viable product (MVP), delivers it to a group of early adopters, collects data on real-world usage, and iterates. The feedback loop is short, ideally measured in days or weeks. This cycle is what gives lean startups the agility to pivot or persevere in response to genuine user feedback.
“The goal of a startup is to figure out the right thing to build—the thing customers want and will pay for—as quickly as possible.” - The Lean Startup
But let’s consider biotech1 as a counterpoint. Whether a startup is developing a novel gene therapy, an immuno-oncology drug, or a medical device, the constraints are wholly different. In biotech:
You Can’t Always Build an MVP
A biotech “product” might involve complex molecules or biologics that simply cannot be slapped together in a garage lab. Producing a new drug candidate requires years of preliminary research, specialized facilities, and compliance with rigorous regulatory standards before you even get to testing in humans.Testing Isn’t a Quick Feedback Loop
Even after the initial design phase, biotech startups cannot release a half-baked product to an audience of early adopters. They must go through multiple phases of clinical trials—often spanning years—before reaching FDA approval or an equivalent regulatory green light. Testing is iterative in a sense, but the timescale is long, and feedback is not from “users” but from carefully structured clinical trial outcomes.Technical Risk vs. Market Risk
The crux of The Lean Startup is that your primary risk is market risk: does anyone want your product? Will customers pay for it? Biotech typically inverts this equation. The “market” for a life-saving drug is often known—patients with certain diseases exist, and many are desperate for treatments. Demand is rarely in question. The real risk is technical risk: can you develop a compound that is both safe and efficacious for a specific disease?
Biotech companies violate just about all the ideas found in The Lean Startup but (some) manage to be very successful. How?
Biotech underscores a fundamental limit of the lean model. For projects where speed to market is restricted by scientific and regulatory complexity, there can be no frequent “release early, release often” loop. Clinical trials also cost tens or hundreds of millions of dollars—an order of magnitude more than what’s needed to spin up a new social media or e-commerce platform. The carefully tuned machine of lean iteration thus falters in the biotech context.
From Internet-Era Startups to Deep Tech: Shifting Risk Profiles
The enduring popularity of lean methods stems largely from their success with software ventures. In the internet age, software startups benefited from:
Established Infrastructure: Many foundational technical problems had already been solved—cloud computing platforms like AWS made deployment cheap, open-source frameworks handled core functionalities, and broadband internet was ubiquitous.
Frictionless Distribution: Startups could release apps and updates instantly, often at minimal cost. This enabled the “fail fast” or “pivot quickly” mantra that defines lean thinking.
Dominant Market Risk: For most software startups, the question was not “Can we build this?” but “Will people want and pay for it?” Hence, small teams with minimal resources could experiment and refine prototypes in real time.
While there are certainly software startups for whom market risk remains central, a significant portion of today’s cutting-edge innovation involves more complex, capital-intensive ventures. Fields like artificial intelligence, quantum computing, space tech, robotics, and clean energy all carry heavy technical risk. As with biotech, these projects cannot simply iterate their way to success in the manner Ries recommends—building an MVP and then adjusting every two weeks is often not feasible.
Moreover, even in “traditional” software spaces, a shift is happening. Much of the easy, obvious innovation—creating a marketplace, building a social platform, spinning up an e-commerce site—has been done. These days, starting a software business from scratch often means competing with giant incumbents. That raises the stakes for both technical ingenuity (differentiating your product) and marketing strategies (cutting through massive noise).
The Declining Relevance of Lean Startup—Or a Need for Adaptation?
Is The Lean Startup still relevant, or do we need a new paradigm entirely?
For pure software startups where market demand is uncertain: Traditional lean thinking can still be powerful. Building quick prototypes, running A/B tests, and pivoting can help reduce waste. There is no need to throw it out altogether.
For ventures with high technical complexity: Lean practices need to be adapted. You cannot “fail fast” if each failure costs millions of dollars or years of effort. Validation must happen in different ways—through careful modelling, simulations, or collaborative research with academic institutions.
For areas beyond software: The key is to identify what kind of risk truly dominates. If market risk is indeed paramount, lean techniques remain useful. If technical risk is more pressing, a different blueprint is needed—perhaps derived from existing scientific or engineering methodologies.
Perhaps “The Lean Startup” needs to evolve into “The Risk-First Startup Model”?
Your startup's strategy depends on where its risks lie. They often lie on a spectrum - technical on one end, market on the other.
Conclusion: Beyond Lean, Toward Smart Risk Management
Ultimately, the best advice is to identify your startup’s dominant risks and mitigate them. That might mean building an MVP and doing rapid A/B tests for a consumer-facing app—or it might mean three years of intense lab research before you ever talk to a regulator. It might mean forming a consortium with universities, or it might mean building a pilot plant. The path you choose depends on the nature of the problem you’re solving and the resources at your disposal.
So yes, read The Lean Startup—it’s a valuable resource. But then close the book and ask: “Where do my real risks lie?” Your answer to that question should determine your strategy, not just blindly following the build - test - iterate loop.
I remember Celine Halioua’s blog “How to Build a Biotech” really challenging my thinking back in the day (https://www.celinehh.com/tech-vs-biotech). You should check it out.