THE SECOND VALLEY OF DEATH: WHY GOOD BIOTECH SCIENCE FAILS IN DILIGENCE

By Kelly Cycon, PhD | 20/15 Visioneers

If you’ve ever watched a term sheet slow down over a question that had nothing to do with your science, this is for you. Every founder braces for the funding valley of death, and an entire ecosystem of grants, accelerators, and tech‑transfer programs exists to pull them across it. Almost no one warns you about the second valley: the operational gap that gets exposed at diligence, IND, and partnership; where fundable science stalls, not because the work was wrong, but because the evidence wasn’t organized.

If you’re a founder, CSO, or investor who’s felt that hollow pause when the data were real but the proof wasn’t portable, read on. The second valley is real, predictable, and unlike the first, almost entirely avoidable.

The meeting that changes everything

Picture the scene.

A founding scientist sits across a conference table from a Series B investment partner. He’s built something real: compelling efficacy data, a differentiated mechanism, a platform that could matter. The science has held up through two rounds of scrutiny. This meeting should feel like a formality.

Then the investor asks a question that sounds simple.

“Walk me through your data integrity controls. If a key technician left today, how would we know that your lead assay results are reproducible by someone else?”

The founder pauses. He and his team ran those assays correctly. He’s certain of it.

But the documentation that would prove it to an outsider? It exists in fragments. Some lab notebooks. Some shared drives that were never really organized. Some protocols written by the technician who left four months ago and never fully handed off.

“We’re in the process of formalizing that,” he says.

The investor writes something down.

The meeting doesn’t end badly. But the term sheet that arrives three weeks later carries a condition: independent quality audit before close. Three months of the founder’s time. A valuation adjustment. A deal that almost didn’t happen at all.

The science passed. The operating system failed. 

If you’ve ever sat in that chair, you know that hollow feeling. The work was real. The risk lived somewhere else entirely.

The Valley Everyone Knows About

The life sciences community has a name for the first existential threat a startup faces: the valley of death. The funding gap between early-stage discovery and institutional capital. The translational gap, where promising science runs out of runway before the market is ready to fund it properly.

It’s a well-documented phenomenon with a support infrastructure built around it: NIH bridge funding, SBIR/STTR grants, seed-stage biotech accelerators, technology transfer programs. We have built an ecosystem to help companies cross that first chasm. 

Most founders know it’s coming. Many survive it.

But there’s a second valley. And almost nobody warns you about it.

Who Should Care About the Second Valley

This second valley shows up for people who are close to the science and feel responsible for its future:

• Founding scientists and CSOs running seed to Series B programs.

  
  

• Biotech investors who care about execution risk, not just biology.

  
  

• Operational and quality leaders asked to “make things audit-ready” while still keeping the lab moving.

If you’re responsible for moving a program from “promising science” to “fundable, partnerable asset,” this second valley is where you lose months and valuation without a failed experiment. 

The Valley Nobody Warns you About

The second valley isn’t a funding problem. It’s an operational one. It appears not at the very beginning, but at the exact moments when a company should be accelerating: during investor diligence, before an IND submission, ahead of a manufacturing partnership, at the handoff from research to development. 

It’s built from a specific set of accumulated gaps:

• No audit trail. Experiments are reproducible in practice but undocumented in a way an investor, regulator, or partner can verify. The data is real. The proof isn’t portable. 

  
  

• Decisions without rationale. Formulations changed, suppliers swapped, methods modified, often correctly, but with no contemporaneous record of why those decisions were made that survives once the decision-maker leaves. 

  
  

• Training that happened but can’t be proven. A signed training record is what turns a standard operating procedure into a legally meaningful control. Without it, the SOP doesn’t count for FDA purposes, and increasingly doesn’t satisfy investor diligence either. You can have the right procedures and still fail the audit. 

  
  

• A quality system assembled under scrutiny. There is a visible difference between a quality system built deliberately, from the ground up, and one bolted together in response to questions. Investors and regulators know the difference. The former signals discipline; the latter signals a team that was caught unprepared.  

This is where good science dies. Not in the lab, where the work is sound. In the diligence room, where the evidence isn’t organized. 

Why it Happens (and Why it's Not Negligence)

The second valley isn’t the product of carelessness. It’s the product of rational prioritization under constraint. 

When a founding scientist is simultaneously running experiments, hiring a first team, managing investor relationships, and trying to extend runway, building a quality management system feels like premature optimization. Documentation infrastructure is a cost with a deferred benefit. The payoff isn’t visible until you need it, and by then, it’s too late to build it carefully. 

At the seed stage, the QMS doesn’t close the next funding round. The data does. So the documentation waits. 

The problem is that operational gaps compound:

• Every week without version-controlled procedures is a week that critical knowledge accumulates in people rather than in documents. 

• Every hire trained verbally rather than through a signed acknowledgment form is a liability that grows with the company.

• Every scientific decision made without a change control record is a gap that will have to be reconstructed later, under pressure, at significant cost.

The founders who navigate transitions cleanly aren’t necessarily better scientists. They’re the ones who quietly built the infrastructure while they still had the time and bandwidth to do it right. 

If this sounds uncomfortably familiar, I want to be clear: this pattern is understandable. It’s all too common. And it’s quite fixable.

What Sophisticated Diligence Actually Finds

For investors conducting real quality diligence, not a cursory document review, but the structured assessment that sophisticated biotech investors increasingly require, the second valley is visible from a long way out. 

The signals are consistent:

• Training matrices with gaps.

• SOPs that exist but can’t be mapped to signed acknowledgment records.

  
  

• Change history reconstructed from email threads.

• Electronic records never validated against 21 CFR Part 11 (the FDA’s rules for trustworthy digital records), even though the company has been using an ELN for two years.

• Supplier qualification that amounts to “we’ve always used them.”

None of these findings is necessarily fatal on its own. But collectively, they tell a story about operational maturity or its absence. They signal execution risk in a domain that is supposed to be infrastructure, not science. And in a funding environment where investors have genuine choices about where to place capital, execution risk in the operating layer matters. 

The inverse is equally true. Companies that arrive at diligence with organized, current, version-controlled documentation don’t just pass the quality review; they compress the timeline. A clean data room is a signal of organizational competence that extends beyond the documents themselves. It tells an investor that this team knows how to build systems, not just science. 

As someone who cares deeply about seeing good science move forward, I think about that distinction a lot. It’s not about perfection. It’s about whether the story your systems tell matches the quality of the work you’re doing.

The fix isn’t a binder of SOPs

The conventional framing of quality systems in life sciences emphasizes compliance: the regulations that must be satisfied, the documents that must exist, the audits that must be survived. That framing is accurate as far as it goes, but it misses something important about why systems fail. 

Most quality systems that fall apart don’t fail because the SOPs were wrong. They fail because the SOPs were treated as a deliverable rather than a living infrastructure. Someone builds a document library, hands it over, declares victory, and moves on. Eighteen months later, the library is out of date, the training records haven’t been maintained, and the system that existed on paper has diverged from how the company actually operates. 

The companies that make it through transitions cleanly have something different: an operational system installed early, one that grows with the company and captures workflows, decisions, and evidence continuously, not in response to a trigger but as part of how the organization runs. 

This distinction matters more than any specific document. A quality system that lives is worth more than a quality system that was perfect on day one and hasn’t been touched since. 

None of this is about making anyone feel judged for not having every piece in place. It’s about recognizing that a modest, living system is more protective than a beautiful static binder you only open when someone calls an audit.

The day-one quality checklist for fundable science

For founders who have never built operational infrastructure from scratch, the question is often: where does this actually start? 

The answer is more concrete than most people expect.

By the end of the first month, a company starting from zero can have the foundation in place: Document control: a simple procedure, numbering convention, and register that assign version numbers, manage change history, and keep everyone on current procedures. From that point forward, every SOP written is automatically controlled. 

• Quality manual: a concise description of how the quality system as a whole is organized; who owns it, what it covers, and how it connects to the rest of the organization’s procedures. 

• Training matrix and signed acknowledgment forms: the legal record that turns configured SOPs into enforceable procedures and proves that training happened, not just that it was intended. 

• CAPA and deviation procedures: a defined way to capture, investigate, and respond when things go wrong. 

• Data integrity controls: practical rules and safeguards for how data is generated, stored, changed, and reviewed. 

• Supplier qualification framework: documented criteria and records for who you trust with critical materials and services and why. 

  
  

The modality-specific layer, the SOPs specific to the biology you’re developing comes in the following weeks. The lab keeps running through all of it. The infrastructure goes in around the work, not instead of it. 

None of this requires specialized software. It requires a decision to treat operational infrastructure as part of the science, not a future administrative problem. 

I’m helping teams put this foundation in place in environments that were already stretched thin. It’s work, but it’s manageable. And the relief on people’s faces when they realize they can answer the hard questions without scrambling is part of why I am doing this.

The Question Worth Asking Now

Back to the founding scientist at the conference table.

He closed the round. But the operational gap cost him the better part of a year: time spent reconstructing documentation that should have existed, working with consultants under deadline pressure, demonstrating competence to an investor who was already skeptical. The science was always fundable. The organization wasn’t ready. 

Every company eventually pays for documentation, traceability, and process capture. The only variables are when and how much. Building the infrastructure early, before the pressure arrives, before the diligence clock starts, before the regulatory submission is six weeks out is almost always cheaper, faster, and less disruptive than building it in response to a crisis. 

A simple diagnostic you can use right now:

• Can you produce a training matrix mapped to current SOP versions?

  
  

• Can you show change history for your lead assay without digging through email?

• Can you point to a documented data integrity procedure that matches how your team actually works? 

If the honest answer to any of these is “not yet,” you’re already in the second valley. You just haven’t been asked the hard questions in a room where the answers affect valuation and timelines.

I don’t say that to alarm you. I say it because the earlier you see it, the more choices you have about how to address it.

The second valley of death is real, predictable, and avoidable. It doesn’t ask whether your science is exciting; it asks whether your organization is ready to carry that science all the way to patients. The founders who avoid it are the ones who decide, early enough, to build a foundation that serves the innovation, protects the people who depend on it, and gives their work a real chance to become the medicine, tools, and therapies the world is waiting for.

Kelly Cycon, PhD, leads the Pharming Services program at 20/15 Visioneers, where she works with incubators and early-stage life sciences companies to implement operational infrastructure that can withstand investor, partner, and regulatory scrutiny. She can be reached at kelly.cycon@20visioneers15.com.