The 40% problem: why most preclinical data never gets published
Around 40% of preclinical research data never sees daylight. After interviewing 23 researchers at UCL and Oxford, we found the cause isn't a lack of willingness — it's workflow friction.
What is the 40% problem?
The 40% problem is the gap between the data a lab generates and the data the wider scientific community ever sees. Around 40% of preclinical research data is never published — not because it failed, and not because scientists are hiding it, but because documenting, structuring and uploading that data sits at the very end of a project, after the funding, people and momentum have moved on. That hidden data is often exactly what another lab needs to avoid repeating an experiment that was already answered. It is where the 40% problem collides directly with the 3Rs — Replacement, Reduction and Refinement.
- Data unpublished~40%
- Researchers interviewed23
- UK licensed establishments134
Why the data never leaves the lab
Three frictions turn most preclinical data into work that never gets finished — none of them about how much researchers care.
01Documentation lives at the very end
Structuring, cleaning and uploading data is the last task of a project — by which point the funding, the people, and the momentum have already moved on. Sharing becomes unpaid, unscheduled work, and it loses every time.
02Spreadsheets capture values, not context
A spreadsheet records numbers, not the links between them — which animal, which treatment, which dose, which outcome. Without that structure, data can't be reused, so it's never worth the effort of publishing.
03Knowledge walks out the door
When the context lives in a PhD student's head and their local spreadsheet, it leaves when they graduate. The next person re-runs the experiment because there was no durable record to build on.
What 23 researchers at UCL and Oxford told us
We set out to build a data-sharing platform. The interviews changed our minds — the problem was never sharing, it was the workflow meant to produce something worth sharing.
Researchers want to share — the process stops them
Across 23 interviews at UCL and Oxford, the blocker was never willingness. It was the friction of reconstructing context, reconciling naming conventions, and formatting data for a repository after the real project was over.
The bottleneck is workflow, not motivation
When sharing is a separate project bolted onto the end of a study, it consistently loses to the next grant, the next deadline, and the next cohort of students.
Hidden data drives redundant animal use
Because data never leaves the lab, experiments get repeated — and labs end up maintaining more animals than the science actually requires. Reducing that redundancy is the 3Rs in practice.
It's an infrastructure gap, not a culture gap
The fix isn't to ask scientists to try harder. It's to make capturing reusable, structured data the path of least resistance — built into the workflow rather than added on top of it.
Why we changed our minds
DaRev's founder, Dr. Hsin Chen, is an Oxford DPhil and pharmacist who worked as a preclinical researcher studying Niemann-Pick disease. We assumed the field needed a better data-sharing platform. After 23 conversations with researchers at UCL and Oxford, it was clear the real bottleneck sat one step earlier — in the workflow that was supposed to generate reusable data in the first place. That insight is why LabArk captures structure as the work happens, instead of asking scientists to reconstruct it later.
Make reusable data the path of least resistance
The reason 40% of data goes unpublished is that it was never structured in the first place. LabArk captures every animal, treatment, dose and outcome as structured data while you work.
Capture structure as the work happens
Every animal, treatment, dose and outcome is recorded as structured data at the moment of the experiment — not reconstructed months later from cage cards, photos and an email thread.
Sharing becomes a byproduct
When data is structured from the start, publishing and reuse stop being an end-of-project task. Compliance and 3Rs reporting can generate themselves from data you've already entered.
Knowledge stays in the system
One durable, structured record per study means context stays in the lab when a PhD student graduates — instead of walking out the door with them.
One record. Every step of the study.
Start where your lab fits. Grow into institution access when AWERB is ready.