The bottleneck is evidence, not discovery

Candidate generation has collapsed in cost. Clinical validation has not. The bottleneck simply relocated — from chemistry to recruitment, sites, and regulatory strategy — and a platform's advantage stops compounding at exactly the point where the bottleneck moved.

Value therefore concentrates in whoever compresses trial cost and cycle time: site networks, recruitment, real-world evidence, and regulatory strategy sold as a product rather than as consulting.

Most software pitched at pharma claims a share of a budget it has no access to. We decompose a phase-two programme line by line and ask which items the product can actually touch. Usually it is recruitment and monitoring, occasionally data management, rarely anything else.

That exercise removes about two-thirds of the addressable market from most decks, and it makes the remaining third believable.

In health delivery the customer is a system with a compliance function and an annual budget. Accuracy alone does not clear procurement. We have a diagnostic in the file that beat the human baseline on published accuracy and spent two years in pilots — signing nothing until it produced documented throughput savings in one named department. The sales cycle then halved.

Lead with the budget line you reduce. Publish the accuracy in the appendix.

Tooling companies get paid whether or not the molecule works. That is a genuinely good business, and it is also a reason to check whether the multiple is pricing discovery volume — which is growing fast — or clinical outcomes, which are not.

If the multiple assumes outcomes, the tooling company is being paid for risk it does not carry.