Start with a written search brief
Write down the mechanism or indication, permitted countries and trial settings, maximum next-experiment budget, exclusions and how recently ownership and status must have been verified. Separate licensing cost from experiment cost. A desired number of leads is a preference, not a qualification rule.
Resolve the asset before assessing it
Treat the molecule, formulation, indication and program as distinct fields. Record aliases and trial identifiers. A company name match is not enough to establish that a publication, patent and discontinued program describe the same asset.
Gate A: establish enough evidence to judge
For each essential fact, retain its source, publication date, last verification date and what was actually checked. An older completed study can still support scientific evidence. An old pipeline page does not establish current availability. Missing ownership or an unexplained discontinuation returns the record to research before scoring.
Gate B: apply six mandatory tests
Check scientific fit; readable completed Phase 1 or Phase 2 human data; an acceptable stopping reason; reachable rights; an affordable credible next experiment; and an endpoint that can answer the question in the required time. Evidence for every required test must stand on its own. A strong score elsewhere cannot compensate for a mandatory failure.
Document why development stopped
Compare dated company disclosures with the completed trial record and published results. Funding pressure and efficacy failure can coexist. Under Hive's strict stranded-asset mandate, a confirmed disqualifying safety or efficacy failure excludes the opportunity. A missed primary endpoint needs explicit adjudication; a favorable subgroup does not automatically reverse the finding.
Map the rights package
List patents, territories, indications, clinical data, regulatory files and drug supply separately. A sponsor or patent assignee is not automatically the current controller of every component. Public partnering language can identify a route to inquiry; it does not prove that the package is available on acceptable terms.
Cost the next answer
State the decision, design, endpoint, follow-up, recruitment and manufacturing assumptions. Itemize the experiment's costs and identify missing quotes. Unknown cost means more research. A supported estimate above the agreed ceiling fails the current mandate. No budget ceiling means no final PASS.
Worked screening example: a hold is useful
Suppose a hypothetical oral candidate has completed human testing and the company announces a portfolio refocus. Its data owner and supply access remain unconfirmed. The correct next step is a rights-and-data investigation. It is not a qualified licensing recommendation, regardless of how promising a biomarker chart looks.
What a decision-ready report contains
Include one clear identity, current owner and scope of access, supporting and contrary evidence for all six tests, a dated decision, the cheapest credible next step, a preliminary cost model and changes since the last review. Reassess when evidence, status or rights change. Never fill a shortlist quota with unresolved records.
Questions people actually ask
FAQ
Does discontinued mean available to license?
No. Current control of the relevant rights, data and supply must be established separately.
Can a trial registry prove current ownership?
A registry helps identify studies and sponsors. It does not by itself establish the complete current licensing package.
What if no assets qualify?
The shortlist stays empty. Report investigation progress and the unresolved decisions without presenting internal leads as qualified assets.
Related
Do the useful thing now