Phase 2 is where most emerging sponsors first feel the gap between what they can run and what the study requires. The site count rises, enrollment becomes a genuine risk rather than an assumption, and the data package starts carrying weight with the next investor.
It is also where CRO selection stops being a procurement exercise and starts being a development decision.
Weight these four things above price
Indication experience, not therapeutic area experience
"Oncology experience" covers an enormous range. A CRO with deep experience in large solid tumour studies may have none in your rare hematological indication, where the site network, the referral pathways and the enrollment dynamics are entirely different.
Ask for studies in your indication, at your phase, in the regions you plan to run. Then ask how those studies actually performed against plan.
Named staff, and their continuity
Team turnover is among the strongest predictors of delivery problems and among the least examined at selection. Ask for the functional leads by name, what else they are committed to, and turnover in that therapeutic unit over eighteen months.
Put the names in the contract with a continuity expectation and a notification requirement. A CRO that resists this is telling you something useful.
Demonstrated enrollment performance
Every CRO will model your enrollment optimistically because you supplied optimistic assumptions. What distinguishes them is what they did on previous studies when enrollment fell behind.
Ask for a specific example. Listen for whether the answer describes a rehearsed process — site-level diagnosis, protocol amendment analysis, additional site activation, targeted outreach — or a general commitment to partnership.
Governance maturity on their side
You will build the sponsor-side governance. But the CRO has to be able to operate inside it: escalate early, bring issues before they are crises, produce metrics that mean something.
Ask how they escalate to a sponsor, with an example. Ask what they would do in month nine if they believed your protocol was the enrollment problem. The willingness to say something unwelcome is worth more at Phase 2 than any capability slide.
The failure modes that show up around month nine
They are consistent enough to be worth naming in advance.
The named team has rotated and nobody told you formally, because the contract did not require it.
Enrollment is behind, the mitigation plan is a list of activities rather than a diagnosis, and no one has asked whether the eligibility criteria are the cause.
Change orders have started arriving, each individually reasonable, cumulatively substantial, with no agreed process.
Governance meetings review the same dashboard every month without producing decisions, and the metrics that would have shown the problem early were never agreed.
All four are foreseeable at selection. Three of them are preventable in the contract.
What a Phase 2 selection process should actually contain
Scoping before the RFP, so the bids answer the study you are running rather than the one in the pitch deck. A written set of assumptions every bidder prices against. A scoring framework agreed before proposals arrive, so the debate is about evidence rather than preference. Bid defense structured around the delivery team, not the business development team. Reference calls with sponsors of comparable studies. And contracting that carries the named staff, the governance cadence and the change control mechanism into the agreement itself.
Eight to sixteen weeks, run properly. The sponsors who compress it to four almost always spend the difference later, with interest.
What to decide before the RFP goes out
Scope the study honestly first — realistic enrollment, realistic site count, realistic duration. An RFP built on the timeline you need for financing rather than the one the study will take produces bids priced for a study that will not happen.
Decide your outsourcing model before you decide your CRO. Full-service, FSP and hybrid impose different demands on your team, and choosing a model that assumes capability you have not built is the most expensive error available at this stage.
Then select against the four criteria above, and accept that the best proposal and the best partner are not reliably the same document.
Phase 2 is usually the first selection that has real consequences. ABRA runs CRO selection for emerging sponsors.
