Before an RFP goes out, a sponsor should speak directly with five to ten experienced investigators and site research leaders about the study. Those conversations test the assumptions that drive timelines and cost, including how many eligible patients a site actually sees, what standard of care looks like, which competing trials are enrolling, and whether the visit schedule is realistic. The result is a stronger protocol and better inputs for every CRO you invite to bid.
Why go to sites before the RFP?
The most consequential assumptions in a CRO proposal, enrollment rate and site count, come from the sponsor's description of the study. If that description is optimistic, every proposal inherits the optimism. In a 45-minute conversation, an investigator can tell you that the eligibility criteria exclude most of their patients, that a key procedure is not standard at community sites, or that three competing studies are enrolling the same population. Learning that before the RFP costs very little. Learning it after the first quarter of enrollment can cost months.
Who should you speak with?
Aim for a mix that reflects where the study will actually run:
Principal investigators who treat the target population
Study coordinators, who know the real burden of each visit
Research directors or site administrators, who know contracting and startup timelines
A combination of academic medical centers and community or private research sites
Sites in each region or country you are considering
Patient advocacy leaders, particularly in rare disease
What should you ask?
Keep the questions consistent across conversations so the answers can be compared. Useful questions include how many patients meeting the draft criteria the site sees each month, which criteria would exclude the most patients, what standard of care looks like today, which competing studies are enrolling, which visits or procedures would be hardest for participants, how long contracting, budget approval, and ethics review usually take, and what would make this study a priority for the site.
How do you run these conversations well?
Put a confidentiality agreement in place first, then share a protocol synopsis instead of the full draft. Schedule calls of 30 to 45 minutes and use a short, standard interview guide. When you engage investigators as advisors, compensate them at fair market value under a written agreement, and keep those advisory conversations separate from any later site selection decision so the process stays clean. Document each conversation in a simple summary that captures numbers, concerns, and suggestions.
What do you do with what you learn?
Use it in three places. First, revise the protocol where investigators identified avoidable barriers. Second, set realistic enrollment and site-count assumptions for the RFP, so every CRO prices a study that can actually be delivered. Third, keep a list of interested sites to share with the CROs you invite, which helps you see how each one approaches feasibility and whether its proposed sites overlap with yours.
Does this replace CRO feasibility?
No. A CRO's feasibility work covers many more sites, often with data from previous studies, and it remains an important part of selection. Sponsor-led conversations serve a different purpose: they let you walk into the bid process with your own evidence, so you can recognize which feasibility findings align with what investigators told you directly and ask good questions where they diverge.
How much sponsor time does this take?
Less than most sponsors expect. Ten conversations of 45 minutes each, plus preparation and summaries, typically add up to roughly 15 to 20 hours of sponsor time spread over three or four weeks. That work can run while the protocol synopsis is being finalized, so it rarely delays the RFP. Compared with a single protocol amendment, which can mean months of rework across sites, ethics committees, and systems, it is among the most efficient investments a sponsor can make before a study begins.
Keep reading: Representative Enrollment Is an Operational Design Problem
