The best sites don't have time for you. Cherish the ones that make time, and use it wisely.

Early in my career, I would come back from a site diligence call enthusiastic when the principal investigator had spent two hours with me. The PI had walked me through the practice, introduced me to the coordinator, talked through the protocol page by page, and asked thoughtful questions about how the study would integrate into their workflow. I would write up the call and tell my team this was a great site. I would be wrong about that almost every time.

The pattern took me years to see. Then one season I noticed it across half a dozen activations: the PIs who had been most generous with their time at qualification were also the PIs whose practices were producing the lowest weekly enrollment. The PIs who had been hardest to schedule, who had given me thirty minutes of clipped, focused conversation between patients and then ended the call abruptly because they had to go consent a real one — those were the PIs whose sites were producing.

The mechanism is not subtle. A doctor's time is the most precise signal of their patient volume that exists. A PI who can spend two hours on a non-clinical conversation in the middle of a Tuesday afternoon is a PI whose Tuesday afternoon was not full. If their Tuesday afternoon is not full, their patient volume is not what their feasibility form says it is. If their patient volume is not what their feasibility form says it is, the study will not enroll at the rate the feasibility form predicted. Every step of the inference is direct.

The best sites, in other words, do not have time for you. They have time for their patients, and the patients are the entire point. This is the inversion that took me too long to internalize. The friction-rich diligence call — the PI who reschedules twice, the coordinator who has to take the call from her car between visits, the visit that gets cut short because clinic is running long — is not a sign of disorganization or disrespect. It is the signal that the site you want is exactly the kind of site that does not have time to perform a smooth diligence relationship for you.

The same signal carries through to the staff. A coordinator who is on the call for forty-five minutes, on Zoom, in the middle of clinic hours, with no patient interruptions and no glances at the door, is a coordinator whose clinic is not running at the volume the practice claims. Coordinators in busy practices interrupt diligence calls. They are pulled away to consent a patient who has just walked in. They reschedule the call because the morning ran long. They take it from their car between visits. The coordinator who is sitting attentively for the entire scheduled hour is the coordinator whose clinic is, on that day, not seeing the patients that would have produced an interruption.

This is the signal CROs systematically reward in the wrong direction. Conventional qualification scoring favors sites that respond fastest to outreach, schedule diligence calls easily, complete questionnaires promptly, and provide attentive availability throughout the call. All of those behaviors correlate with low patient volume, not high. The site that takes three weeks to schedule the call, reschedules twice because the clinic schedule shifted, and ends the call abruptly when the PI gets paged is the site that is going to enroll your study.

So far, this is a warning. The harder and more important part of the principle is the corollary. The good sites — the ones that do not have time for you — will, occasionally, make time for you anyway. The PI who normally has three back-to-back consults will block off thirty minutes after clinic ends. The coordinator who is usually pulled away mid-call will sit down with you on a Wednesday afternoon she has rearranged her week to give you. These moments are rare and they are precious, and the discipline that distinguishes a CRO that succeeds at sites like this from one that does not is what the CRO does with that time when the site has chosen to give it.

The default mode is to fill the time. Walk through the slide deck, hit the milestones on the agenda, ask the questions on the qualification form, get through the protocol page by page. This is the wrong response. The good site has, by definition, more value to deliver in thirty minutes than a marginal site has to deliver in three hours, and treating the time the same way — slide deck, agenda, page-by-page — wastes it. The good PI's thirty minutes is where you ask the question whose answer is going to determine whether the study works at this site, and you ask it directly, and you listen to the answer the way the answer deserves.

What does that look like, concretely. You arrive prepared in a way you would never be prepared for a marginal site. You have read every public paper the PI has written that touches the analyte. You have looked at the practice's claims footprint and have a specific question about a population sub-segment you saw in the data. You know the names of the lead coordinator, the phlebotomist, the front-desk supervisor. You have a draft kit specification that takes a position on the questions a competent PI is going to raise, and you bring it to the meeting rather than treating the meeting as the moment to start the conversation. The preparation is the respect. The PI who has cleared their afternoon notices the difference between a CRO that came ready and a CRO that came to read off a slide.

You also ask different questions. Not the qualification-form questions; those can be answered by the coordinator over email. The questions worth asking when a serious PI has cleared their afternoon are the ones that surface the operational reality of the practice in a way the form cannot. What changed in the patient population in the last twelve months. Which of the practice's referral sources has gotten weaker, and which has gotten stronger. Where the consent flow currently breaks down on similar studies. Which of the protocol's eligibility criteria the PI privately expects will be the binding constraint on enrollment. These are the questions the PI is qualified to answer in a way no one else is, and the moment when they have given you their time is the moment to ask them.

And you finish on time. The PI who has cleared thirty minutes for you should not be asked for thirty-five. The discipline of ending the meeting at the time it was scheduled to end is itself a signal that you understand whose time you are using and what it costs them. CROs that run over on diligence calls are CROs that the good PIs quietly stop giving time to, the same way good restaurants quietly stop taking reservations from the customers who routinely arrive late.

I am still uncomfortable telling sponsors this directly. The reflex is to want to brag about how engaged our partner sites are, how attentive the PIs, how cooperative the staff. The honest version is the opposite: the engaged, attentive, cooperative sites are not, on average, the sites we want to activate. The sites we want — the ones with full clinics and busy PIs and coordinators who interrupt our calls to consent real patients — are the ones whose studies hit timelines. And when those sites give us time, we use it carefully, because we know what it costs them. The CRO that can sit with that paradox without defaulting back to the easier-to-measure signal is the CRO whose site lists actually enroll, and whose sites stay loyal across studies because the relationship is built on respect for the work the site is actually doing.