A Society for Clinical Research Sites survey found that 58 percent of sites held six months or less of operating cash, and nearly all carried accounts receivable older than ninety days. Read those two numbers together: much of the industry's trial infrastructure runs closer to the edge than anyone likes to admit.
There is little reason to think the picture has improved since. Trials keep getting more complex, and the negotiation cycle that precedes every study keeps stretching: budget turnaround at US sites rose roughly sixty percent in two years, from about 32 days to 52. A site can be excellent at medicine, enrollment, and data quality and still run out of runway, because the failure mode is structural. That fragility is also the backdrop to the consolidation wave buying up independent sites.
The structure is a squeeze between when costs happen and when revenue does. A site's spending is front-loaded: staff hired and trained, systems stood up, regulatory work done, all before the first participant is screened. Revenue then arrives per completed visit, netted up quarterly, after sponsor review, minus a holdback of five to ten percent that may not release until database lock, years later. We have seen proposed terms under which even the startup fee, payment for work finished before screening opens, would have been paid at study close-out. We rejected that, on a principle worth stating plainly: costs should be paid when they are incurred.
A budget line that never becomes an invoice is not revenue; it is decoration.
The revenue sites never invoice
The second leak is quieter. Every negotiated budget contains items payable only if the site invoices for them: unscheduled visits, serious adverse event reporting, re-consents after amendments, pharmacy fees, long-term record storage, translation costs. Capturing these is genuine work, and at most sites it is nobody's explicit job, so it simply does not happen. Across a portfolio, uninvoiced items can represent hundreds of thousands of dollars of income the site already earned. A budget line that never becomes an invoice is not revenue; it is decoration.
The tooling does not help
Conventional site software makes this worse by measuring the wrong thing. Trial management platforms are competent at tracking payments due and weak to nonexistent at showing a site its actual profitability, its workload by study, or its performance by sponsor and therapeutic area. A site that cannot see which studies make money cannot fix its mix; it can only feel busy. Closing that visibility gap was a founding reason we built our own platform instead of licensing one.
What financially durable sites do differently
The defenses are unglamorous. Document standard costs and fees before any sponsor asks, in a master fee schedule, so every negotiation starts from your numbers instead of reconstructing them; sites that do this cut weeks from budget finalization. Negotiate payment terms as hard as fee amounts: startup paid at startup, holdbacks challenged rather than absorbed, record retention priced for the twenty-five-year reality rather than the boilerplate. Give invoiceable capture an owner and a routine, so the budget's conditional items flow into bookkeeping instead of memory. And review per-study profitability while the study is running, when the mix can still be steered.
All of it rests on one precondition: the budget has to be right in the first place. A study priced below the site's real cost structure cannot cash-flow, no matter how well it is administered afterward, which is why the discipline starts with a line-by-line budget model before signature, not with collections after. For sites that want that review done independently, it is a service we provide.
Research sites are the load-bearing layer of clinical development, and the industry prices them as if they were infinitely resilient. They are not. Six months of cash is not a business model. It is a countdown, and the sites that endure are the ones that stopped treating finance as paperwork and started treating it as an operational discipline equal to the science.