The pilot went well. The clinicians loved the product. The department head sent an enthusiastic email saying she wanted it across the whole service. The founders updated their investor deck with a new logo.
Then the contract went to procurement, and nothing happened for eleven months.
Anyone who has sold to hospitals will recognise some version of this story. It is usually told as a tale of bureaucracy: slow committees, endless forms, risk-averse buyers. That telling is satisfying and mostly wrong. Procurement is not failing health startups. It is doing exactly what it was built to do. Most startups simply never designed their product, their pricing or their evidence for the people who actually sign.
The buyer is not the user
Health startups are usually founded by people who understand a clinical problem. They build for clinicians, test with clinicians and collect enthusiasm from clinicians. That is the right way to build a good product. It is the wrong way to understand who buys it.
In most hospitals, the clinician who wants a product does not hold the budget. The decision passes through some combination of a value analysis committee, the IT and information security team, legal, finance and procurement. Each of them has a veto and a different definition of risk. The clinician champion who loved the pilot may never be in the room.
In the United States, the purchasing landscape adds another layer. According to industry figures cited by the Government Accountability Office, somewhere between 96% and 98% of US hospitals buy through group purchasing organisations. These GPOs negotiate contracts on behalf of member hospitals and, as GAO reported in 2014, are funded mainly by administrative fees paid by vendors, which under federal safe-harbour rules must either be 3% or less of the purchase price or be specified in the GPO’s agreement with its members. A startup that is not on the relevant GPO contract can find that even an eager hospital has no easy route to buy from it.
Procurement is designed to compare like with like
Here is the core problem. Public and institutional procurement exists to spend money fairly, defensibly and at a good price. It does that by defining a category, writing a specification and comparing suppliers against each other.
That works well for gloves, syringes and infusion pumps. It works badly for a product that does not fit an existing category. If your software replaces part of a nurse’s workflow, reduces readmissions and needs integration with the electronic record, which budget does it come from? Which category is it in? Who are the comparators? When procurement cannot answer those questions, the safest option is to do nothing. And doing nothing is almost never punished.
In the UK, the Procurement Act 2023, which took effect for new procurements from 24 February 2025, was meant partly to make public buying more flexible, including a new competitive flexible procedure. That may help at the margins. It does not change the basic incentive: buyers are accountable for spending public money correctly, not for adopting innovation.
The uncomfortable incentive
Put yourself in the procurement lead’s position. If they approve a new supplier and it goes well, the clinicians get the credit. If it goes badly, with a data breach, a failed integration, a supplier that goes bust, the procurement lead gets the blame. The rational response is caution.
Startups add to that caution without meaning to. Many are loss-making, venture-funded and one bad funding round away from closure. From a buyer’s point of view, a young company is a continuity risk as much as an innovation opportunity. A hospital that builds a clinical pathway around a startup’s software needs to know it will still exist in three years.
So procurement’s scepticism is not irrational. It is a reasonable response to a real risk that most startups never directly address.
There is also a quieter asymmetry. The savings a startup promises usually land in someone else’s budget. A tool that shortens length of stay frees bed days, which matters to the chief operating officer. A tool that prevents readmissions may, under some payment arrangements, actually reduce the hospital’s income. The cost, meanwhile, lands on a specific line that a specific manager defends every quarter. When benefits are diffuse and costs are concentrated, the person holding the concentrated cost will always ask for more proof.
What eleven months actually costs
It helps to put rough numbers on the stall, because founders tend to feel it as frustration rather than as arithmetic. The figures below are illustrative, not drawn from any particular company.
Take a seed-stage startup with twelve staff and a monthly burn of around $250,000. It has eighteen months of runway. Its first enterprise contract, if signed, is worth $300,000 a year. The pilot ends in month four. If procurement takes eleven more months, the contract starts in month fifteen, leaving three months of runway to show the revenue that the next round depends on.
Now look at the same deal from the other side. For the hospital, the contract might be a fraction of one percent of its non-pay spend. Waiting another quarter costs the buyer almost nothing. For the startup, the same quarter is a sixth of its remaining life. That mismatch in clock speed is the real story behind most “procurement killed us” post-mortems. The buyer is not being cruel. It simply does not share the seller’s deadline, and nothing in its incentives asks it to.
The practical conclusion is that a founder’s runway plan should assume the slowest plausible close, not the champion’s optimistic estimate. If the business only survives on the fast path, it is not a business yet. It is a bet on one department head’s influence.
Designing for the signature
The startups that get through procurement tend to have done a few unglamorous things early.
They mapped every veto holder before the pilot started. They know who will need to approve the contract, what each person cares about and which documents each will ask for. That usually means meeting the information security lead in week one, not month six.
They priced to fit a budget line. A product priced per user per month may fit an IT budget. One priced per patient episode may fit a clinical service budget. A price that straddles both, or that requires a new line altogether, can stall a deal indefinitely while two directors argue over whose money it is.
They built the security and integration file in advance. Information governance and security reviews are a frequent place for deals to stall. Having standard documentation ready before anyone asks removes one of the easiest reasons to wait.
They designed the pilot to answer procurement’s questions. Clinician satisfaction is nice. Procurement wants cost, measurable outcome change, implementation effort and what happens when the pilot ends. A pilot with a pre-agreed success threshold and a pre-drafted contract that triggers when the threshold is met turns a vague “let’s see” into a decision with a date.
And they addressed continuity risk openly. Escrow arrangements, data export guarantees and clear exit terms reassure a buyer who is worried the company might not survive. Saying “here is exactly what happens to your data and your pathway if we fold” sounds like an admission of weakness. To a procurement lead it reads as maturity.
None of this is exciting. All of it is the difference between a pilot and a contract.
The strongest objection
The best counterargument deserves a fair hearing. It runs like this: some procurement really is dysfunctional. Frameworks are renewed on multi-year cycles, so a startup that misses a window may wait years for the next one. Requirements get copied from old tenders and quietly exclude new approaches. Committees meet monthly and defer items they have not had time to read. No amount of founder preparation fixes a process that does not have a door.
That objection partly holds. There are buyers whose processes add delay without adding protection, and it would be glib to tell founders the fault is always theirs. But notice what the objection implies. If the door is structurally shut, the answer is not to push harder on it. It is to find a different door: an existing framework held by a larger partner, a GPO contract, a reseller relationship with an incumbent vendor, or an EHR vendor’s app marketplace that already sits inside the hospital’s approved supplier list. Each of those gives up margin and some control. Each also borrows someone else’s procurement history. For many startups that is a better trade than burning a year on a direct route that may never open.
Second-order effects
The procurement filter shapes the market in ways that are easy to miss.
It favours incumbents, obviously, but it also favours a particular kind of startup: one that sells a narrow product into an existing budget category, looks like something the buyer already purchases, and can plug into an established framework. Genuinely new categories, the ones that might change a care pathway, face the hardest route. That means the procurement system quietly selects for incremental innovation.
It also pushes consolidation. When selling directly is slow, being acquired by a company that already has hospital contracts becomes attractive. Some of those acquisitions keep the product alive. Others fold it into a larger suite, where it loses its sharpest edges.
And it changes what investors fund. If direct hospital sales take years, capital drifts towards models that avoid the hospital buyer altogether: selling to employers, to insurers, or directly to patients. Some of those models are good. Some exist mainly because the hospital door was too expensive to open.
What each side can do
For hospital CEOs and chief financial officers, the useful move is to create a route with a clock. A small innovation budget with a defined decision window, and a named owner for new-category purchases, gives procurement permission to say yes without inventing a process each time. The point is not to lower the bar. It is to make the bar visible.
For procurement leads, publishing what a complete supplier file looks like, including security questionnaires and data protection templates, costs little and removes weeks of back-and-forth. Buyers who do this tend to get better-prepared suppliers.
For ministries and national health agencies, the lever is shared assurance. When every hospital runs its own security and clinical safety review of the same product, the system pays for the same due diligence dozens of times. A national or regional assessment that local buyers can rely on shortens the path for everyone and concentrates scrutiny where it is most competent.
For clinicians who champion a product, the job is to become a translator. Bring procurement the numbers it needs rather than the enthusiasm it discounts.
For founders, the message is the one this whole piece has been arguing. Treat the signature as a product requirement.
The takeaway for investors too
Investors share some of the blame. Pilot counts and logo slides make for good pitch decks, and they are often rewarded as evidence of traction. But a pilot is not a sale, and a portfolio full of pilots that never convert is a portfolio headed for trouble. Better questions at the diligence stage are simple: how many pilots converted to paid contracts, how long did each conversion take, and who signed?
What to watch
Over the next year or two, a few signals will show whether the picture is shifting. In the UK, the first full cycles under the Procurement Act will reveal whether the competitive flexible procedure is actually used for new categories or left on the shelf. In the US, watch whether GPOs expand dedicated innovation contracts for digital products, and whether hospital systems keep building their own venture and innovation arms, which often function as a side door around standard purchasing. Watch too for how buyers treat AI tools, which raise new questions about liability, monitoring and validation that existing procurement templates were not written to answer. The first buyers to publish a clear assessment route for those tools will attract the best suppliers.
Procurement is not the obstacle standing between a good product and the patients who need it. It is part of the market. Startups that treat it as a design constraint, from the first version of their product and pricing, will sell. Those that treat it as paperwork to be endured after the clinicians say yes will keep collecting enthusiastic emails and very few signatures.
If you work in hospital procurement, what is the one thing startups consistently get wrong when they come to you? We would love to hear from the buyer’s side in the comments.
