The MedTech startup paradox is this: you need product-market fit before you commit to a regulatory path, because regulation locks your product shape for years — but you cannot legally ship a medical device to customers until that same regulation is satisfied. Classic SaaS discovery (build rough, ship, iterate with paying users) is illegal. Classic regulatory work (build the file first, then sell) bankrupts you before you know if the market wants the thing. The resolution is not picking one side. It is running a disciplined parallel track: verify the market without shipping the technology, recruit clinical partners on Day 1, and stage the project in two phases so every regulatory commitment is made after the market has spoken.

By Felix Lenhard and Tibor Zechmeister. Last updated 10 April 2026.

TL;DR

  • In MedTech you need PMF before regulation and regulation before shipping — a paradox that breaks founders who try to solve it with a SaaS playbook.
  • You cannot test the technology with customers before CE marking, but you CAN verify the market — through interviews, letters of intent, buyer economics, and a rolodex of friendly clinicians.
  • Clinical partners belong on Day 1, way before you sell. They are the customer-discovery channel the Regulation itself sanctions via Articles 62–82 and Annex XV.
  • The business model must check out BEFORE regulatory work starts. If the hospital will not pay for it and the numbers do not work, a Notified Body certificate does not save you.
  • The paradox resolves through a two-phase approach: Phase 1 verifies the market and locks intended purpose; Phase 2 builds the certifiable device against a known buyer.
  • If your product is not 10–20x better for the BUYER, not just the user, do not start the MDR project. The Regulation is too expensive to burn on an undifferentiated device.

The 18-month silence

Felix once spent a year and a half begging a founder to go out and talk to customers. Just talk. Not sell, not pitch, not close — talk. The founder had a product idea, a technical plan, and a pitch deck. What the founder did not have was a single conversation with a buyer. Not one. Felix kept asking. The founder kept finding reasons. The product needed to be further along. The market was not ready. The conversations would come "when we have something to show."

Eighteen months of silence. No revenue. No letters of intent. Nothing to put in front of an investor that said a human being on the buy side had ever seen this product and wanted it. The company dissolved. Runway gone, confidence gone, credibility gone. The technology may have been fine — nobody will ever know, because nobody was ever asked.

Felix can name at least four similar stories off the top of his head. This is not an anomaly. It is a pattern, and in MedTech the pattern is worse because the founder has a tempting excuse hardwired into the domain: "I cannot test with customers because of the Regulation." That excuse is half true and entirely fatal. It is the trapdoor that the MedTech startup paradox opens underneath first-time founders, and the ones who do not recognise it fall through.

The paradox stated plainly

Here is the paradox in its cleanest form.

Side one: you need PMF before regulation. Commit to a regulatory path before you know the market wants your device and you lock yourself into a product shape that will have to carry the full weight of the MDR for years. Intended purpose, classification, clinical strategy, QMS scope — all of it solidifies around a guess. If the guess is wrong, you do not get to pivot cheaply. You rebuild the file, which means you rebuild the company.

Side two: you need regulation before you can ship. A medical device as defined in Regulation (EU) 2017/745, Article 2(1), cannot be placed on the EU market without conformity with the Regulation. The SaaS move of shipping an ugly first version to paying users and learning from their behaviour is not available to you. If the thing you are building meets the Article 2(1) definition, shipping it without a CE mark is not a lean experiment. It is an illegal placing on the market.

Most first-time founders see one side of the paradox at a time. The technical founders see side two and freeze — they build for years in the lab because "we can't ship until we are certified." The commercial founders see side one and drift — they keep avoiding regulatory commitments because "we need PMF first." Both groups run out of runway. The first group runs out waiting for certification of a product nobody ordered. The second group runs out waiting for PMF signals they are not allowed to collect.

The resolution is that both sides of the paradox are true at the same time, and the plan has to honour both.

Why the SaaS playbook breaks in MedTech

The standard startup discovery loop — build, measure, learn — assumes you can put your actual product in front of actual users, observe what they do, and iterate. In MedTech that loop is blocked at almost every step.

You cannot put the actual product in front of actual patients without a clinical investigation under Articles 62–82 of the MDR, with the technical and ethical requirements in Annex XV. (Regulation (EU) 2017/745, Articles 62–82, Annex XV.) Clinical investigations are not rapid iteration. They are structured, ethically reviewed, competent-authority-notified studies that take months to set up and have nothing in common with the "deploy on Friday, measure on Monday" rhythm of a SaaS team.

You cannot put the actual product in front of actual hospitals for commercial use without satisfying Regulation (EU) 2017/745 in full — the placing-on-the-market bar under Article 5 is not negotiable, and "we're just testing" is not a category the Regulation recognises for a commercially-intended device. Pilot studies that are actually disguised commercial use are one of the red lines that end companies and careers.

You cannot safely "fake it until you make it" by building a lookalike front-end and pretending the backend works, because the whole point of a medical device is that the backend has to work on real physiology under real clinical conditions. A mock-up that hides the risk is a mock-up that hides the thing you most need to learn.

The SaaS playbook works because the cost of being wrong is one sprint. The MDR playbook does not forgive that kind of wrong — the cost is a Notified Body fee, a clinical investigation, and a year of runway. The first thing a first-time MedTech founder has to do is accept that the playbook they read about on technology blogs was written for a different game.

The key principle: you cannot test the tech, but you CAN verify the market

This is the line Felix comes back to every time he coaches a MedTech founder: you cannot test the technology with customers, but you can verify the market. Those two things are not the same, and mixing them up is where the 18-month silence starts.

Testing the technology means putting the device in a patient or clinician's hand and measuring what it does. That is clinical investigation territory, and the Regulation is clear about how it happens.

Verifying the market does not require the device to exist yet. It requires you to find the people who would buy the device if it existed, and get them to tell you — on the record — whether they would, how much they would pay, and what they would stop using to pay for it. Every one of these activities is completely legal, completely free of regulatory burden, and completely ignored by the founders who get stuck.

Market verification in MedTech looks like this:

  • Buyer interviews with hospital procurement, clinic directors, and department heads — the people who actually sign the purchase order.
  • User interviews with the clinicians and staff who would use the device — separately from the buyer interviews, because the buyer and the user are not the same decision-maker.
  • Letters of intent and written endorsements from clinical sites, collected before a single euro is spent on regulatory work.
  • Unit economics spreadsheets that the founder can defend in front of a sceptical CFO: how much time the device saves, how many patients per day, what that time is worth in the hospital's labour and throughput math.
  • A rolodex of friendly clinicians who have committed, in writing or at least in a warm email, to participate in the eventual clinical evaluation or investigation.

None of this is shipping. None of this is "placing on the market." All of this is allowed, all of this is cheap, and all of this gives you a defensible answer to the only question that matters before you commit to the MDR spend: does the buyer want this?

The bar Felix uses is blunt. If the product is not 10–20x better for the BUYER — not just the user, the buyer — do not build it. Understand to the last detail: how much time it saves, how many patients per day it improves, what the daily hospital impact is. If those numbers do not embarrass the alternative, the Regulation is too expensive to be worth it. The business model must check out BEFORE regulatory work starts, not after.

Clinical partners on Day 1 as the PMF path

The single highest-leverage move a first-time MedTech founder can make is this: get clinical partners on board on Day 1, way before you sell anything.

Clinical partners are hospitals, clinics, and individual clinicians who agree early to work with you on developing and evaluating the device. They are not customers yet. They are the bridge between "I have an idea" and "I have evidence a buyer will pay." They serve four functions at once, each of which is irreplaceable.

Function one: reality check. A clinician who will actually use the device will tell you, within one conversation, whether the workflow fits their day. Most lab-grown MedTech products fail this test in the first fifteen minutes of the first real conversation. Finding out early is the entire point.

Function two: clinical evaluation input. MDR Article 61 and Annex XIV lay out the clinical evaluation requirements — the evidence that the device delivers the intended clinical benefit at acceptable risk. (Regulation (EU) 2017/745, Article 61.) Clinical partners you recruited early are the people whose judgment, literature access, and eventual participation will shape that evaluation. Recruiting them late is expensive and sometimes impossible.

Function three: clinical investigation readiness. If your path includes a clinical investigation under Articles 62–82, the partners who will host it need to be identified, aligned, and ethically engaged long before the investigation starts. Finding investigation sites while the runway burns is one of the most common failure modes of underprepared Class IIb and Class III startups.

Function four: first friendly customers. Clinical partners who participated in development and in the clinical investigation become the first commercial customers the moment certification lands. They already know the device, they have seen it work, and they are advocates inside the hospital's procurement process. The founders who recruit clinical partners early end the certification project with a pipeline already warm. The founders who skip this step finish certification with a certificate and zero customers — and then burn the next year trying to build the sales motion they should have built during development.

This is the bridge across the paradox. Clinical partners are your customer-discovery channel, but they operate inside the regulatory frame the Regulation itself recognises. They are both PMF signal and clinical evidence path. They satisfy both sides of the paradox at once.

The wellness / non-device beachhead as a PMF probe

There is a second move that every first-time founder should evaluate honestly: the beachhead strategy. If there is a legitimate version of your product that is NOT a medical device — a wellness product, a workflow tool, a data product, an unregulated adjacent thing — shipping that version first lets you generate real revenue, real usage data, and real market signal without touching the MDR at all.

This is not a loophole. It is a legitimate strategic choice, and it is allowed only when the intended purpose and claims of the non-device version genuinely place it outside the Article 2(1) definition of a medical device. The moment you claim diagnostic or therapeutic benefit, even in a blog post or a sales deck, the wellness framing collapses and you are back inside MDR scope.

Done right, the beachhead is a PMF probe with teeth. You find out whether anyone will pay for the product before you commit to the regulated version. You build brand, channels, and customer understanding. And when you eventually cross into the regulated version, you do it with a real user base telling you what matters. See the beachhead strategy post and the wellness first, medical device later post for the details on how this gets executed without tripping the Regulation.

A short counterpoint for the optimistic founder: the EUR 50K first quarter Felix coached in an unregulated education space is the cleanest positive PMF example in his notebook. Landing page, customer conversations, iteration, then the product built after people had already paid. Fifty thousand euros in the first quarter, zero marketing spend, because the founder had answered the one question buyers actually had. The lesson translates to MedTech, but the translation is strict: in an unregulated domain you can ship the answer and learn. In MedTech you can only ship the verification work — the conversations, the letters, the clinical partnerships — and defer the actual product shipping until the regulatory file is ready. The discipline is the same. The surface is different.

The two-phase approach as the paradox resolution

The two-phase development approach is the structural answer to the paradox. It splits the project into a phase that honours the PMF side and a phase that honours the regulatory side, and it puts a real gate between them.

Phase 1 — Exploratory. Intended purpose is not yet locked. Classification is not yet committed. The team is doing market verification, running buyer and user interviews, recruiting clinical partners, building the business-model case, and avoiding any design decision that would force an irreversible regulatory commitment. You are not building the MDR file in Phase 1. You are building the evidence that the MDR file will be worth building.

Phase 2 — Committed. Intended purpose is written and defensible. Classification is decided and traceable to Annex VIII. You are now building the device you will certify, not the device you might certify. Every design decision runs through the regulatory lens. The QMS starts building up. Risk management starts capturing every hazard. The technical file grows with the product, not after it.

The gate between the phases is the thing most founders skip. You need a deliberate, explicit decision — based on the Phase 1 evidence — that the market wants this product, that the buyer economics work, and that a specific intended purpose and classification are worth committing to. That gate is where the paradox resolves. Before the gate, you are allowed to be flexible and you are not allowed to ship. After the gate, you are allowed to commit heavy regulatory spend and you are no longer allowed to drift.

See the two-phase development approach post and when to start MDR regulatory work for the operational version.

The Subtract to Ship angle

Subtract to Ship applied to the paradox looks like this: every activity in Phase 1 has to earn its place by producing a signal about whether the market wants the product. Every activity in Phase 2 has to earn its place by tracing to a specific MDR article, annex, or harmonised standard obligation. Activities that fit neither are waste, and waste in a first-time MedTech project is what converts a manageable runway into a death spiral.

The founders who spent eighteen months in silence were not doing the wrong work. They were doing no work on the question that mattered. The founders who spent EUR 1.8 million on a product that worked in the lab and failed in the real world were doing very hard work on a question the market never asked. Both failures are the same failure: activity that did not trace to a signal or to a requirement. Cut it. Do the work that does.

Read the Subtract to Ship framework for MDR post for the full methodology.

Reality Check — Where do you stand?

  1. Have you talked to at least ten actual buyers — the people who would sign the purchase order, not the users — in the last ninety days? If not, what is your excuse, and does it survive reading this post?
  2. Do you have a written, defensible number for how much time, money, or throughput your device saves the buyer per day? Would a sceptical hospital CFO accept the number?
  3. Do you have at least two clinical partners who have agreed, in writing or at least in a warm committed email, to work with you on development and eventual evaluation?
  4. Have you mistaken "building the technology" for "verifying the market" at any point in the last six months? If yes, how much of the runway went to the wrong question?
  5. Is your current intended purpose locked, or is it still flexible? If locked, can you defend it against the Phase 1 evidence you have gathered — or did you lock it before gathering any?
  6. Have you evaluated whether a non-device or wellness version of the product could be your PMF probe, or did you skip straight to the regulated version?
  7. If you had to cancel the MDR project tomorrow, how much of the Phase 1 learning would survive as a real asset? If the answer is "none," Phase 1 is not happening; something else is.

Frequently Asked Questions

Can I run a clinical investigation as a customer-discovery tool? No, not in the SaaS sense. A clinical investigation under MDR Articles 62–82 and Annex XV is a structured, ethically reviewed study with a specific scientific aim, conducted under a detailed plan and approved by a competent authority and an ethics committee. It is not a rapid-iteration discovery channel. It is the formal evidence-generation step for a device whose intended purpose is already decided. Use buyer interviews, clinical partner conversations, and letters of intent as your discovery channels — not investigations.

Why can't I just ship an early version and fix it later like a SaaS company? Because a medical device under MDR Article 2(1) cannot legally be placed on the EU market without conformity with the Regulation. Shipping an unfinished version to paying users is an illegal placing on the market, not a lean experiment. The Regulation does not have a "beta" category for commercially-intended devices. If you try to ship early, you are not doing lean MedTech — you are doing the kind of thing that ends companies and careers.

Is MedTech actually harder than SaaS or just different? Both. It is harder because the regulatory cost of being wrong is orders of magnitude higher, the feedback loops are longer, and the capital requirement is bigger from Day 1. It is different because the discovery channel is clinical partners and buyer interviews, not A/B tests and cohort analytics. Founders who try to run MedTech on a SaaS playbook fail predictably. Founders who accept the difference and run the two-phase approach survive. See why MedTech needs more capital than SaaS for the financial side of the same argument.

When should I lock my intended purpose? At the Phase 1 to Phase 2 gate, not before. Before the gate, the intended purpose is a working hypothesis you are refining based on buyer and clinician feedback. At the gate, you commit — and the commitment means every downstream activity in the MDR project is tied to that specific intended purpose. Locking it before the gate is how first-time founders end up building regulated products nobody asked for.

What if my Phase 1 evidence says the market doesn't want it? Then the two-phase approach just saved you the cost of Phase 2, which in MedTech is the expensive part. Walk away, pivot, or restart with a different product idea. The founders who most regret their MDR projects are the ones who kept going into Phase 2 despite Phase 1 evidence that was telling them to stop. Phase 1 exists precisely to give you permission to stop.

Sources

  1. Regulation (EU) 2017/745 of the European Parliament and of the Council of 5 April 2017 on medical devices, consolidated text. Specific references: Article 2(1) definition of medical device; Article 61 clinical evaluation; Articles 62–82 clinical investigations; Annex XV clinical investigations technical and ethical requirements. Official Journal L 117, 5.5.2017.
  2. Regulation (EU) 2023/607 of the European Parliament and of the Council of 15 March 2023 amending Regulations (EU) 2017/745 and (EU) 2017/746 as regards the transitional provisions for certain medical devices and in vitro diagnostic medical devices. Official Journal L 80, 20.3.2023.

This post is part of the MedTech Startup Strategy & PMF series in the Subtract to Ship: MDR blog. Authored by Felix Lenhard and Tibor Zechmeister. If you are somewhere in the eighteen-month silence and you recognise yourself in it, the right next step is one buyer conversation this week. Not ten. One. Then another. That is the whole resolution to the paradox, written as a to-do list.