Back
Finding Product Market Fit in the Lausanne Startup Ecosystem
SmartInterview Team
The Short Answer
Finding product market fit in Lausanne is shaped by three local realities: much of the startup base comes out of EPFL research, so the search is often for an application rather than for a solution to a known problem; the Swiss home market is small enough that early traction can be misleading; and the country's language borders mean validation in Suisse romande does not automatically transfer to Zurich.
Decide early: validate in Suisse romande and expand later, or design the study for a larger market from the first interview. Both work. Drifting between them does not.
Deep tech inverts the usual order. You start with a capability and go looking for the customer, which means running several application hypotheses in parallel instead of iterating on one.
The support structure is unusually dense. EPFL's transfer and startup services, EPFL Innovation Park, Innovaud, Innosuisse and Venture Kick cover coaching, space and early non-dilutive funding.
Watch the sales cycle trap. Swiss corporate buying is slow and relationship-heavy, so a pilot can absorb a year of runway and still tell you nothing about demand.
What the Local Structures Actually Give You
Suisse romande has an unusually complete support stack for early companies, but the parts do different things. Most of it helps you build and fund. Very little of it validates demand for you, which is the thing founders most often assume is covered.
Organisation | What it is | What it gives you at the validation stage | What it does not do |
|---|---|---|---|
EPFL technology transfer and startup services | The university's own route from lab result to company, covering IP, licensing and an official spin-off label | A clean legal starting point, credibility with investors, and access to people who have taken the same path | It does not tell you which market wants the technology. Licensing terms say nothing about demand |
EPFL Innovation Park | A technology park on the EPFL campus. It describes itself as a deep-tech hub hosting startups alongside large companies and corporate R&D units | Space, proximity to labs, and useful density: corporate R&D teams on the same site are reachable first customers or design partners | Proximity to a corporate R&D unit is not proximity to its operating business, which is where budget lives |
Innovaud | The canton of Vaud's innovation and economic promotion agency, connecting companies to coaching, financing and hosting | Orientation across a confusing landscape, and introductions to coaches and cantonal instruments you would not find alone | It is a connector, not a customer. Introductions accelerate a validated thesis rather than produce one |
Venture Kick | A philanthropic funding initiative for startups coming out of Swiss universities, structured in three stages with a jury at each step | Non-dilutive early money and, more usefully, forced deadlines: you have to defend the business case in front of people with no stake in it | Winning a stage is validation by a jury, not by a market. The two correlate loosely |
Innosuisse | The federal innovation agency, funding innovation projects run with a research partner and offering startup coaching | Funding for the technical de-risking that deep tech needs before any customer can realistically evaluate it | It is oriented toward innovation projects with research partners, so it can reinforce a technology-first path if you let it |
The pattern is consistent. The ecosystem is very good at getting a technically strong company to the point where it can be evaluated, and largely silent on whether anyone wants to buy it. That gap is yours to close, and the general method for closing it is in how to find product market fit without fooling yourself.
The Decision Nobody Makes Explicitly: Local First or Large Market First
Switzerland has roughly nine million people, and French-speaking Switzerland accounts for around two million of them. For most B2C products and a good number of B2B ones, that is not a market you can build a venture-scale company on. It is, however, big enough to produce encouraging early numbers, which is exactly what makes it dangerous.
Every Lausanne founder faces the same fork within the first year, and most of them never decide it deliberately. They default to selling to whoever is nearby and call it strategy eighteen months later.
The case for validating locally first
Proximity is a real advantage while you are still learning. You can meet customers in person, iterate weekly, and reach decision-makers through two introductions in a network where reputation travels fast. Swiss buyers, once they commit, tend to stay, which gives you clean retention data uncontaminated by churn from casual signups. For deep tech in particular, having your first pilot within an hour of the lab matters more than the size of the local market.
The condition is that your local segment has to be representative of the larger one you intend to enter. Validate a compliance product against Swiss regulation and you have learned about Switzerland. Validate a lab instrument against a Lausanne research group and the finding probably holds in Boston, because the users are the same species.
The case for targeting a larger market from day one
If your product is software with no local regulatory dependency, validating in Suisse romande can build a company shaped around a market that is a rounding error in your eventual plan. Pricing calibrated to Swiss purchasing power does not survive contact with Germany or Spain. Positioning tuned to a market where everyone knows everyone does not survive contact with one where nobody has heard of you.
Practically, targeting large means running discovery in your target market from the beginning, which means recruiting respondents you have no personal network to reach, in a language you may not work in daily. That is a real cost and it is smaller than the cost of discovering in year three that your first hundred customers were unrepresentative.
How to decide
Is the problem Swiss-specific? If it involves local regulation, Swiss healthcare structures, cantonal administration or the Swiss financial sector, local validation is the real thing. If not, the local market is a convenience sample.
Would your local buyer look like your target buyer abroad? Research groups, hospitals and industrial engineers translate reasonably well across borders. SMEs and consumers translate badly.
Does your price hold? Take the price a Swiss customer accepted and test it in your largest target market. If it collapses, your unit economics were validated on a market you will not be selling to.
Can you reach the foreign segment at all? If you cannot recruit twenty German or French buyers for interviews now, you will not sell to them later. Test that route early, while the cost of finding out is a few weeks.
Three Local Frictions That Distort Your Signal
Switzerland is not one market
A Lausanne startup that validates in French has validated in French-speaking Switzerland. Zurich, which holds the larger share of Swiss corporate budgets, is a different commercial culture with different buying habits, different reference customers, and a different language of business. Founders regularly discover that the Suisse romande playbook that worked, warm introductions, French-language material, a relationship-led sale, produces almost nothing east of the Sarine.
The practical consequences are concrete. Your reference customers carry less weight in a region that has not heard of them. Your material has to exist in German, and machine-translated German reads as foreign to a Swiss corporate buyer. Sales cycles restart from scratch because the network you built does not extend. Treat Zurich as international expansion with the travel costs removed, not as the rest of your home market. What that market looks like from the inside is the subject of finding product market fit in the Zurich ecosystem.
If you are running any kind of structured research across the language regions, run it in all the relevant languages rather than defaulting to English. English-only instruments in Switzerland systematically over-sample younger, more international, more urban respondents, which is precisely the group least representative of a corporate buyer in Ticino or a retail consumer in Fribourg.
Swiss enterprise sales cycles absorb validation time
Selling to a large Swiss company is slow. There is usually a long evaluation, several stakeholders, a procurement process, and a strong preference for suppliers that already have a Swiss reference. Twelve to eighteen months from first meeting to signed contract is not unusual for anything touching core operations.
For a startup this creates a specific trap. You have a genuinely interested contact, a pilot in discussion, and a plausible reason to keep waiting. The whole thing consumes a year of runway and, at the end, a signature tells you that one organisation was willing to try something. It does not tell you the market wants it.
Two defences work. First, run several enterprise conversations in parallel, never one, so that no single account can hold your roadmap hostage. Second, define in advance what would count as a real signal inside a pilot: usage by the operational team rather than by the innovation unit, a named budget for the following year, or a second department asking for access without you prompting it.
The innovation-budget false positive
Swiss corporates, and the ones with R&D presence around the EPFL campus in particular, run innovation programmes with budgets specifically for trying things. Getting funded by one is genuinely useful: it is money, a logo and a real deployment. It is also the single most common source of false positives in the region, because the buyer's motivation is to explore, and exploration budgets renew regardless of outcome.
The test is whether the paying party is the operating business. When the department that would actually use the product pays out of its own budget, that is a purchase. When innovation pays, that is a subsidy with a logo attached, and you should size it accordingly in your own forecasts.
Deep Tech PMF: Finding the Application, Not the Feature
A large share of Lausanne companies start from a research result. The founder has a capability, often a genuinely strong one, and no specific customer. This inverts the normal validation sequence, and applying standard startup advice to it produces bad outcomes.
In a conventional startup, you have a problem and you iterate toward a solution. In a research spin-off, you have a solution and you are searching for a problem worth solving with it. The search space is wide, the cost of exploring one branch is high, and the technology's own elegance is a constant distraction from the question of who is suffering enough to pay.
Run applications in parallel, not in sequence
The common failure is committing to the first plausible application, usually the one suggested by the research context, and spending two years on it. Treat applications as a portfolio instead. List every domain the capability could serve. For each, spend a bounded amount of time, a few weeks, on cheap disqualification: is there an existing budget line, who owns it, what do they use today, what would have to be true for them to switch, and what regulatory path applies.
Most branches die quickly and cheaply on one of those questions. That is the point. You are not looking for the application you like best, you are looking for the one where the pain is acute, the buyer is identifiable, and the path to deployment is shortest.
Different questions from a normal discovery interview
Deep-tech discovery has to cover ground that software discovery does not.
What is the incumbent process, precisely? Not "how do you do this" but what equipment, what throughput, what error rate, what it costs per unit. You need numbers because your value proposition is a comparison.
What performance threshold changes the decision? Many technical improvements are real and irrelevant. Ask what level of improvement would justify requalifying a process, because below that line nothing happens regardless of how good the science is.
What is the cost of switching? Recertification, retraining, revalidating a production line, requalifying a supplier. In industrial and medical contexts this frequently dominates the price of the product itself.
Who has to approve it, and what is their timeline? Regulatory and procurement gates determine whether your addressable market is reachable within your runway.
These conversations are best run in the buyer's own language and in depth, which is expensive when you need them across several candidate applications and several countries. This is where running discovery at scale earns its place: platforms like SmartInterview exist to put the same open questions to a few hundred practitioners in their own language and probe each answer with a follow-up, so you can compare three application hypotheses against real practitioners instead of against your own intuition. It complements the twenty conversations you should still run personally, described in how to validate a product idea with real customers.
If you would rather have a specialist run the fieldwork, the region has an established research sector: see the main market research companies in Switzerland and, for the local providers specifically, market research agencies in the Lausanne region.
Frequently Asked Questions
Is the Swiss market too small to validate a startup in?
It depends on whether your local customers resemble your eventual ones. For products tied to Swiss regulation, Swiss healthcare or the Swiss financial sector, local validation is the real thing. For general software, French-speaking Switzerland is a market of roughly two million people and functions better as a convenient early sample than as proof of a business, particularly for pricing, which rarely survives the move to a larger market.
Does validating in Lausanne tell me anything about Zurich?
Less than founders expect. The language of business differs, the corporate culture differs, and your Suisse romande reference customers carry limited weight in a region that has not heard of them. Plan for Zurich as a separate go-to-market with its own references and German-language material, not as an extension of your home region.
What does the EPFL ecosystem actually help with?
Technology transfer, credibility, space, coaching and access to early non-dilutive funding through instruments such as Venture Kick and Innosuisse. What it does not systematically provide is evidence that a market wants what you built. That gap is the responsibility of the founding team, and the density of technical support around EPFL can make it easy to postpone.
How do I know if a corporate pilot is real demand?
Look at who is paying and who is using. If an innovation or R&D budget is funding it and the operating team is only lightly involved, treat it as exploration. Real demand shows up as usage by the team that owns the process, a budget line for the following year, and other departments asking for access without you prompting them.
Should I run customer research in French, German or English?
In the language your respondent works in. English-only research in Switzerland skews toward younger, more international and more urban respondents, which distorts both consumer and corporate samples. If you are studying more than one language region, run each in its own language and compare, because differences between regions are often the most useful finding in the study.
How long do Swiss enterprise sales cycles take?
For anything touching core operations, plan for many months and several stakeholders, with a strong preference for suppliers who already have a Swiss reference customer. The strategic implication matters more than the exact number: never let one enterprise conversation define your roadmap, and always run several in parallel so that a single slow account cannot consume a year of your runway.
