> ## Content Index > Fetch the complete content index at: https://icopilots.com/llms.txt > Use this file to discover other available public pages before exploring further. # πŸ‡ͺπŸ‡Ί How to build a European incubator in 2026 - #3 The selection problem - URL: https://icopilots.com/european-startup-incubation-selection-problem-first/ - Published: 2026-05-08T09:14:21.000Z - Updated: 2026-05-08T09:14:21.000Z - Description: Parts 1 and 2 argued for a network-anchored geographic model and a private consortium funding structure. Part 3 addresses what happens when founders actually walk through the door (which is, in practice, where most serious incubation programs tend to underperform). - Author: Philippe MEDA - Tags: Innovation Strategy, #european startup, #incubation model, #deepetch, #europe, #problem The standard European incubator model selects founders the way a theatre school selects students: based on the quality of a 20-min stage performance. To be fair, this also includes a compelling deck, some data points, and a founder who consistently hit a few important buzzwords. For consumer internet, this proxy is not entirely wrong (in B2C markets, the ability to tell a convincing story to a room of strangers is slightly predictive of the ability to acquire customers at scale). But for industrial deeptech and B2B infrastructure, it often selects the wrong people, and, more precisely, the wrong kind of thinking. So why does every serious incubator keep doing it? Mostly because nobody has agreed on what to replace it with, so they have to make their own onboarding quotas fast and consistently. The consortium model we discussed [in Part 2](https://icopilots.com/how-to-build-a-european-incubator-in-2026-2-minimal-public-funding/) implies a structural inversion of this logic. **When corporate investors define the problem space and the program's mandate, the selection process will be skewed toward measuring the fit between a founding team and a predefined industrial problem.** Shy, nerdy introverts can now apply. ## No, not a startup studio The obvious objection is that this sounds like a startup studio. The studio model (practiced by Entrepreneur First and BCG Digital Ventures, among others) does something adjacent: it recruits individuals before they have an idea, forms teams institutionally, and builds companies around problems the studio itself has identified. It has produced genuinely valuable companies. It also tends to produce founding teams with no authentic conviction in the problem they are working on, because the problem was handed to them rather than found. I would argue they are the real 'intrapreneurs' everyone is talking about. Genuinely entrepreneurial profiles, but on a contract. This usually leads to a different failure mode from bad pitch selection, but it's another discussion entirely. What the consortium model requires is distinct: **the industrial partners define the problem space, but with varying degrees of certainty.** From the ***known unknowns*** to the ***less-known unknowns***, and maybe a dash of ***unknown unknowns*** to spice things up (and increase optionality). In this approach, founding teams self-select into it competitively and retain full decision authority over how they solve it. ### Track 1: direct problems with procurement commitment **The first entry track is for startups willing to work on problems the consortium's industrial partners have listed for the current year** (specific, documented, and *budgeted*). The model for structuring this commitment is stolen from how Deutsche Bahn (DB), the German state-owned rail operator, ran its open-innovation incubator in Berlin back in the 2010s. Winning ventures entered a deliberately structured procurement relationship in which, after successfully exiting the incubation program, DB would commit to 80% of their revenue in year one, down to 50% in year two, and no more than 30% by year three. What makes this architecture interesting is that it creates a deliberate revenue dependency designed to decay at a rate that forces the startup to build an independent commercial base and learn more from the market, while the institutional revenue gives them enough runway to do it without existential pressure. The equity terms, IP rights, and exit conditions are specified [before selection begins](https://icopilots.com/startup-apply-corporate-incubation-program/), not negotiated once a team is already committed and the power asymmetry has become inescapable. Which is both tremendously important and vastly overlooked in most of these setups. This track looks most like a typical open innovation program, which means corporations will be fairly comfortable with it and can also reuse their own processes within it. ### Track 2: vertical problems with defined market pathways The second track is more open-ended and operates on a three-year horizon. The consortium defines problem areas at sector or capability level (sustainable packaging logistics in European food manufacturing, predictive maintenance in rail infrastructure, AI-assisted regulatory review in pharmaceutical development) and invites startups to compete within those verticals without specifying the solution approach. Here,the procurement commitment is lighter at the entry and heavier at the PoC milestone: the industrial partners are buying optionality rather than contracted output. **What they commit to upfront is a clear signal of what they would be willing to adopt if the solution works**, and a pre-agreed evaluation framework so the startup knows in advance what "working" means in commercial terms. The absence of this framework is, in my experience running programs of this kind, the single most common cause of technically promising PoCs that produce case studies but very few business contracts. And when I say 'absence', I mean there's an evaluation framework, usually even rightly so, at a PoC level, but it's a technical one, guided by TRL and R&D principles. In truth, the technology is rarely lacking; the business focus is. ### Track 3: open problems with asymmetric reward **The third track is for startups proposing problems outside the predefined verticals. Problems the industrial partners did not think to list, or did not yet know they had.** The risk profile is genuinely higher: less institutional cover to begin with, no guaranteed procurement pathway, and lighter capital allocation from the consortium at entry. In exchange, the consortium takes less equity on the logic that if the founding team has identified a real problem that major industrial players have missed, the value created upon validation belongs disproportionately to the people who saw it. But the track is here, and the industrials reckon they can't predict the future (I know, it's a cultural high bar I'm aiming for!). This track is where the program maintains its capacity for genuine surprise and maximizes [optionality](https://icopilots.com/what-is-the-roi-of-an-innovation-program/). It's the part that most consortium-funded programs quietly eliminate in the name of focus, and then wonder why the portfolio looks like if a junior manager had written a procurement catalog. ## Problem quality as the hard switch Here's the key part, though. **Across all three tracks, the selection criterion is the same: the quality of the problem statement, not the quality of the proposed solution.** How well is the problem documented? Is it scalable within a single industrial partner or across the consortium? Does the team have measurable indicators of the problem's current cost or frequency? Is it high-frequency and low-intensity, or low-frequency and catastrophic (because [these call for entirely different venture architectures](https://www.icopilots.com/the-only-four-problems-that-innovators-can-solve/?ref=icopilots.com))? The [quadrant logic we use to determine which problem type requires which strategic posture](https://www.icopilots.com/framing-the-problem-you-are-solving/?ref=icopilots.com) applies directly to entry evaluation here. The reason this criterion produces a cultural shock (and it does, reliably) is that the teams most likely to apply to an industrial deeptech program are engineers and PhDs whose entire professional formation has trained them to lead with the solution, or worse (from an innovation standpoint), with the science behind it. A team presenting a novel sensor array for detecting micro-fractures in rail infrastructure will enthusiastically spend 40+ minutes on the technology and a reluctant 4 minutes on the cost to a train operator when a micro-fracture goes undetected for 6 months. Inverting that ratio is the boldest and most effective move you can make to change your whole incubation logic. And it will be hardcore difficult. It is a fundamentally different way of thinking about what problem ownership means, and [precisely the reason ideation-first programs consistently produce technically impressive outputs that nobody buys](https://www.icopilots.com/the-problem-with-ideation/?ref=icopilots.com). Ask any innovation director who has sat through a demo day and tried to explain to their CFO why none of it converted. They know. This is the model we run with Merck KGaA at the moment. The German science and technology company, which has 68 projects in this year's program (approximately 300 team members), is submitting against a multi-layered canvas of needs ranging from tightly specified strategic priorities for the current year to deliberately open problem areas for the next three. Admittedly, it's not exactly where I would want it to be, but we're at 80% there, and the first season has been quite a resounding success, with selected teams capable of fluently speaking to ROI and customer pricing regarding their value add. Were the submitting teams happy initially? Not. In. The. Slightest. At first cut, selection evaluates only the problem statement: how well it is documented, whether it scales within the organization or beyond it, and whether the team has identified key metrics for the problem's current cost, frequency, and strategic weight. Teams of highly skilled technicians and managers find this at best disorienting. But here's a huge payoff. The ones who adapt quickly are, without exception, the ones who eventually build something worth scaling (which is, naturally, the only outcome a consortium of serious industrial investors should be prepared to accept). The key elements of this logic are mapped out here, step-by-step: [Value-Driven Innovation TrainingFrom R&D logic to solid business case in 3 weeks.![](https://static.ghost.org/v5.0.0/images/link-icon.svg)innovation copilotsPhilippe MEDA![](https://storage.ghost.io/c/8f/a3/8fa3eb46-8b94-4e7c-85af-c9fe45614239/content/images/thumbnail/photo-1614935151651-0bea6508db6b)](https://icopilots.com/value-driven-innovation-training/) --- ### More on this: [πŸ‡ͺπŸ‡Ί How to build a European incubator in 2026 - #1 Forget the cityMost public incubators and startup programs built in Europe over the last fifteen years share one structural flaw: they were designed as local answers, infamously trying to be the β€œnew Silicon Valley.” Worse, two cities 50 km apart would try to be the one. Forget the city, build a network![](https://static.ghost.org/v5.0.0/images/link-icon.svg)innovation copilotsPhilippe MEDA![](https://images.unsplash.com/photo-1607269910705-63c248ac5715?crop=entropy&cs=tinysrgb&fit=max&fm=jpg&ixid=M3wxMTc3M3wwfDF8c2VhcmNofDIwfHxldXJvfGVufDB8fHx8MTc3NTc0NjEzMXww&ixlib=rb-4.1.0&q=80&w=2000)](https://icopilots.com/how-to-build-a-european-incubator-in-2026-1-forget-the-city/) [πŸ‡ͺπŸ‡Ί How to build a European incubator in 2026 - #2 Minimal public fundingWhile pondering how I would build a proper startup incubator in 2026, a few principles immediately came to mind. The first one we already discussed was letting go of the local binding to a city. Trying to make Madrid or Berlin the next Silicon Valley is rather clueless in the![](https://static.ghost.org/v5.0.0/images/link-icon.svg)innovation copilotsPhilippe MEDA![](https://storage.ghost.io/c/8f/a3/8fa3eb46-8b94-4e7c-85af-c9fe45614239/content/images/thumbnail/photo-1546610072-90a8cdd6aa4d)](https://icopilots.com/how-to-build-a-european-incubator-in-2026-2-minimal-public-funding/)