The Research Phase: What the Lab Owes the Company
How Decisions Made Before Incorporation Impact the Entire Lab-to-Market Journey
Before the market – before the actual company – there is the lab.
It’s not always a prestigious one. It’s not always well-funded. Sometimes it’s nothing more but a shared space in a university basement, or a cluster of desks behind a national institute's main facility. But in almost every Deep Tech venture that has ever mattered, the starting point was a group of people, most of whom are highly educated, working on something with an uncertain outcome, asking questions with no commercially useful answers.
This is the Research phase of the Lab-to-Market journey. It sits at TRL 1 to 3, pre-MRL, pre-CRL. And it is a most consequential part of the Lab-to-Market journey.
For many, Lab-to-Market journeys start only once the group is incorporated, with early lab experience treated almost as a precondition for the journey, rather than the beginning of the actual journey itself.
But this may be wrong.
Assumptions made at TRL 1 and 2 can and often do impact every subsequent phase. Team dynamics that form around a research group shape what the company can eventually become. How questions are asked, how failure is treated, how honestly the gap between the science and what the world wants to hear is managed – all of this reveals the ultimate intellectual culture of the venture, and is often difficult to reverse.
This essay identifies five important considerations that Deep Tech research groups should address before incorporation.
1. The Structure of a Research Group
Most Deep Tech entities begin with a research group and not a founding team in the conventional startup sense. Structure at this early stage matters because it determines not only what gets discovered but who discovers it, how, and with what behaviours and/or habits of mind.
In their early stages, those groups that function well typically contain several distinct roles, each contributing to collective, intellectual work. These are generally led by a Principal Investigator, who provides scientific direction, institutional authority, and access to funding (via 3-5 year grant cycles). Post-doctoral researchers play key roles and carry a large share of the experimental work and lab management activity. PhD students bring the kind of deep, sustained attention that comes from spending years on a single problem. Masters students and undergraduates provide breadth and fresh questioning, including the occasionally naive perspective that challenges assumptions.
At its best, such a structure produces a genuine diversity of scientific perspective, which is very much the point. PI's are there to create the conditions under which great questions get asked and honest answers spring forth. The result is a dynamic human system, bound by curiosity and a quest for truth.
Such environments are optimised for actual discovery work, which is the intention of TRL 1-3, far away from any thinking about commercialisation or scalability. Being wrong during this early phase is expected, and in the best groups, genuinely welcomed.
But research group models are not company models. Therefore, those considering incorporation should think carefully about which parts of the structure translate well into a company and which do not. During the Research phase, early-stage funding comes from grants rather than target-driven investors. Success is generally measured in publications and peer recognition, and not commercial milestones. Time horizons are different, as are accountability structures and the tolerance for open-ended exploration. Not all of this translates easily.
But successful transitions do happen. The groups that navigate it best are those that are clear-eyed in their understanding of what is worth leaving behind versus what should be taken forward.
2. The Essential Scientific Method
The Scientific Method is generally taught as a procedure: hypothesis, experiment, observation, conclusion, repeat. At the frontier of Deep Tech research, it is also a commitment to prioritising evidence over belief.
While this sounds straightforward, it is not. Research groups are human systems, and human systems have a tendency toward confirmation bias. When a team has spent three years developing a hypothesis, securing funding on its basis, and publishing preliminary results that attract attention, the psychological and institutional pressure to find evidence confirming a hypothesis, rather than re-testing it, can be immense.
As such, countermeasures are important, the most important of which is the practice of actively seeking falsification. ‘What would have to be true for our hypothesis to be wrong, and have we genuinely tested for it?’ reflects the attitude required as part of hypothesis fortification. This may be a function of values and behaviour more than intellect, and is a harder discipline to consistently demonstrate, particularly when careers, funding, institutional reputation, and personal dreams are riding on particular outcomes.
The irony, of course, is that rigorous scientific method – the discipline that feels like it might slow progress – is actually what creates durable commercial value. Technologies that have been honestly and ruthlessly battle-tested at TRL 1 to 3 can arrive at TRL 4 and 5 more stable, with better tested assumptions, better understood failure modes, and a rigour that is missing from those that don’t.
It’s the difference between: ‘have we genuinely tried to break this? Or have we just been trying to make it work?’
3. When to Incorporate?
The case for staying unincorporated is stronger than many groups appreciate.
Incorporation introduces a set of obligations and incentive structures that can be actively harmful at TRL 1 to 3. Once a company exists, external stakeholder expectations begin to shape behaviour. Founders feel pressure to demonstrate progress in commercially legible ways. Equity structures create personal financial stakes that often distort scientific judgement. The company's needs begin to compete with the research group's needs, creating tensions that can fracture a team that was functioning well as a research entity. For those that are still genuinely in discovery mode – where the dominant question is ‘does this phenomenon exist and can we characterise it?’ rather than ‘can this phenomenon be harnessed for a specific commercial purpose?’ – staying unincorporated can often be the right choice.
The counterargument is that incorporation can create clarity and commitment that accelerates progress. With allocated equity, roles are formalised and teams can work with greater urgency and focus. External funding can also be secured on commercial terms, bringing with it useful networks and advice. Competitive pressures and dynamics from the market can also be a spur. But, from our experience, this makes the most sense only when the technology concept is sufficiently validated, and that the central remaining questions are more to do with engineering and application, rather than fundamental science.
Of course, there are domain considerations too. In Life Sciences and Bioengineering, regulatory pathway means that early IP protection can be critical, leading to earlier incorporation compared to other domains. In Quantum Computing, frontier science is generally so far from commercial application that incorporation at TRL 1 to 3 is rarely sensible, given that the risk of locking into specific technical architectures too early is high. In Advanced Materials, where the core innovation is often a process, the distinction between research and early manufacturing development can blur quickly, and incorporation may make sense as soon as pilot production becomes a realistic near-term objective.
Incorporation decisions, therefore, should be less about ‘how excited we are about the technology’ but ‘what kind of work are we doing now, and is the structure of a company better suited to that work than the structure of a research group?’ When the answer to the second question is ‘yes’, incorporation may be warranted. When it is ‘no’, remaining a research group may be more valuable.
4. Who Should Lead?
Put another way, who should become the CEO?
There is no universal rule here. But there are patterns and questions worth considering.
The academic PI is a common founding CEO profile, and in the right domains, often the right choice. Where scientific credibility is the primary currency of progress – where funders, strategic partners, and potential hires are making decisions based on the intellectual reputation of the leadership – the PI can carry the institutional authority, the peer network, the grant track record, and the scientific reputation that the early company needs to exist at all. This is especially true across Life Science and in domains like Quantum and Nuclear Fusion, amongst others.
Of course, there is a risk that the skills which make a great PI – deep domain mastery, intellectual independence, long-horizon thinking, comfort with open-ended exploration – are not the same skills that a company in Phase II requires (for a clear understanding of the different Lab-to-Market phases, take a look at our essay, ‘The Four Risk Phases’). PIs who become CEOs and remain effective are those who either evolve quickly into the translator and organisational designer role that Phase II demands, or who have the self-awareness to bring that capability in around them before it becomes the bottleneck. The PIs who struggle are those who cannot make that transition and will not, or cannot, acknowledge it.
Post-doctoral researchers are also popular founding CEOs, as are PhDs. Both have domain depth without the institutional tenure that can make some PIs resistant to certain kinds of change. Many have genuine operational experience – including running experiments, managing equipment, training students, dealing with the day-to-day chaos of a real laboratory – that many PIs have long since delegated. And they tend to have a more flexible professional identity, which makes the psychological transition to a commercial role somewhat less challenging. Of course, the same risks highlighted above apply within this talent pool.
What about bringing in a professional CEO at incorporation?
An executive with commercial and operational experience who has not been part of the original research group is sometimes advocated and occasionally effective. But from our experience, it is rarely the right answer during the Research phase. A professional CEO brought in too early often struggles to earn the trust of the scientific team, makes decisions that reflect commercial logic before the technology can support it, and adds a layer of translation overhead that actually slows, rather than accelerates progress. Importantly, commercial skills are not yet the bottleneck during the Research phase. Instead, in our view, the better time for a professional CEO transition – if one is needed – is Phase II or Phase III.
As with so much in the Lab-to-Market framework, the transition question is best answered in terms of phase of risk, not time. From what we have observed, a founding CEO should remain in the role for as long as their particular profile matches the requirements of the dominant risk. The question should never be ‘has the founding CEO failed?’ but instead, ‘what kind of leadership does the current phase demand?’
5. Integrity and the Internal Locus of Control
Deep Tech has produced a generation of researcher-founders who have watched peers become celebrated figures, raise large rounds, and command significant media attention at early stages of their journey. The pressure to tell the story compellingly, to project confidence, to run slightly ahead of where the science actually sits, is persistent. Resisting this, remaining genuinely anchored to what the evidence shows while still communicating with conviction, is what adherence to the Scientific Method demands.
This is what we refer to by internal locus of control. It is the discipline of keeping evidence-based judgement insulated from forces that might corrupt it. And it is this quality that is essential in founder CEOs and Lab-to-Market cultures, especially during the Research stage.
We have identified five founder CEO archetypes. Scientific Stewards are one of them, and the most common to emerge from research labs. With backgrounds as PIs, post-docs, or PhDs, these are often the perfect founder CEOs for Deep Tech spinouts. But even these types can shift in nature to believe, partially or wholly, that their narrative of the technology is truer than what the data shows. Such founders end up squaring the cognitive dissonance between evidence and aspiration in favour of aspiration – and then proceed to build an organisation around that resolution. There are many examples of those who have failed this ‘test’, including Elizabeth Holmes of Theranos and Jan Hendrik Schön of Bell Labs.
At the same time, it is important to appreciate that Lab-to-Market journeys, by their nature, select strongly for ridiculously ambitious people. Building something truly novel that survives the scepticism of peers, the patience of investors, and the conservatism of industrial customers, requires a level of personal conviction that most people cannot sustain under pressure over a period of years. Ambition is vital.
But the best leaders direct their ambition at the problem, not at the narrative surrounding the problem. They remain open to being wrong, taking care not to personalise it or to see it as a failure, because being wrong reveals something new. And no amount of commercial brilliance or domain mastery compensates for a leader who has stopped being honest with themselves and/or their stakeholders about what the evidence shows.
Building any Lab-to-Market company is one of the most difficult things anyone can do. But starting in the right way, including the delivery of a proven technology concept, a credible IP pathway / position, a strong early team, and a founding leader and culture that treats evidence with greater authority than belief – all of these help to ensure a more successful start.
Deep Tech Leaders is building the ‘data-first’ operating manual and talent network for companies navigating the Lab-to-Market journey. Through domain-specific data & insight, long-form analysis, in-depth conversations, and our world-leading talent network, we surface how Deep Tech companies actually progress – and why so many fail. This work is underpinned by our executive search practice focused on placing proven leadership talent into roles where phase, risk, and capability align.
