Somewhere around 55 percent of deep tech companies do not reach their fifth birthday, and most of the reasons recorded afterwards are not scientific. They are failures to be understood, by buyers, by investors, or by the person inside a large organisation who has to explain the technology onward once the founder has left the room.
That figure and the reading of it come from Hailey Eustace, who works on communication for deep tech companies. Her argument is that if you take the top ten stated reasons for failure, eight of them come back to communication, either failing to listen to what the market needed or failing to make the value understandable to people who are not specialists.
Why communication failures get recorded as something else
Nobody writes lack of clarity on the post-mortem. What gets written is lack of product-market fit, funding constraints, or the business model did not work.
Look at what sits underneath those. Lack of product-market fit often means the company built for a problem it had not verified with enough people, which is a listening failure. Funding constraints frequently mean investors did not understand why this mattered quickly enough to fund the next stage, which is an explaining failure. A weak business model often means nobody established what the technology was worth to a buyer, which requires having asked.
None of those are failures of engineering, and all of them are recorded as something more respectable than we could not make ourselves understood.
I want to be careful here, because this argument can be pushed too far and usually is by people selling marketing. Some deep tech companies fail because the physics does not scale, or the cost curve never crosses the incumbent’s, or a competitor got there with a better approach. Those are real and no amount of communication would have changed them. The claim is not that everything is a communication problem. It is that a large share of what gets labelled as something else turns out, on inspection, to be one.
The specific failure mode
The one I see most often is a company that can explain itself perfectly to another specialist and not at all to anyone else.
This is invisible from inside, because the people the founders speak to most are people who understand the field. The problem only shows up in the moments nobody witnesses: the investor who reads a deck and does not follow up, the procurement lead who visits the website and does not reply to the email, the corporate engineer who was genuinely interested and could not carry the argument into their own management meeting.
Hailey Eustace’s version of this test is the sharpest one I have heard. If the people in the room cannot explain your value proposition after you have left, the sale is already dead, whatever they said while you were there. And you never find out, because nobody writes to tell you they could not repeat your pitch.
The habit that produces it is well-intentioned. Founders send the full technical case, because in academia the complete argument is the respectful thing to send and the reader is assumed to want depth. Sending a 50 page body of work to an investor is generous behaviour in the culture the founder came from, and unreadable in the one they have entered.
The three sentence test
The practical diagnostic is to write down, in three sentences a non-specialist can repeat, what your company does, who buys it and why they would choose you.
Most technical founders find this uncomfortable, and the interesting part is why. It is not usually a wording problem. When the sentences will not come, it is generally because one of the three answers is not settled. The buyer is still a category rather than a named type of company. The reason to choose you rests on a technical advantage nobody has priced. The application is one of four possibilities the team has not chosen between.
That is why the exercise is worth doing even by companies with no interest in marketing. Distilling the message is a stress test of the commercial strategy, and it fails in exactly the places the strategy is thin.
Where this becomes visible
The website, mostly, because it is the one place every stakeholder goes and the one place nobody is present to explain.
A site that opens with the mechanism rather than the consequence is telling a buyer that this company thinks of itself as a research group. A site with no evidence that any serious organisation has engaged is asking a procurement team to be the first, which they are trained not to be. A site written for an investor audience and then used for two years as a sales asset is answering questions nobody in the sales conversation asked, which is the subject of the pitch that raised your round will not close an enterprise contract.
None of that means the website is the cause. It is where the underlying message problem becomes legible, and it is the cheapest place to fix it once the message is right. Fixing the site without fixing the message produces a better looking version of the same failure.
The hiring version of the same mistake
There is a related pattern worth naming, because it wastes a year when it happens.
A company recognises it has a commercial problem and responds by hiring one junior marketer, handing them the technical documentation and asking for corporate pilots. Eustace’s version of this is the forty thousand pound hire expected to build a global market presence single-handedly.
It does not work, and the failure is not the person. Translating advanced science into a commercial argument is a strategic job connected to the business model, the fundraising plan and the product roadmap, and it needs the chief executive involved. Nobody at that level of seniority can do it alone from the outside, and when it fails the company usually concludes that marketing does not work for them.
The comparison I would make is that you would not hire a graduate to run the R&D programme and conclude, when it went badly, that research is ineffective.
What to do about it
Start with the three sentences, and test them on somebody genuinely outside your field, then ask them to repeat it back a day later. What comes back is your actual message, regardless of what you wrote.
If the sentences will not settle, the problem is upstream and the work is customer discovery rather than copywriting. If they settle easily and your materials do not reflect them, that is a production job and a fairly quick one.
And if you are about to hire one person to solve this, consider what you would need to be true for that to work, and whether it is.
Our website and branding work usually begins with the three sentences rather than with design, because the design decisions follow from them. The Albotherm project is the clearest example of that order, and a 20 minute call will get you an honest read on where your own gap sits, including if the answer is that your message is fine and the problem is elsewhere.
Frequently asked questions
What is the main reason deep tech startups fail?
Rarely the science. The reasons usually recorded are lack of product-market fit, running out of funding and a weak business model, and underneath most of those sits a failure to understand what the market needed or to make the value understandable to people who are not specialists.
How long does a deep tech company usually last?
The commonly cited figure is that around 55 percent do not reach their fifth birthday. Deep tech has longer development timelines than software, so a company can spend those five years making genuine technical progress and still run out of money before it has commercial revenue.
Is a website really a factor in whether a deep tech company survives?
Not on its own, but it is where the failure becomes visible. Buyers and investors check a company before replying, and a site that reads like a research abstract tells them nothing about whether this can be bought. The underlying problem is usually the message rather than the design.
What is the three sentence test?
Explaining what your company does, who buys it and why in three sentences a non-specialist can repeat accurately. If that cannot be done, the gap is usually in the commercial strategy rather than the wording, which is what makes it a useful diagnostic.
Are there failures communication cannot explain?
Yes. Some companies fail because the technology does not scale, because the economics never beat the incumbent, or because a better approach arrived first. Those are genuine and no communication work would have saved them. The point is that a large share of failures recorded under other headings turn out to be communication problems on closer inspection.