Most new tech companies scale using the lean startup model: create a minimum viable product and then incrementally improve it. This tends to work best for consumer or software businesses, but it’s much more difficult when it comes to deeptech — the technology takes longer to get right and may have no immediate market to slot into.
This means it can take some time (and funding) before revenue is unlocked. A semiconductor business that needs to build a prototype, or a new materials business that needs to show the longevity of its materials, could easily go more than five years from inception without generating any revenue — and it could feasibly take a further five to reach profitability.
So how can deeptech startups give themselves the best chance of scaling successfully in the near to medium term?
In part one of this three-part series on deeptech we looked at the initial investment phase and the questions VCs need answers to before investing. Here in part two, we’re looking at the growth phase and how to scale.
How to find the first customer
To scale effectively you need to be completely clear on what problem you are solving and for whom. Customer understanding is the most important challenge and opportunity.
It’s a chicken and egg problem: technical milestones need to be reached before a customer will buy your product, but you may need that ‘real world’ customer to try, test and build your product with you along the way. This means you need at least one customer to take the leap and do the technical validation, allowing you to reach the wider market with some real results in your pocket.
You will need to have a minimum of 100 conversations to find your champion customer.
Because you will be selling a promise rather than reality, you need a champion customer who shares your vision, and you will have to kiss a lot of frogs to find them. More than 100 conversations is the baseline.
You should also think about what carrots you can dangle in front of these champions. You may need to waive some costs, though you will always need to charge something, otherwise they aren’t really a customer. Alternatively, you may need to tailor your technical and product roadmap to make it work for them — but not so much that it becomes bespoke, as we will come on to later.
We’ve also seen consistently that a demo, with a real, tangible prototype showing the customer something that they can’t do themselves now, is a compelling reason for them to have a conversation with you — particularly when it solves a specific problem or pain point.
There are three broad approaches to think about your first customers:
1. Develop
The first option is to keep developing your technology until it reaches the required technical milestones for your biggest or most important prospective customers. This will mean being pre-revenue for longer, but it doesn’t rule out talking to future customers, forming relationships, getting feedback and staying current with market developments.
Remember that perfection can be the enemy of progress and a highly technical team will often solve for a 100% perfect product, delaying its release and overriding the customers’ immediate needs.
Founders will benefit from a mindset where imperfection is acceptable and prioritise getting something into the customers’ hands sooner rather than later, especially given the risk of running out of cash.
Take Dogtooth’s strawberry picking robots. They needed to be physically on farms, being used by growers to understand, for example: the best height of the chassis to pick strawberries, how far apart the two arms should be to get maximum fruit coverage, or whether the wheels work in muddy conditions — and so on.
2. Intersect the market
Is there a set of customers whose needs match your tech where it's at now? If you can sell to this segment without distracting from your more long-term product development goals, you can start getting revenues and market traction sooner.
We’ve seen this approach in quantum hardware startups. Whilst working towards building the most powerful computer on the planet, they can also offer simulation services on their hardware whilst it's in development and charge service-style revenues for this.
3. Joint development
A joint development agreement (JDA) is another way to start working with customers before your product is fully ready.
A JDA allows you to formally develop a product or technology with a potential customer, though you need to make sure any jointly developed IP doesn’t encumber you and is properly dealt with in the contract. Not only can you get revenues from these products, but you will have built a customer relationship.
You’ll be learning all the way, tapping into your customer's evolving needs, while significantly increasing the chance of being included in their roadmap.
One of our portfolio companies is currently jointly developing a ‘localisation product’. Under the agreement, each party bears its own cost of development, but the bigger corporate acts as the systems integrator, selling the product while the portfolio company provides the software for a very technical piece of the puzzle — in this case, estimating where a vehicle is located using camera inputs.
Build for scale not bespoke solutions
Given how hard it is to find the first customers for deeptech startups, it can be tempting to agree to build bespoke solutions. But be careful not to get distracted.
Does the bespoke iteration you are building for one short-term prospect actually take you towards the bigger vision and product? If not, don’t do it. You don't want to become a consulting service, solving one niche problem at a time, just to generate short-term revenues.
You don't want to become a consulting service, solving one niche problem at a time.
That said, interacting with and building for customers as early as possible has huge upsides and can be foundational for future success. Discipline is the answer. One company we know worked to a £10m cut-off point — if the bespoke development asked for was worth less than £10m to the business, they would say no.
At each iteration, continue to ask yourself: 'Will this be beneficial to my product as a whole?' and 'Are there more customers who would benefit from me doing this?'
Ultimately, the aim is to build a product that’s repeatable, that ‘sells in your sleep’ so to speak. Your product might be niche, but it should not be bespoke.
For some deeptech companies, revenues from a single customer may be in the tens of millions and scaling may be based on only four or five large customers. In these cases, the temptation to go down the bespoke route can be strong. It may be required for your biggest clients. In fact, any high touch customer account worth millions would expect a level of support and servicing, whatever the sector. But in deeptech the stakes and costs are higher and may involve the build and upkeep of specific, expensive technology.
Ultimately, you must remain loyal to your vision and avoid bending your company out of shape. Have some faith that your customers love what you’re doing and are along for the ride, but in the passenger seat, not behind the wheel.
A humble approach can pay off
A brutal reality of deeptech is that superior tech won’t always win out. Big corporations can have long and complex procurement cycles and may not have the flexibility to incorporate you.
Many potential customers may also be tech companies themselves, with their own research and development teams. This can sometimes result in the classic ‘not built here’ problem, whereby organisations have a negative attitude towards knowledge or technology developed externally.
Don’t insist on the fact that your product is superior, even if it’s true.
The antidote to this is an awareness of your customers’ context and a humble approach. Being sensitive to the fact that threatening a status quo, possibly even throwing light on their shortcomings, will make your approach more tactful and productive.
Don’t insist on the fact that your product is superior, even if it’s true. Rather frame it in terms of 'We could be useful here', or 'Perhaps our technology would enhance yours'. There’s always more than one way to approach a technical problem anyway, so this won’t be false humility.
When it comes to procurement cycles, you want to be the ‘need to have’, not the ‘nice to have'. An understanding of your customer’s strategic priorities is crucial here. If you can show that your technology converges with their roadmap, providing a potential shortcut, your case will be infinitely more compelling. Although timing — another word for luck — will inevitably be part of this.
How to get ready for your Series B funding round
All companies want to raise their next round at a higher valuation and lower risk than the one before it, but deeptech companies are unlikely to be doing the things ‘normal’ companies do at Series B.
A good growth rate with $5-10m annual recurring revenue and clear unit economics probably won’t feature in your deck.
A good growth rate with $5-10m annual recurring revenue and clear unit economics probably won’t feature in your deck.
This fact isn’t always factored in by the wider market at Series B, but fortunately, there are VCs who recognise this reality and know that it’s not unusual for a deeptech business to do more than one Series A round at the same valuation. Syndication of funding rounds is advantageous to give you a broader base of investors to support and share multiple rounds.
Key things deeptech investors look for at a Series B round:
- A large market opportunity. This should have been outlined in earlier rounds but you’ll want to demonstrate increased conviction.
- A path to large scale adoption with tangible, believable evidence. For example, do you have a waitlist for your software development kit; is your open source project contributed to by hundreds of maintainers; is your product being verified by a team in a large corporate that have earmarked this technology for multiple projects and/or for use by multiple teams?
- Reduced technology risk, supported by external validation of your technology by prospective or actual customers.
- Core IP identified and protected. This can often be patents, but the company may have a strategy for keeping their trade secrets and knowhow undisclosed. Either way, an understanding of the core IP, a strategy around it and how it is going to be protected — whether formally or informally — is required.
- A unique team. As investors, we need to believe that yours is only one of a handful, if not the only team on the globe who could genuinely tackle this technological problem.
- Referenceable customers in a beachhead customer segment where a repeatable problem set has been identified.
- Grants or public funding. These are a further sign of endorsement.
This is who you need to hire
Scaling won’t happen without the right people in place. Building out your team and having the best people at the right time are enormous challenges given the specificity of the skillsets needed. From day one you need to grow your network and sell your vision so that talented people will want to leave comfortable positions to join you.
An experienced chair can be transformational as an early hire.
As you scale, you’ll want to find people who not only understand the tech but can translate its value and excitement to a far wider set of stakeholders, some of them non-tech minded, such as potential customers, investors and other employees. These people should be able to explain in a few sentences what your product does and the problem it solves.
In our experience, an experienced chair can be transformational as an early hire.
Beyond that, critical first hires are a commercial lead who can go and secure those early sales, likely in an outbound, direct and nonscalable fashion — whilst the technology is probably still in development. Then a product lead who can ensure sales and tech teams work in harmony to please their initial customers.
In summary, scaling any business is challenging, but deeptech presents unique barriers. With some forward planning and an experienced VC backer, they can be overcome. And then you will be onto the next set of problems: when the company is scaling fast and you need to start thinking about exits. We’ll tackle that one in part three.
