“If you’re not embarrassed by the first version of your first product, you’ve launched too late,” said LinkedIn founder Reid Hoffman in 2009. While Silicon Valley is full of stock phrases about how to successfully grow a company, Hoffman gets to the heart of one of the entrepreneurial dilemmas — the growth versus product conundrum.
I’ve witnessed firsthand how this has come to bear in the automotive industry since it began its dalliance with deeptech over 20 years ago. At Fernride, my day-to-day is spent leading the development of our human-assisted autonomous driving technology for the logistics sector.
On the surface, Hoffman’s equation is a simple one: how do you increase market share without jeopardising the quality of the service provided? Pulling it off is less straightforward — and when you’re dealing with robotics, fast-moving cars and trucks, the stakes are high.
Advertisement
What’s missing from the “release early, release often” mantra is the pivotal distinction between Day I, the engineering behind a product; and Day II, how it actually runs. Because when you’re dealing with enormous R&D overheads, troubleshooting your tech stack on the fly can be a death sentence.
There is a big difference between being able to iterate and winging it. I’ve seen both and I can tell you from experience neither your investors nor your customers will take kindly when too much is left to chance. Data may be the new oil but it’s still flammable when mishandled.
Here are my top 10 tech stack mistakes to avoid as you scale.
1. The “best” tech does not always win. The best operations on the other hand…
Remember Betamax? Maybe, maybe not, But everyone has heard of VHS, its technologically inferior competitor which ended up dominating the video market from the1980s to 1990s. The lesson from their contrasting fortunes is that success isn’t all about the tech, it’s about understanding your target domain and the constraints you’re working with.
This means keeping tabs on your supply-chain partners as well as the regulatory and safety standards. You also need to ensure your software is embedded into existing systems, customer profit margins and broader unit economics. Whether you’re working in a highly regulated, safety-first domain such as autonomous driving or a more consolidated B2B tech offering, all of your moving parts should be able to adapt to get you to market and help you stay there.
2. CVs aren’t everything
Deeptech pyrotechnics on a polished CV or LinkedIn profile aren’t the be-all-end-all, especially when the challenge ahead is as much commercial as it is technical. Experience is one thing but the ability to implement the right solutions in the right context is often another — and it’s not always immediately evident on paper.
If you ignore the realities of your supply chain, which are more unpredictable than ever before, they will not ignore you back
Look for senior personnel who can adapt to the exigencies of your product in a short space of time. That means balancing expertise in development processes, coding languages and verification methods against experience in bringing a product to market amid a changing regulatory environment.
3. Manufacturing the golden sample
Engineering a smooth passage from R&D into manufacturing is critical and comes in three stages. First, there’s your golden sample, which is all about shepherding your product through multiple iterations.
Stage two is troubleshooting and delivering a golden batch to understand the problems when scaling internally. Think of it as a market stopgap, which will give you more data to finetune problems and ambiguities.
Finally, there’s stage three: your production line. While the first two stages are all about agility, stability and predictability are key in this final stage. Consumers need to know what they’re getting and that they’ll get it on time. Weave the R&D and manufacturing stages into one continuous flow with the right timing, and everything else will fall into place.
4. Watch out for the long tail
The “long tail” refers to those rare but critical corner cases to which a system must be able to safely respond. Whether or not you are building autonomous technologies like robotaxis, you must ensure your tech stack has the demonstrable system capabilities to handle all eventualities. Tiny margins of error are easy to handle if you spot them in your prototype, but as you scale they can start piling up fast. Establish verification frameworks and close feedback loops with a well-trained operations team and external industry partners.
Advertisement
5. Beware of data silos
Establish data flows between engineering and operations so that your product can adapt in real time. Customers expect the product to be functional from the get-go, which means you need to determine the key diagnostics, timestamps, logging and tracing packages that are necessary to maintain a scaled product. You can then integrate technical factors — including connectivity, cloud computing and storage — into your business plan more effectively.
6. Your supply chain defines your launch
If you ignore the realities of your supply chain, which are more unpredictable than ever before, they will not ignore you back. Build a supply chain that supports how you want to launch, as well as your scaling targets, phase by phase. As you achieve more market penetration, you’ll have access to more suppliers and partners who want to work with you, so build this into your strategy.
7. Pay attention to the release cycle dilemma
Customers are paying for product features, not an overall engineering environment, so make sure you have the building blocks for your “vertical slice” locked down. This is an end-to-end impression of every layer of your tech stack that runs the full hardware-to-software gamut — and can be a problem when they move at different paces, which they invariably do. Your software might be updated multiple times daily, while your hardware lifecycle might be yearly, so ensure these two domains interface with one another.
8. Take “conventional wisdom” with a pinch of salt
Don’t be overly focused on what the textbooks and consultants say. You need an organisational design and workflow that suits your team and your product. Developing a website is radically different to building a safety-critical robotics system. Don’t treat them as if they are the same.
Don’t get hung up on being “agile”. Blindly applying the term to everything is completely ridiculous. Be “post-agile” — respond to market realities, the need for revenue generation and overall product functionality.
9. Trust your people
Scaling a company means continuously changing it. As you scale, the skills and experience needed to succeed in each role will evolve. When hiring, focus on people’s skills and value systems. Pay attention to talent density: hiring the best of the best and paying more for them, but hiring in smaller numbers and across multiple levels.
And remember, sometimes, the right person for a particular role is already working for you.
10. Passion is everything
There will be highs, lows and times you may even think you’ve hit a wall but it will be your sense of passion that keeps you going. So when there is something to celebrate, celebrate. These milestones and successes are everything. Sometimes they might even feel like failures but if they are instructive, share in them. Build a team that has that inner motivation to keep things turning on all cylinders.
Scaling a business is a marathon. When the chips are down, press on. The road is long but it’s worth it.
