Startups have to care about productivity again

When it’s raining money and a dollar today is worth about the same as a dollar seven years from now, growth really is the most important thing for a startup. Scaling the team and spending large on marketing is the name of the game, and so long as you’re doing those two things then burn rate doesn’t matter too much—the next raise is always around the corner.

Under this paradigm it doesn’t really matter what tech stack you pick. While your chosen stack will influence your time to market, customers don’t really care that you’re using Rust on the backend. Customers care that you’re solving their one very specific pain point, and solving it well. You can pick whatever complicated stack you want and so long as your MVP achieves product-market fit, you’re off to the races.

Who cares if your chosen stack requires twice as many engineers to ship a feature? Just hire more heads and keep the wheel turning! Profitability is a nice-to-have right before your distant IPO.

Over the past year the game has changed. Growth at all costs was a zero interest rate phenomenon. Today the trick is to deliver real value to customers with meaningful financial performance. Caring about unit level economics is cool again.

I talk to a lot of people in the startup space, and am constantly amazed by the complexity they’ve taken on. Complicated microfrontend architectures, sprawling distributed monoliths masquerading as microservices, and stunted reinventions of commoditized PaaS offerings are commonplace in the tech startup world. These approaches worked when you could rely on doubling your engineering team every year, but now they’re real barriers to improving your product.

It’s 2023, and it’s not hip to write Terraform files anymore. The startups that will win over the next decade are streamlining their software development lifecycle by following these guidelines:

1. Don’t develop a native app until market conditions dictate it. All things considered, the web is a pretty mediocre app platform. On the other hand, multiplying your engineering effort to two native codebases is a disaster and app store discoverability has been broken for a long time. Building on the web lets you move quicker, and experiment faster.
2. Don’t self host. Running servers competently is way harder than you think it is. Delegating this responsibility to AWS or Vercel lets you narrow your hiring requirements and run leaner.
3. Hire full-stack developers. You need engineers who can develop and ship features end to end. Silos between frontend and backend are anathema to high performing teams. Simplifying your infrastructure, cloud, and operations requirements makes it easier to hire great product engineers with deep frontend and backend experience.
4. Hire designers that understand the web. If your product designers are stuck in Figma, they’re going to drag down productivity. Designers need to understand the constraints of the medium, or else there’s going to be tension between engineering and design. If your designers can write HTML and CSS, there’ll be a lot less back and forth over ‘impossible’ Figma designs.

Now’s a good time for founders and technology leaders to do a deep dive on product team productivity. Where are the pain points, and where are things getting bottlenecked? What would it mean for your customers if you were able to ship more things, and ship those things faster and with fewer people?

Did you implement your processes and pick your tech stack based on practical considerations, or because they were hyped up in a FAANG blog post?

Things are pretty messy right now in the startup ecosystem. The forecast says things will get even worse before they get better. Smart founders are already neck deep in this problem.

It’s a lot easier to do this work now than next year.