Announcing our $32 million Series B
gradient
Startup lessons from 2 years building Airplane.dev

Startups

6 min to read

Startup lessons from 2 years building Airplane.dev

Written by

Ravi Parikh, Founder at Airplane.dev

Published on

Dec 20, 2022

As of December 2022, we've officially been working on Airplane for 2 years. In that time, we've grown from 2 co-founders to a team of 20, brought on hundreds of paying customers, and massively expanded the scope of our product.

We're still very early, but there are already a few lessons learned that might be helpful to other early-stage startups.

1) Launch early…

We founded Airplane in December 2020 and publicly launched the product later in July 2021. We got more customer feedback in the few days after the launch than we'd gotten in the preceding 7 months. There were so many insights and learnings about our product-market fit that we could have acquired a lot sooner if we'd launched sooner.

The typical reason people wait to launch is that the product isn't "ready." But for a pure software SaaS product, you can typically scope down the MVP a lot more than you think you can, and if your core feature/differentiation is solving a real problem, people will use even a very bare-bones product. Last year, I wrote a blog post where I noted that recent SaaS launches were a lot more polished than they used to be back in 2013, when I founded my prior startup. I think this is a mistake. In retrospect, we could probably have launched Airplane to the public in Jan or Feb 2021 if we'd really been ruthless about prioritization, and I believe we'd be 5-6 months further along today if we had.

2) …and launch often

Since our initial launch, we've done more product launches around features that represented a step-function change in functionality.

Startups often neglect product launches as a potential marketing channel since they seem like one-time events that don't "scale," unlike more steady-state marketing channels like ads or content. But they're incredibly impactful if done right. Each of our product launches has driven a huge spike in user signups and attention, and even after the one-time spike died down, we saw a permanent increase in signups that was higher than the daily average prior to the launch.

Also, make sure to re-explain your core value proposition with any feature launch. You aren't Apple releasing the next iPhone–you should assume that most of the people that see your new feature launch are also seeing the rest of your product for the first time too.

Our strategy for these feature launches has, paradoxically, been the opposite of my above advice to launch as early as possible. For our most recent Views launch, we spent several months iterating on the feature before launching a fairly polished product to the world. However, we got to "have our cake and eat it too." We released an MVP as fast as possible to our existing customers in a private beta. We then saved the public unveiling for a few months later when we had confidence in its product-market fit.

This strategy, of staggering a marketing launch after a private "product launch" of your feature to existing customers, has worked really well for us. That being said, I only recommend this strategy for at-scale new feature launches. For the initial launch, I'd get things out there as soon as possible as mentioned above. The mistake I see with so many startups building "in stealth" is that they're trying to do this as well, but this doesn't work when you don't have an existing customer base already.

3) Treat the money you have as if it's the last you'll get

When I started my previous company, Heap, in 2013, we were fortunate to go through Y Combinator and raise a $2M seed round. After that, we assumed that this money was the last money we'd ever get and that we had to get to profitability on that fundraise alone. We knew that it was theoretically possible to one day raise a Series A, but we never felt that it was a milestone we could predictably hit. We did end up getting cashflow positive from there, and then eventually did raise a Series A, but not out of necessity.

That mentality wasn't unique to Heap. Most founders I knew in 2013 assumed that VC funding was unpredictable and never relied on a "guaranteed" next round if they hit certain metrics.

Over the last few years, tech companies seem to have un-learned this mentality. Many companies have 18-24 months of runway and no "Plan B" if a VC doesn't bail them out when they have 6 months of cash left. With the current tech downturn, people are learning this lesson quickly, but the lesson was still just as true last year.

Luckily, we've run Airplane conservatively. Even in 2021, when VC money was abundant, we hired slowly relative to our cash position. We recently announced our Series B fundraise. When we raised that round, we still had over 80% of our capital remaining from our Series A, so raised opportunistically on good terms, rather than out of need.

4) Be thoughtful about fundamental decisions, be fast about everything else

Airplane is a code-based, developer platform for building internal tools. Every app or workflow you create in Airplane is fully expressible as normal JavaScript/Python/React/etc. code that you own, can store in your codebase, and extend however you'd like.

This decision about Airplane is fundamental to Airplane's value and differentiation. We understood from Day 1 that it was important to get this decision right, both in terms of the large feature bets, but also more detailed product decisions. I'm glad we invested the time to get a lot of this (mostly) right early on, saving a lot of thrash later.

However, we also spent a ton of time overthinking other decisions or overbuilding features that ultimately did not matter as much. For example, we've redone our product onboarding flow from the ground up 3 times and we'll probably change it again soon. Early on, trying to gameplan exactly how users would go through the journey of learning Airplane was a waste of time, and we would have been better off doing whatever was simplest to implement.

5) Your business model requires more thought than just saying "PLG"

There are several startup tropes or strategies that many of the "hot" startups seem to employ, and other startups emulate these strategies even if they don't make sense for that particular product or market.

One example we encountered: seemingly every new SaaS company wants to be a PLG (product-led growth) company. Founders want to create a beautiful product that gets bottoms-up adoption within small and large companies alike and then spreads virally within the company. When we started Airplane, we had the hope that Airplane would work this way too.

But not every product lends itself to exactly that process. It's important to be honest about what a realistic buying process will look like for your product, and align your go-to-market process to that. At Airplane, we do consider ourselves to have characteristics of PLG, but we've learned over time that we can't do as "pure" of a PLG process as companies like Zapier or Figma.

Instead, we have a process where anyone can signup for Airplane and use it on our free/self-serve plans, but there almost always needs to be at least some lightweight sales process for companies above a certain size. There are a few reasons for this:

  • Airplane requires write access to production data to be useful. It inevitably requires some level of security approval, which can involve things like filling out security questionnaires or talking to security teams.
  • Airplane is a platform that lets you build internal tools. This is inherently vague and the potential use cases are really wide. Most people come to us with 1-2 concrete use cases in mind, but a strong sales-led discovery and demo process can get a customer to think of several more.
  • Airplane doesn't have much utility as a single-user product. It requires an entire team to buy in. A good sales process can help with team-level buy-in.

Some pure PLG companies can avoid the sales process entirely, or put it a lot later in the customer journey (eg only reaching out after there are 20 active users at a company), but the inherent characteristics of our product and market require more of a sales process early on for larger customers. So we built out a sales team earlier than a pure PLG company might need to.

Overall, Airplane is early on its journey, and I'm sure we'll have more (and different) lessons as it evolves. However, triangulating across my previous startup and Airplane, it seems like these lessons continue to apply to early stage startups across all kinds of market environments.

Share this article:

Subscribe to new blog posts from Airplane.