#93: Opportunity Cost & Sourcegraph Bet
August 3, 2025 – DevTools Brew #93
I’m Morgan Perry, co-founder of Qovery. Every week, I share the raw, often uncomfortable lessons from building and scaling a tech startup from 0 to 1 and beyond.
In today's edition:
Opportunity Cost
→ My biggest learning after growths and plateaus? “Judgment”. Knowing what’s truly worth the effort and what only feels like it is.
The Sourcegraph Bet
→ Why Sourcegraph won where others didn’t. Three boring but brilliant execution bets that made all the difference.
Let’s dive in.
Opportunity Cost
“Every decision costs you something. The risk is thinking it doesn’t.”
As a founder, you go through phases.
The early excitement. The traction high.
Then the plateau… when things feel heavier, quieter, slower.
That’s when the question isn’t what should we do next?
It’s: what are we doing today that’s holding us back?
Because most stalls don’t come from lack of effort.
They come from effort spread too thin, or spent in the wrong places.
After early traction, 20+ hires, many tests, shifts, false starts…here’s one thing I’ve only recently started to get right:
Judging decisions.
Not just acting fast. Not just doing more.
But weighing: what’s truly worth it?
That’s how I define opportunity cost.
Not the textbook version.
The real one.
→ The silent cost of doing X… when Y would’ve created 10x more value.
It’s not just what does this give me?
But:
→ What does it replace?
→ What does it block?
→ What do we gain if we don’t do it?
→ What do we risk if we do?
→ And: is the potential gain greater than the potential risk?
That’s the core.
It’s not about doing less. Not doing more.
It’s about choosing what compounds and letting go of what doesn’t.
One concret example we’re living through now:
Big public-sector client.
Strong logo. Deep in the funnel.
But it requires custom work.
Old me: “Let’s ship it. We’ll figure out the rest later.”
Now? the lens has shifted:
→ Can we reuse it?
→ Will it compound?
→ Who maintains it in 6 months?
→ Is it a wedge — or a one-off?
And above all:
What does this choice displace?
Because when the upside is real and repeatable, then the effort is worth it.
But if it’s just a one-time win that drags us later, it’s not.
Same goes for everything else:
→ Hiring: Are we bringing in leverage or burning energy trying to make the wrong fit work?
→ GTM: Do we chase something new or deepen where we already win?
→ Product: Do we build more or strip back what’s not pulling its weight?
Every move has a cost.
Even the ones that look like progress.
And the tricky part?
That cost hides behind “busyness.”
You feel active but drift off course.
Six months later, nothing’s changed.
—
So how do you get better?
Not with formulas.
But with judgment.
→ Short-term relief vs. long-term gain
→ Energy spent vs. leverage created
→ One-time fix vs. compounding system
At Qovery, this changed how we:
- Hire → for fit, not speed
- Prioritize → on bets that scale
- Sell → where we can land and expand
- Say yes → only when the upside is real, repeatable, and worth the cost
Because yes: 95% of the game is execution.
But execution without judgment? That’s just noise.
So now I ask more often:
→ If we do this, what do we gain?
→ If we don’t… what do we keep?
It’s not about doing less or more. But about choosing better.
And that’s where leverage begins.
Sourcegraph Bet
Most devtool startups build great tech.
Very few build trust.
Even fewer survive long enough to be trusted.
Sourcegraph did.
Not because of a flashy launch. Not because of open source buzz.
But because they made 3 boring, smart, compounding decisions…EALRY:
1. They stopped asking for too much trust, too early.
Early Sourcegraph was cloud-only.
Upload your whole codebase to their servers.
Guess what? Nobody did.
They realized the mistake:
→ Even if your product is good, it won’t matter if trust is the bottleneck.
Fix:
They went full self-hosted, even before it was cool.
docker run → try it locally → zero friction, zero trust required.
That unlocked their first real customers.
And gave them something even more valuable: credibility.
2. They picked simplicity over “modern” stack decisions.
Most infra startups overengineer delivery.
Sourcegraph did the opposite.
They built it like any good open-source tool:
- Postgres, Redis, object storage.
- No exotic dependencies.
- One Docker image that “just works”.
Result:
→ Easy to deploy, easy to trust, easy to adopt.
→ Works in any environment, no excuses.
They didn’t try to impress.
They tried to ship.
3. They didn’t pretend open source would do the heavy lifting.
Open-sourcing Sourcegraph didn’t change everything.
It barely changed anything.
Why?
Because open source ≠ distribution.
It ≠ product-market fit.
And it ≠ community by default.
They still had to:
- Market the product like a business.
- Convert usage into revenue.
- Define clear product lines between free & paid.
They treated open source as a trust enhancer, not a growth channel.
That’s what made it durable.
—
Lessons for founders in devtools:
→ Don’t assume your users trust you. Make it easy not to.
→ Build like an open-source tool, even if you’re not one.
→ Open source is a wedge, not a strategy. Don’t rely on it to grow.
→ Say no to custom requests you can’t generalize.
→ The best infra is boring. That’s how it scales.
→ Don’t over-optimize your tech but optimize your path to usage.
You don’t see Sourcegraph everywhere.
But in the places it exists, it’s critical.
That’s not because they shouted louder.
It’s because they made the right kind of decisions, the ones that compound.
Final Thoughts
Opportunity cost isn’t just a founder concept. It’s a team skill. A company muscle.
And Sourcegraph is a perfect example of what it looks like when it’s applied early and well.
They didn’t try to do everything.
They made deliberate bets:
→ No cloud before trust → No overengineering → No illusions about open source doing the job alone.
That clarity gave them momentum.
And that momentum compounded because they protected their focus.
In a way, that’s the biggest lesson from both stories:
→ At Qovery, we’re learning to ask: what do we gain if we don’t do it?
→ At Sourcegraph, they kept asking: what do we need to say no to — to earn long-term trust?
Two different phases. Same mindset.
Don’t do more. Don’t do less. But do what compounds.
That’s how you protect your edge and scale without burning it.
That’s it for me today! :)
Thanks for reading and Happy Sunday!
— Morgan
///
Do you like personal lessons like this? More insights/stories from other devtool founders? Let me know, I’m always open to feedback.
You can reach out to me on LinkedIn.