Building in the Open: Our First 90 Days
What we've learned, what we've broken, and what's next for Dilmune.
Ninety days ago, Dilmune was an idea and a handful of prototype services. Today, we have a cloud platform, a cyber security product, and this blog you're reading right now. Here's what the journey looked like.
What Went Right
Choosing our stack early and committing to it. Astro for the blog, Sanity for content, Tailwind for styling. These decisions compound — every day we don't spend debating tools is a day we spend building product.
The other thing that went right: building the brand early. Too many startups treat design as a later concern. We invested in typography, color, and voice from day one, and it's paid off in ways we didn't expect.
What We'd Do Differently
We underestimated how much time design takes. Not the visual design — the information architecture. Figuring out how to organize content, what belongs on which page, how navigation should work. These decisions take time and iteration.
We also learned some hard lessons about deployment:
1# What we thought deployment looked like
2git push origin main
3# Done! ✓
4
5# What deployment actually looks like
6git push origin main
7# → Build fails (missing env var)
8# → Fix env var, push again
9# → Build succeeds, deploy fails (wrong region)
10# → Fix region, push again
11# → Deploy succeeds, but images are broken
12# → Fix image pipeline, push again
13# → Everything works!
14# → ...on desktop. Mobile is broken.
15# → Fix responsive layout, push AGAIN
16# Total time: 4 hours
17# Estimated time: 5 minutesBy the Numbers
- 3 products launched (Cloud, Cyber, Blog)
- 847 git commits
- 12 major pivots in product direction
- 1 production outage (it was DNS. It's always DNS.)
- ∞ cups of coffee
What's Next
More of the same, but better. We're doubling down on the things that work and cutting the things that don't. The next 90 days will bring public APIs, expanded documentation, and — if things go well — our first paying customers.
Stay tuned.