Will Your Architecture Hold?
This tells you whether the way your system is put together will carry you forward, or whether it’s quietly setting up a rebuild you can’t afford. Answer a handful of questions about how your software is structured, where it’s coupled, and what happens when one part fails, and you’ll get a read on where the structural risk actually sits. It’s the first thing I look at on any audit, because architecture is the layer where mistakes get expensive slowly and then all at once.
How to read your result
The result isn’t a grade on your engineering team. It’s a map of where the structure will bend before it breaks. A high-risk flag almost never means “your code is bad” — it means the shape of the system makes a particular kind of change dangerous, and that danger compounds as you grow.
The pattern I see most: a system that works fine today but is built so that everything touches everything. One database every service reaches into directly. A single deployment where a change to billing can break login. No clear seam between the parts. It runs, the demo is clean, and then you try to change one thing and three unrelated things fall over. That’s the risk a high score is pointing at — not that it’s broken now, but that it’s expensive to change, and it’s only getting more so.
A low score means the opposite, and it’s worth saying plainly: it means leave it alone. Clear seams, parts that fail independently, changes that stay contained — that’s a system you can grow into. Don’t let anyone talk you into re-architecting something that’s holding. The most common over-engineering mistake I see is a team rebuilding a perfectly sound system because a louder architecture looked more impressive on a diagram. Solve the real problem; if there isn’t one here, don’t manufacture one.
The middle band is where judgement matters most. It usually means one or two specific weak points — a part that’s become a bottleneck, a dependency you can’t easily replace — sitting inside an otherwise sound system. That’s good news. It means a targeted fix, not a rebuild.
What I’d do about it
If you scored high, resist the instinct to rebuild from scratch — that’s almost always the wrong move, and the most expensive one. The play is to find the worst seam and cut it cleanly: pull the most entangled part out, give it a clear boundary, let it fail on its own. You do that incrementally, on a system that keeps running, the same way you’d renovate a house you still live in. A full rewrite throws away years of hard-won knowledge baked into the existing code and bets the business on a build that’s late more often than not.
Underneath every high score is usually the same root cause, and it’s worth naming: the parts of the system you own least are where the risk concentrates. A platform you’ve over-relied on, a vendor whose API dictates your data model, a rented service wired so deep into your core that you can’t move without its permission — that’s where the structure gets brittle and the options run out. The fix points the same way every architecture decision should: own the parts that are actually your business, keep the seams clean so you can swap the rest, and don’t let a tool you rent become load-bearing in a way you can’t undo. The full reasoning on that line lives in build or buy.
If you scored in the middle, fix the one or two weak points and move on. Don’t gold-plate. If you scored low, you’re done here — go build something.
Talk it through
If your result was worse than you expected, or you want a real set of eyes on the structure rather than a handful of questions, an architecture review is exactly that: an honest read on whether what you’ve built will hold, and what to do if it won’t. You can also tell me about it directly — twenty minutes on the specific weak point usually makes the next move obvious.
Two related reads if you want them: the tech debt calculator puts a cost on the friction a brittle structure creates, and is your project at risk? checks whether a build in flight is heading the same way — both sit alongside this one in my diagnostics. And if what this surfaces is a system big enough to need a standing team to re-shape and run it, that’s what my agency, Digital Royalty, is for.