What a codebase audit actually finds
By Wernich Bekker
A lot of my work starts the same way: someone has software that used to move fast and now doesn’t. A founder inherited a codebase from a contractor who’s gone. A team can’t ship a change without three things breaking. The question is always some version of “is this fixable, or do we start over?”
A codebase audit answers that. Here’s what it actually looks at — and what it tends to find.
What the audit looks for
- Can a new person be productive quickly? How long from cloning the repo to running it locally and shipping a small change. This one number tells you more about a codebase’s health than any metric.
- Where does change get expensive? Every codebase has hot spots — files everything depends on, logic copy-pasted in six places. These are where bugs and delays cluster.
- What’s the blast radius of a mistake? Tests, types, and deployment safety. Can you ship on a Friday? If not, why not?
- What’s the operational reality? What it costs to run, what breaks at 3am, and whether anyone would know if it did.
What it usually finds
Rarely a single catastrophe. Almost always an accumulation of small, reasonable decisions that made sense individually and added up to a codebase that’s hard to change:
- No local setup story, so every developer’s machine is subtly different.
- A “temporary” workaround from two years ago that everything now depends on.
- Tests that exist but don’t test the things that actually break.
- One person who’s the only one who understands a critical piece — a bus-factor of one.
Fixable, or start over?
Almost always fixable — and far more often than people expect. Rewrites are seductive because a clean slate feels faster, but they throw away years of hard-won edge-case handling and take longer than anyone plans. The honest recommendation is usually: keep the code, fix the three things making it painful to change, and get back to shipping.
The point of the audit isn’t to grade the previous team. It’s to give you a clear, prioritised picture of where you are and the shortest path to moving again. Sometimes the most valuable thing I deliver is permission to not rewrite.