Talk to us

Blog · 16 November 2026 · Denver White

What Five Days With Your Code Actually Buys You

What we actually look at

Security and access control. Whether one user can reach another's data. Where credentials are stored. What happens at the boundaries where the outside world sends you things. This is where we find the most, and it is where the consequences are largest.

The database. How the data is structured, whether the relationships make sense, whether the constraints that should prevent impossible records exist, and whether the indexes match what the application actually asks for.

Correctness under pressure. What happens when two people act simultaneously. Whether multi-step operations are grouped so they cannot half-complete. The bugs that only appear once you are busy.

Scale. Which parts get slow as data grows, and roughly when. This is usually predictable to within an order of magnitude, which is enough to plan around.

Maintainability. Whether a competent developer could pick this up. Whether anything is documented. Whether you are locked to a platform you cannot leave.

Operations. Backups, and whether anyone has ever tested restoring one. Logging. Whether you would find out about a failure before your customers did.

What we routinely find

Being honest about the pattern, because it is remarkably consistent.

Nearly always: missing ownership checks somewhere, and missing indexes. These two turn up in the large majority of what we look at.

Frequently: operations that can half-complete, no useful logging, backups nobody has tested, and error handling that stops at the happy path.

Sometimes: credentials committed into the codebase, a data structure that will genuinely need reworking, and dependencies that stopped receiving updates some time ago.

Occasionally: nothing very much. It happens, and it is a good day when it does.

What you get at the end

A written report, in English, that your non-technical co-founder could read.

Findings sorted into three groups: fix now, fix soon, and worth knowing. Each with what it is, what could actually go wrong, and roughly what it would take to put right. No severity matrices, no scoring system, no forty pages of automated tool output.

You also get a straight answer to the question everyone actually wants answered: is this fixable, or should I start again? In our experience the answer is nearly always fixable, and we will say so plainly when it is not.

The report is yours. Take it to another developer if you would rather they did the work. That is a perfectly reasonable thing to do and we will not be offended.

What it is not

It is not a penetration test. We read the code and the structure rather than attacking a running system.

It is not a rewrite, and it is not a sales funnel dressed up as a review. Plenty of audits end with a short list and no further work, which is a good result.

It is also not a certificate. Five days is enough to find the things that matter and give you an accurate picture. It is not enough to prove the absence of every possible problem, and anyone claiming otherwise is selling you something.

Who it is for

Founders with an AI-built application that is starting to carry real weight. Anyone who has inherited software and needs to know what they have taken on. People heading into a funding round or a sale who would rather find things themselves than have someone else find them. And anyone with the nagging feeling that they cannot answer questions about their own product.

What you are really buying is the second half of the job, the half that a brilliant typist and a terrible architect was never going to do for you.

Questions before you commit to anything? Get in touch for a no-pressure conversation.

Tell us what you've got

It goes straight to Rob and Jase. You'll hear back from one of us, usually the same day, and we'll tell you honestly whether we're the right people.

Just want it checked? Code audit, £250 + VAT