Two days · Fixed fee · A report you can act on
Architecture Review
A written report of how your system actually behaves and why it was built that way — so every plan on the table gets argued against evidence rather than against what everyone remembers.
$500, fixed, agreed in writing before anything starts.
Two ways people arrive here
You inherited it
The people who built it have gone, the documentation stopped being true a while ago, and a rewrite has been proposed with a number attached that nobody in the room can defend or refute.
- Something in the estate is slow, expensive or fragile, and the diagnosis changes depending on who is asked.
- Budget is about to be committed to a modernisation programme on the strength of how the code feels to work in.
You are about to grow it
Funding has landed, or a new market has a date on it, and the plan quietly assumes the platform will cope. Nobody has measured whether it will, or which part gives first when it does not.
- A rebuild for scale is on the table and the load it is sized against is an estimate.
- The expansion deadline is fixed, and the engineering sequence behind it is not.
Either way it is the same two days: instrument the system, find what is load-bearing, then decide. Only the question at the end changes — which parts of this rewrite are justified, or what breaks first at ten times the load, and what has to be built new.
What lands on your desk
One document, written to be read by an engineering manager and survivable in front of a finance director. Four things in it:
01
What is load-bearing, and what merely still runs
Two very different things that look identical from the outside. One can be retired this quarter. The other is holding up revenue through a path nobody has drawn in three years.
02
Which debt is structural, and which is an expired decision
Most of what gets called technical debt is a decision made under a constraint that has since lapsed, and that is cheap to undo. Genuine platform coupling is not, and the two get quoted at the same price.
03
Which parts of the proposal the evidence supports, and what they cost
Line by line against the plan already on the table, whether that is a rewrite or a rebuild for scale. Some of it will be justified. The report says which parts, on what measurement, and what the defensible number looks like.
04
The three highest-value changes available without a rewrite
Ranked, scoped, and each one deliverable by your own team. If the review earns its fee anywhere, it is usually here.
Why this is cheap to say yes to
Four things settled before you ask. They are also the reason this is a decision you can make this week rather than next quarter.
Analysis only. No code changes.
Nothing we do can break anything, because we are not touching anything. The output is a document and the conversation about it.
No telemetry yet is a finding, not a blocker.
Where the signal exists, we read it. Where it does not, the report says exactly what to log to get it — and a system that cannot answer basic questions about itself has already told you something worth knowing.
Fixed scope, fixed fee, two days.
The scope is written down before it starts and does not move. Two days is short enough that the answer still applies when it arrives.
Approvable without procurement.
Priced and scoped to sit inside what an engineering manager can already sign off. Nobody has to build a business case to find out whether the business case is real.
What the report opens up
Nothing further from us. Your own team runs the list, and that is a fine outcome.
Book the review
A short call first, to confirm the system is a fit, agree the scope in writing, and work out what we need from your team. If it is not a fit, you will hear that on the call.