Ecommerce audit

Continue, fix, or stop: one answer, with the evidence behind it.

A review of what has actually been built, commissioned by you and answerable only to you.

A fixed-scope review, delivered in weeks, examining the software, the plan and the way the work is run.

What does an independent review settle?

An independent ecommerce audit is a review of a project by someone with no stake in its continuation. We examine the software that exists against the scope that was agreed, the way the project is being run, and whether finishing it would still serve the business. The answer comes back as one of three: continue as you are, fix specific things and continue, or stop.

Three things we look at, in this order

What was built against what was agreed

The scope on paper, the software that exists, and the distance between them. Working software is the evidence here, so we run it, read it, and check what happens under the conditions the business actually creates.

The scope on paper, the software that exists, and the distance between them. Working software is the evidence here, so we run it, read it, and check what happens under the conditions the business actually creates.

How the work is being run

Estimates against outcomes, how decisions get made and recorded, how change requests are handled, how much of the knowledge is in one person. A project rarely fails on the code alone.

Whether the plan still leads anywhere

The original plan was written when less was known. The question is whether finishing it would give the business what it needs now, at a cost worth paying.

‍

The review covers the technical side, including a Magento audit or the equivalent on your own platform, and the commercial and organizational conditions around it. We work from the artifacts: the repository, the tickets, the environments, the contracts, the meeting records. We also talk to everyone involved, your team and your suppliers alike, and check what we hear against what the work shows.

The verdict: continue, fix, or stop

Continue

The work is sound, the plan holds, and what you have been feeling is the ordinary discomfort of a project in its difficult middle. It is a real outcome, and a review that could never return it would be worth nothing.

Fix and continue

The direction is right, and specific things have to change: the environments, the testing, the way scope is agreed, and sometimes how responsibilities are divided. The report says which, in what order, and what each one costs to leave alone.

Stop

Finishing would cost more than starting again, or the thing being built stopped being the thing the business needs. Stopping is the hardest answer to hear, and acting on it early is usually cheaper than finishing. It comes with a list of what to salvage, since a stopped project rarely loses everything.

‍

Each answer arrives with the evidence behind it and with the conditions under which a different answer would have been correct. You get one recommendation rather than a menu, because a review that ends in options hands the decision back to you unchanged.

Independent of everyone in the room, including us

Our own position is stated in writing at the start: our engagement is the review itself, and any later work is a separate decision you make once the verdict is in.

That matters because people inside a project see it from their own position, however good their intentions. The supplier building it, the team that commissioned it and whoever hopes to take on the next phase each bring a view shaped by their role, and a review adds one from outside all three. We have told clients to rework a plan before building on it, when the build would have been ours, and told others that a rebuild they were ready to fund would leave the actual problem in place.

The other half of independence is what goes in the report. It names what happened rather than who to blame, because a report that reads as blame gets argued with instead of acted on. The recommendations are written to be carried out by the team and suppliers you already have, and we are glad to work with them.

How the review runs

The question, agreed in writing

What decision the review has to support, and what you would do with each of the three answers. Agreeing this takes an hour, and it decides what the review examines.

Evidence gathered

Repository, environments, tickets, contracts, plans, and the software itself under real conditions, plus conversations with the team and the suppliers.

Findings tested before they are written

Anything serious is checked twice and discussed with the people it concerns before it reaches the report. A finding that holds up in that conversation is one you can rely on.

One recommendation

Short enough for the person who signs to read in full, with the working behind it available to whoever has to act on it.

Before a project starts, while the decision is still open, the right engagement is a discovery.

Systems Architecture & Discovery →

Reviews that changed what happened next

Our Workplace

Our Workplace

The review came first, and the rebuild it recommended never happened.

A compliance platform that had been running for years and could no longer carry what the business needed. The review found a framework and a language version past their supported life, an absence of automated testing, and hosting unsuited to the load. Our recommendation was to build the replacement alongside the running system and switch off the old parts one at a time, with security ahead of any new feature, since the platform holds people's identity documents.

Maxilia

Maxilia

A replatforming prepared on rebuilt requirements.

A promotional products business whose shop, ERP, order management and customer records had grown into a single application. The review and the workshops that followed found that the existing replatforming plan described the old system more than the business it serves. We recommended rebuilding the requirements first, then prepared the replatforming on that basis, alongside interim technical leadership and the delivery team.

‍

Bring us the project as it stands today.

Book a project review →

What arrives at the end

The verdict, on one page

Continue, fix, or stop, with the reasoning, written for the person who has to decide.

The findings, with evidence

What was examined, what was found, and how serious each thing is, separated from opinion.

What to do about it

Ordered by what it costs to leave alone, with an estimate against each and a note on what can wait.

What is worth keeping

The parts of the work that hold up, so a decision to stop is a decision about direction and not about everything.

What a verdict changes

The conversation moves from recollection to evidence. One document exists that everyone in the room has read, with the evidence attached, and the discussion moves on to what to do about it.

The spending changes too. A project that should stop stops costing money, and one that should continue is funded against the same document.

Where we fit, and where we do not

The fee is small next to what it settles: money leaving every month, progress you have no independent way to judge, and a decision worth more than the cost of putting evidence behind it.

A marketing performance problem needs a different specialist: we review whether the software and the project behind it are sound, and questions about conversion rate, campaign performance or search rankings belong with people who do that for a living.

FAQ

What does an ecommerce audit cover?

The software as built, the process building it, and whether the plan still fits the business. On the technical side that means the codebase, the architecture, the environments, the testing and the platform's own health. On the commercial side it means scope against spend, how decisions are made, and what finishing would cost against what it would return.

What do we do when a project has stalled or failed?

Stop adding budget until you know what you have, because money spent inside an unexamined plan buys more of whatever went wrong. Get an independent view of what is built and what the remaining distance actually is. Then decide once, with evidence, instead of every month in a status meeting.

What is a code audit?

A structured read of the codebase and its surroundings: how it is organized, what it depends on, whether it is tested, whether it can be safely changed by someone new. It answers whether the work can be built on, which is a different question from whether it is elegant. Most of our audits are broader than this, because code is rarely the only thing that has gone wrong.

Can you review a project while it is still running?

That is cheaper than waiting for it to stall, and we do it. We have run sprint-level oversight alongside another supplier's build, checking requirements at the start of each sprint and the work at the end. That arrangement catches the drift that produces a rescue engagement a year later.