What clients should receive when a website is declared launch-ready

Most website launches end with a Slack message. "We're live 🎉" — and then a silence where the paperwork should be. The site is up, the invoice is out, and the client owns a thing they cannot describe, backed by verification they never saw, with open questions nobody wrote down.

It works fine until it doesn't. Three weeks later a form breaks, or a competitor points at a page, or the client's board asks what exactly was checked before launch — and the agency's answer is a memory. The launch was real; the handoff never happened.

A launch-ready handoff is a package. Not a ceremony, not a hundred pages — a short set of artifacts that together answer the four questions a client is entitled to ask: what did we decide, on what evidence, what is still open, and what happens next?

The decision, in writing

The centrepiece is one paragraph most agencies never write down: the launch decision itself.

Every clause is doing protective work — for both sides. Which build pins what the decision covers. Shipping as known converts future discoveries from surprises into references. Outside it marks where the agency's verification ends and the client's ongoing ownership begins.

The score, with its meaning attached

A score earns its place in a handoff only when it arrives with what it measures and what would move it. A bare number invites the client to treat 82 as a school grade; the same number as a worked ledger reads as engineering.

Note the held-back row: things only the client can confirm are reported but not scored. Folding unresolved questions into the number would make the score claim knowledge nobody has — and the first time a client works that out, the number stops meaning anything.

Blockers resolved, and blockers remaining

The handoff lists both — what was found and fixed during the review, and what ships open. Agencies hide the first list out of instinct: why show the client the mistakes? But the resolved list is the QA work made visible. It is the difference between "we checked it" and evidence that checking found things worth finding.

The open list is smaller but sharper: each item carries an owner and a date.

Those two labels do a specific job here: they mark the items a client might otherwise rediscover as "bugs" — the intentionally hidden staging area, the form that needs their CRM credentials to finish — as decisions with names attached.

Strengths worth preserving

One section almost nobody includes, and the cheapest goodwill in the document: what the site does well, stated as things to protect. The clean metadata, the fast product pages, the accessible checkout. Two sentences — but now the next redesign conversation starts from "preserve these" instead of from zero, and the client learns their agency can tell good from lucky.

Decisions the client still owns

Every launch ships with questions that were never the agency's to answer: the tone of the legal pages, whether the founder's photo stays, the analytics consent posture. The handoff names them and hands them over explicitly. An unnamed open decision defaults to the agency owning it forever — silently, and for free.

The developer tasks, with acceptance criteria

For the open technical items, the handoff's implementation layer is the same discipline as any good ticket: the evidence, the location, and a criterion someone can verify without re-arguing the finding.

Contact form error state: replace the raw 502 response with the designed error component. Acceptance: submitting while the endpoint is unreachable keeps the visitor's message on screen and shows the retry state. Evidence and capture in the launch report, finding 7.

Tasks written this way survive the handoff — including into a coding assistant, which is increasingly who picks them up first.

Before and after, when fixes were made

When the review found things and the team fixed them, the handoff should show the movement — the state at first inspection against the state at launch. A before/after is the single most persuasive artifact an agency can attach, because it captures the thing the client actually bought: not a scan, but the improvement between two scans.

This is also the honest boundary of a one-pass review. A single inspection can only describe a moment; demonstrating improvement needs the second pass after fixes. That is exactly the gap the Launch Audit tier exists for — its shipped scope is the expanded audit plus a re-scan after fixes with before/after scores, which is this artifact as a product.

The verification path forward

Last page of the package: what happens after launch. Which parts of the site the client can change safely, what should trigger a re-check, and how the next inspection compares against this one. One paragraph — but it turns the handoff from a full stop into the start of a maintenance relationship, which is where agency QA stops being a cost and starts being a line item.

The whole package is five artifacts: the decision, the scored evidence, the open items with owners, the preserved strengths, and the path forward. An afternoon's work with the right structure — and the difference between a launch the client remembers as a message, and one they can defend to their own board.