The most expensive thing about changing agencies is not the transition fee. It is that eighteen months of accumulated knowledge about a specific site — what was tried, what broke, why a particular redirect exists — usually does not transfer at all.
The new team starts by rediscovering it, often by breaking the same things. The client pays twice for the same lesson and it never appears on an invoice.
What actually needs recording
A handover document is not a report. Reports describe performance, which the analytics already do. A handover describes decisions and their reasons, which nothing else captures.
The redirect map, with reasons. Every non-obvious 301 should say what it replaced and why. Redirects with no explanation are the single most common thing a new team removes and then has to restore.
Anything deliberately noindexed, and why. A page excluded from search on purpose looks identical to one excluded by mistake. Without a note, the next person will either fix a non-problem or leave a real one.
Migrations, with dates. URL structure changes, domain moves, platform replatforms. Almost every strange pattern in a site's history traces to one of these, and knowing the date makes an inexplicable chart legible in thirty seconds.
Manual actions and recoveries. If anything was ever filed, what it was and what resolved it.
Things tried that did not work. The most valuable section and the one nobody writes, because it reads as a list of failures. It is the only thing that stops the next team spending a quarter repeating them.
Would someone competent, arriving with no context, make a worse decision without this? If yes, write it. If they would reach the same conclusion from the site itself, skip it.
Access is not documentation, but it is where handovers fail
The practical half is duller and blocks more work. Who owns the Search Console property, and is it a domain property or a URL-prefix one. Which analytics property is authoritative when there are four. Where DNS is managed. Where the site is actually hosted, which is frequently not where everyone believes.
Ownership matters more than access. An agency-owned Search Console property disappears when the relationship ends, taking the historical data with it. The client should own the properties and grant access, not the reverse, and that is much easier to arrange at the start than during a separation.
The site's own history
Some context cannot be reconstructed from the site because it is not in the site.
Why certain pages exist for legal or regulatory reasons. Which pages convert despite negligible traffic. Which URLs have external links pointing at them from somewhere that matters. Seasonality that looks like a decline if you have not seen it before.
That last one causes real damage. A business whose demand halves every January will look like it is collapsing to anyone reading the chart without knowing the pattern, and the reaction to an imagined collapse is usually worse than the dip.
Keeping it current without it becoming a project
Handover documents rot because they are written once, at the end, when nobody wants to write them.
The version that survives is short and appended to as things happen — a dated line when a redirect is added, a paragraph when a migration completes, a note when something is tried and abandoned. Five minutes at the time replaces a week of archaeology later.
It also changes the quality of decisions while the work is still running. A team that has to write down why it is doing something reconsiders a certain number of those things before doing them.
Who this is really for
It is easy to frame this as protecting the client from the agency. It is at least as much the other way round.
An agency inheriting a site with a real history attached can be useful in week one instead of week six. And an agency that leaves a proper record behind is judged on the work rather than on the confusion that follows its departure — which, for anyone doing the job honestly, is the better trade.
Inheriting a site with no history attached?
We will reconstruct what was done, what changed, and what is still load-bearing, so decisions stop being guesses.
Get in Touch →Does this apply to your site?
Reading about it is one thing. Point the scan at your own site and see whether this applies to you, and what it is worth fixing.
Free and unlimited. No account, no card, and you get every finding rather than a teaser.