Capture first, file later
Evidence is usually gathered when someone sits down to write the notice. By then the page has had days to change. The capture belongs at the moment the page is found.
Most evidence for a takedown is gathered at the wrong time. The page is noticed on a Tuesday. Someone is asked to deal with it on Thursday. On the following Monday they open the page to take a screenshot for the notice, and the page is no longer the page that was noticed — or is no longer there.
The notice then documents whatever the page happened to be on Monday. If that differs from what ranked on Tuesday, the notice is wrong in a way nobody will discover until it fails.
Pages do not wait
A page built to rank on someone else's brand is not a fixed object. It is edited when it starts attracting attention. It is moved to a new address when it is reported. It is taken down and put back up. The person running it has every reason to make the version that gets examined differ from the version that ranked.
Even a page that nobody is actively managing changes: the template updates, an advertisement rotates, a paragraph is trimmed. The passages that matched the brand's material last week are not guaranteed to be there this week.
The consequence is that a capture made after the decision to file is a capture of a different page. Sometimes the difference does not matter. There is no way to know in advance which times those are.
The record outlives the page
The only capture that documents the page that ranked is one made when the page was found — before anyone has decided whether to file, before anyone has contacted anyone, before the page has any reason to change.
That capture has to be complete on its own, because it may be the only one there is. It carries:
- the page as it appeared, as an image, not a description;
- the passages on it that match the brand's material, quoted;
- the query that produced it, and the country, language and device of that search;
- the time of the capture;
- the page as the crawler sees it and as a visitor sees it, so that a page showing different things to each is caught then, not later.
Once that exists, the decision to file can wait. The evidence does not degrade. If the page is taken down before a notice goes out, the record shows that it was there. If it comes back at a new address, the new capture can be compared with the old one.
A page removed later cannot be captured afterwards. An uncaptured page is one you can no longer prove existed.
What changes in the workflow
Reversing the order — capture, then decide, then file — changes three things.
The first is that capturing becomes routine rather than exceptional. Every unauthorised page found in the results is recorded the same way, whether or not anyone intends to do anything about it. Most captures will never be used. The ones that are used will be correct.
The second is that the notice becomes a document written from the record, rather than a document that goes looking for a record. The passages are already quoted. The coordinates are already there. The writer's job is to assemble, not to investigate.
The third is that the record accumulates. A page that appears, is captured, disappears, and reappears three weeks later under a new domain is not three unrelated findings. It is one page with a history, and the history is itself evidence of intent.
What this does not solve
A complete capture does not make the host remove the page. Removal is the host's decision, and a notice with everything attached can still be declined. What the capture does is remove the most common reason for a notice to be ignored: that the reader could not verify it without work.
It also does not replace re-checking. After the notice is filed, the search that found the page keeps running. A page that returns under a new address is found and captured again, and the cycle repeats. Filing is a step. Removal is the outcome, and the outcome is something you check.
Read next