Check · 25 minutes · 5 questions
“You now rank in the top ten for 47 keywords, up from 12.”
Ask for these first
In writing, before the meeting. The request is half the check: everything below is one line to supply if it exists.
Your own name, product names and misspellings of both. You ranked for those before the engagement and will after it, so they measure the supplier's keyword list rather than their work.
If insteadThe list is not supplied, or arrives with branded terms mixed in and unmarked.
A top-ten position on a phrase queried twice a year costs almost nothing to obtain and returns exactly that. The count rises, the business does not.
If insteadVolume is described as “long tail” rather than given, or the totals are quoted without it.
Two instruments, one simulated and one recording real appearances. Close agreement means the tracker is configured sensibly. A tracker showing positions the engine's own data does not corroborate is measuring a search nobody performed.
If insteadSearch Console access is not offered, or the comparison is declined as “not comparable” without saying why.
A rank counts the ten links. An AI overview, a map pack, a shopping row and four ads sit above them, and a third place under all of that is at the top of a list the reader never reaches.
If insteadNobody has looked. Ask them to screenshot the actual result page for the three keywords they are proudest of.
The point of a position is a visitor. Positions improving with clicks flat, over a period long enough to matter, means the improvement happened where nobody was looking.
If insteadClicks are reported from a different tool than the one used last month, or the comparison period is chosen to start at a trough.
Our own reporting on this domain
We sell ranking and citation reporting, and we run no rank tracker on our own site. Everything we know about our positions comes from Search Console — for the twenty-eight days to 16 August 2026, 1,950 impressions, 6 clicks, an average position of 35.9 across 192 queries. That is a better instrument than a tracker, because it records where we actually appeared rather than where a datacentre simulated it, and until this week we had not published anywhere what the difference is.
SinceThe vocabulary now says it: rank tracking, average position, impression and click each carry what they measure and what they cannot. Our own 0.3% click-through rate is the worked example on the CTR entry, chosen because a flattering one would have demonstrated less.
A supplier who cannot supply a base and a period has told you something, and it is not that they are disorganised. We do this for a living and the questions land differently when they come from outside.