Average position 41.6 on the best day: why position moves later than discovery

Discovery and ranking run on two separate clocks. Being found needs a page indexed and judged relevant, which can happen within days. Holding a place near the top needs standing a page has to accumulate, which does not. The reading on our proof page is the pair of clocks caught mid stride: on 11 October 2026, the best day in the aiknowsus.com export taken 2026-10-14, the domain was shown 2,685 times at an average position of 41.6. The first clock had clearly moved. The second had not started.

What an average position of 41.6 is a reading of

It is not a rank you hold. Search Console averages the position of your best result across every impression a query produced, in every location and on every device, then averages that across every query in the window. So 41.6 is a weighted average over thousands of appearances, most of them several screens down, not a statement that any particular page sits at number 41 for anything.

That distinction matters for a new domain more than for an old one, because a new domain's impressions come from a wide spread of queries it has only just become eligible for. Each of those contributes a position to the average. The figure is a description of the whole spread, which is the right unit at this stage and the wrong one for diagnosing a single page.

Why the two clocks cannot be made to run together

Indexing is a decision about eligibility. Ranking is a decision about preference, and preference is informed by evidence that takes time to exist: other pages linking to yours, people choosing your result when they do see it, and a history of the site answering the query well. A nine day old domain has none of that evidence yet, not because anything is wrong with it but because nine days is how old it is.

So the order is fixed. Impressions appear, the set of queries widens, position tightens, and clicks arrive last. Reading that sequence backwards, and treating flat clicks in week two as a verdict, is the most common way a working programme gets stopped early.

The practical consequence for a new site

While the second clock is still catching up, the work that pays is the work that widens eligibility rather than the work that chases position:

• Publish against questions nobody has answered properly, because thin competition is where a new domain can be served at all.
• Count distinct queries week over week. A widening query set is the early signal, and it moves before position does.
• Leave the position number alone for the first month. On a spread this wide it moves for reasons that have nothing to do with the last thing you changed.
• Fix the snippet once impressions are real, because that is the one lever that converts an impression without needing a position change. Worked examples of that rewriting sit in before and after copy for every description field.

What the figure would have to be before clicks follow

Click share falls steeply with every place down the page, so the distance between an average of 41.6 and a position where choosing you is a realistic option is most of the journey. That is why 3 clicks sits beside 5,965 impressions in the same nine day export and why neither number is surprising. The honest description of that window is that the domain became findable. A description that implied it had started ranking would be a different claim about the same two columns.

Knowing that also fixes what to watch for next, and it is not clicks. It will be the average tightening as the earliest pages gain a little standing and stop entering queries from the bottom, followed by the first handful of queries crossing onto a screen people actually read. Watching for clicks before that happens is watching the wrong figure, and watching the position average without the query count beside it is watching a number that two opposite events can move the same way.

How to report a position figure so it survives a sceptic

Three things travel with it or it will be misread. The window, because an average over nine days and an average over ninety are different statistics. The export date, because Search Console revises recent days. And the impression count it was averaged over, because the same position average computed over 5,965 impressions and over a few dozen is not the same evidence at all. Across 102 build spec conversations we ran on ChatGPT and Gemini on 15 October 2026, 80 asked specifically for the sample size to be attached to the figure, which is the same request in a different vocabulary.

Questions about average position on a new domain

Is an average position in the forties bad?

For a domain that is nine days old it is the expected reading, and the useful comparison is against the same figure a month later rather than against somebody else's site. For an established site on commercial queries it is a signal to work on the pages sitting between the second and third page, where a few places move click share most.

Why does my average position get worse when things are going well?

Because every query you have newly become eligible for joins the average at the place it entered, which is usually a long way down. The average can therefore slip in the same month the site gets stronger. Read it with the query count beside it and the apparent contradiction disappears.

Can I speed the second clock up?

You can shorten it by being specific. Pages that answer a narrow question completely compete against fewer alternatives, and standing accumulates faster where the field is thin. Breadth on contested head terms is the slow route for a new domain.

Does a position average tell me anything about one page?

Very little. Filter to that page and that query before drawing a conclusion, and take a manual reading for a single query you care about. The site wide average is a trend instrument.

Why publish this figure at all?

Because we sell the measurement and the pages that follow it, and a company selling measurement that only publishes its flattering numbers has not demonstrated the thing it is selling. 41.6 on our best day is the figure that makes 2,685 impressions on the same day believable.

What the next reading of this figure has to show

A position average is only worth publishing if the next one is published too, so it is worth saying in advance what would count as the second clock starting. Three things, in this order: the average tightening while the query count holds or rises, the first queries appearing with a position inside the top twenty, and clicks arriving on those queries specifically rather than spread thinly across everything. If the average tightens because the query count fell, that is the opposite of progress wearing the same number. Writing that down before the next export is the difference between a measurement and a figure chosen after the fact, and it is why the window and the export date sit on the published version of this one.