Six fields that decide whether a published Search Console chart means anything
Check them in this order on somebody else's chart and most of what gets published in this category stops being an argument within a minute.
Published by AI Knows Us · Updated 19 October 2026
Six fields decide whether a published performance chart carries information: the window, the days with data, the export date, the clicks, the impressions and the average position. Checked in that order, the first missing field stops the read, because every field after it depends on the ones before. Below is each field, what its absence hides, and the one worked example this site publishes with all six present. That example is the aiknowsus.com case on our proof page, and it is used here because a sequence of checks is only demonstrable against a file that passes it.
The order, and why it is not negotiable
The order runs from the fields that define the measurement to the fields that report it. A clicks figure with no window attached is not a weak number, it is not a number: there is no period it describes. So the sequence starts with scope, moves to completeness, then to the recording date, and only then looks at what was counted.
| Order | Field | What its absence hides |
|---|---|---|
| 1 | The window, both dates | Which period the figures describe, and whether the reader can reproduce the filter |
| 2 | Days with data | The divisor behind every daily average on the chart |
| 3 | The export date | How settled the data was when the file was pulled |
| 4 | Clicks | Whether anybody arrived, as distinct from being shown |
| 5 | Impressions | Whether the pages are in the results at all |
| 6 | Average position | Where in the results, which is what converts one column into the other |
Fields one to three define the measurement
The window has to be two dates rather than a duration. A chart labelled last ninety days was true on the day it was made and is unreproducible afterwards, because nobody knows which ninety days were meant.
Days with data is the count of days inside the window that carried a row. It is the divisor, and it goes missing more often than any other field on this list.
The export date is the day the file was pulled, and it is a third thing again. A performance report does not carry the most recent day or two, so a chart whose window runs to the export date is showing rows that were still settling.
Fields four to six report it, and only together
Clicks and impressions answer different questions and a chart with one of them is a chart with an argument in it. Impressions say the pages were in the results. Clicks say somebody chose them. Average position is the bridge, because position is what turns one into the other.
A chart showing impressions alone reads as success and may be reporting that nobody arrived. A chart showing clicks alone reads as modest and may be hiding that the site is shown constantly and chosen rarely, which is a different problem with a different fix.
The worked example, all six fields present
Here is the aiknowsus.com case with each field named, so the sequence can be run against something real rather than described.
- Window: 2026-10-04 to 2026-10-12.
- Days with data: 9.
- Export date: 2026-10-14.
- Clicks: 3.
- Impressions: 5,965.
- Average position: 41.6, recorded on the best day of the window, 11 October 2026, which carried 2,685 impressions.
Run the six checks on that and the result is a short, defensible reading. The pages are being shown. Almost nobody is arriving. The position figure explains why, and the day count says the domain has been in existence for nine days. Every one of those statements is available from the fields and none of them requires a word from us.
What the sequence does to a chart that fails early
The value of a fixed order is that it stops you arguing about the wrong thing. If field one is missing, there is nothing to discuss about fields four to six, and the right response is to ask for two dates rather than to debate the size of the number.
That saves the conversation most of its length. 101 of 102 build spec conversations run on ChatGPT and Gemini on 15 October 2026 asked for a before and after with the measurement window stated, and 102 of 102 asked for a date, a last verified line and a refresh frequency. A chart missing field one is missing the most requested item in the entire specification, and no amount of detail further down compensates.
For what each of the six figures actually measures rather than how to audit it, the Search Console guide to the reports that matter is the reference to keep open beside a chart you are checking.
The one surface where none of the six fields exists
This sequence works because Google records the data and hands you a file. Ask ChatGPT, Gemini, Perplexity or Claude which product to buy for the thing you sell and there is no console, no window and no export: there is one written answer naming two or three companies, and nothing to pull. A figure about that surface exists only if somebody puts the question and writes down what came back, with the date and the apps named.
That is the measurement we sell, and the comparison worth reading first is what Search Console can measure against what an AI visibility tracker can, because the two answer different questions and only one of them produces the six fields above.
Questions about auditing a published Search Console chart
Which field is missing most often?
Days with data. It is visible on the screen the chart was made from and it does not survive being passed to whoever builds the slide, so it disappears without anybody deciding to remove it.
Field six came back as a single site wide figure. Can I use it?
Use it for direction over months and not as a location. The figure is averaged across every impression inside the filter, in every place and on every device, so acting on one query means filtering down to that query first. A site wide number is a trend line wearing the clothes of a rank.
What if the only figure offered is a percentage increase?
Ask what the first number was and what the last number was. A percentage contains both and discloses neither, and the two absolute figures contain the percentage, so the pair is strictly more information than the percentage was.
Can I run these six checks on a competitor?
Not from the outside, because a performance export needs verified ownership of the site. What the six fields are for is reading charts other people have published about themselves, which is most of what gets shown in a sales meeting.
How long does the sequence take in practice?
Under a minute on a published page, because the first three fields are either printed above the chart or they are not. The part that takes longer is the conversation afterwards, which is why having the fields ready on your own pages shortens a review rather than extending it.