Knowledge
Why aren’t the changes visible straight away?
There are four overlapping delays between the publication of a change and its appearance in the model’s response. Consequently, the effect is not visible on the day of publication or during the same week, even if the work has been carried out flawlessly.
Last updated: 7 September 2026 · reading time: approximately 6 minutes
Mechanism
Four delays that overlap
Each one on its own is minor. Taken together, they amount to a timeframe measured in months rather than days.
01
Publication
The change is on your site. The model doesn’t know about it yet, as she doesn’t check it regularly.
Days
02
A return visit
The bot must return to the page and download the new version. The frequency depends on how often the website changes.
Days to weeks
03
Inclusion in the replies
The new content must be added to the dataset from which the model generates responses. Not every retrieval automatically means this.
Weeks
04
Signals from outside your domain
Reviews, mentions and rankings must first be created, then discovered, and only then do they begin to carry weight. This is where the biggest delay lies.
Months
The fourth point is the most important. The first three relate to your website and can be completed within a few weeks. The fourth concerns what others say about you, and it is this that determines whether you are recommended.
Dates
How long does each task take?
The bars show not only the duration but also the start time. Activity on one’s own site begins immediately, whilst external signals only start once they have been generated.
Amendment to an existing subpage
2–4 weeks
New subpage for the query
4–8 weeks
Company details, bot access
2–4 weeks
Reviews on independent websites
1–3 months
Entry in the category list
2–6 months
Publications and presentations
3–12 months
03 months.6 months.12 months.
This is a rough guide, based on observations from our scans. The actual rate depends on the category and on how often models visit your site.
That is why the first report after a month usually shows movement only in content-related metrics, whilst visibility in category-based queries remains static. This is not a lack of effect, but simply a different trend. We describe this in the article knows it, but doesn’t recommend it.
Earlier indications
What you can see before the main sequence begins
The work usually yields results sooner than the single figure at the top of the report suggests. You just need to know where to look.
First scan after implementation: what usually works and what doesn’t
↑
Domain citation index
The model starts to refer to your website, even if the brand hasn’t responded yet.
↑
Audit rating
Technical and editorial changes are immediately apparent, as they are tracked by your own website.
→
Ranking amongst the brands listed
Sometimes a brand moves up from tenth place to fifth, yet still fails to make it into the top three.
–
Visibility in category-based queries
Main frame. It usually remains there for the first one or two scans, and that’s normal.
This is the practical reason why the report breaks the result down into several metrics rather than giving just one. With just a single figure, the first two months appear to show no progress whatsoever.
Practical experience
How not to lose your patience along the way
The greatest risk is not a lack of results, but giving up the work in the third month, just before external factors start to have an impact.
Measure the same thing every monthRhythmThe same set of questions, at the same time. Without this, each measurement measures something slightly different and the conclusions are random.
Look at the breakdown, not just a single figureReadingCitations and the audit process begin before visibility. It is these that determine whether the work reaches the models at all.
Do two things at onceScopeContent on your own website and a single action outside it. In that case, the shorter and longer curves rise in parallel, rather than one after the other.
In this article, we explain how often it makes sense to take measurements and why once a month is the ideal frequency How often should visibility be measured?.
Please note
When something really isn’t working
Patience has its limits. Four situations where a lack of results isn’t just a matter of time.
Four scans with no movement in any of the recordsIt is not just visibility that is at a standstill, but also citations and audits. This is a sign that the changes are not reaching the models, for example due to a block on bots.
All work on your own domainThree months of articles without a single action taken off the website. In that case, the lack of results is to be expected – it’s no surprise.
Tasks completed only partiallyStarted but not finished: a profile has been created but is empty; a subpage has been published without any specific details. The model treats this as non-existent.
Revised set of questionsIf the queries have changed between scans, the two results are not comparable and conclusions regarding the absence of an effect are unfounded.
FAQ
Questions about time
The questions that are most commonly asked between the first and second scans.
Why aren’t the changes visible straight away?
This is because there are four delays that overlap: publication, the crawler’s return visit, the inclusion of content in the results, and the generation and detection of signals from outside your domain. The last of these takes the longest.
How long does it take to see the first results?
Changes to existing subpages may take two to four weeks to become visible. Backlinks from outside your own domain take months to take effect. Therefore, the first reliable indication comes from the second full crawl.
We’ve done everything we can, and the result stands. What next?
Check the breakdown by metrics. Citations and the audit start earlier than visibility. If they’re also at a standstill, the problem lies elsewhere, not with the timing.
Will measuring more often speed up the results?
No. Taking measurements doesn’t change anything, and more frequent scans produce noise rather than a trend. We discuss this in the article How often should visibility be measured?.
Is there any way to speed this up?
To some extent. It’s quicker to work on your own website and gather reviews. There’s no way to speed up inclusion in rankings and publications beyond the pace at which their authors respond.
Why did one model detect the change, whilst the others did not?
The models differ in how often they consult fresh sources. Perplexity usually responds the quickest, as it most frequently uses the latest search results.
When should you conclude that something really isn’t working?
If no metrics, including citations and the audit, change over four scans, this suggests a technical issue – such as a block on bots – rather than a timing issue.
Have a look at what’s moving underneath
You start with 250 AeloCoins, without a card. That’s enough to start tracking and test the modules. Once you’ve used them up, you switch to a paid plan. Measurements work straight away, and a full report on your brand is included from the Starter plan onwards. The results are broken down into several metrics, so you can see changes even when the main figure remains the same.
