The honest ranges
Different parts of this work move at very different speeds, which is why a single answer is always wrong and usually a sales tactic.
| Typical time to show | Why it takes that long | |
|---|---|---|
| Technical fixes | Days to weeks | Crawler access and indexing changes are picked up quickly |
| Schema and entity data | Two to six weeks | Needs recrawling, then cross-referencing against other sources |
| Profile completeness | Two to eight weeks | Feeds Gemini and AI Overviews fairly directly |
| Content restructuring | One to three months | Has to be recrawled, then chosen over what is already cited |
| Reviews and reputation | Three months upward | Depends on your customers acting, not on you |
| Third-party presence | Three to twelve months | Depends on other people publishing things |
The pattern is the useful part: anything you control directly moves in weeks; anything requiring other people to act moves in months. Almost all disappointment comes from expecting the second group to behave like the first.
Why the timeline differs from SEO
With SEO you can watch a position climb. Progress is visible, even when slow.
With GEO there is no position. You are either named in an answer or you are not, and the same question can produce different answers on different days to different people. That makes real progress look like noise.
A business typically moves from never mentioned, to occasionally mentioned, to usually mentioned — and the middle stage is indistinguishable from random variation unless you are recording it consistently. This is exactly why the baseline matters: without one you cannot tell genuine progress from a lucky answer, and businesses abandon work that was starting to succeed.
What should happen, month by month
- Month one — diagnostics and foundations. Crawler access verified, indexing confirmed in Google and Bing, citations audited, profile completed, baseline recorded. You should see nothing at all in the answers yet, and that is correct. If month one is content production instead, the order is wrong.
- Months two to three — fixes take effect. Details start agreeing across sources, restructured pages get recrawled, and the first changes in how you are described appear. Often the detail changes before the presence does: you get mentioned with the right phone number before you get mentioned more often.
- Months four to six — the slow inputs land. Review flow and third-party presence begin to register. You should be named at least occasionally for your main questions.
- Months six to twelve — compounding, or a clear signal. Either it builds, or you have enough evidence that something structural is wrong.
Leading and lagging indicators
The mistake is judging month two by the metric that only moves in month six. Watch these in order:
| Moves first | What it tells you | |
|---|---|---|
| Crawl and index coverage | Weeks 1–4 | The engines can now see you |
| Detail accuracy in answers | Weeks 3–8 | Your entity data is resolving correctly |
| Branded search volume | Month 2 onward | People are hearing about you somewhere |
| Presence in answers | Month 3–6 | The work is landing |
| Enquiries mentioning AI | Month 4 onward | It is reaching actual customers |
Branded search is the most underrated of those. When an assistant recommends someone, the next thing that person often does is search the name — so a rise in branded impressions with no other obvious cause is one of the few early signals you get. The measurement guide covers how to separate it from the rest.
What makes it faster or slower
Faster: an established business with existing reviews, a complete profile, consistent details and pages that already rank. You are adjusting something that works rather than building from nothing.
Slower: a new domain with no reviews, few citations and little authority. Not because GEO is harder in that situation, but because everything is — there is simply less to work with. Trust is the slowest input to acquire and nothing substitutes for it.
Not a factor: budget, beyond a point. Technical and entity fixes happen quickly at any spend. The slow part rests on other people, and money does not reliably accelerate it — which the cost guide is blunt about.
If nothing has happened after six months
Stop and diagnose rather than continuing. Six months of genuine work with no movement is not bad luck, it is evidence. The usual causes, in order of likelihood:
- Crawler access was never actually verified — only assumed from robots.txt.
- Details still contradict each other somewhere you have not looked, commonly an old directory listing or a stale profile.
- The pages still market rather than answer, so there is nothing quotable.
- The named competitors have reputation you have not begun to match, in which case the timeline was never six months.
A GEO audit works through those in that order, and it is a cheaper answer than another six months of the same.