How long does SEO take to work?

Every answer to this question quotes the same range: three to six months. It is repeated so consistently that it has become the industry’s standard reply, and it is wrong in a specific and damaging way.

It is wrong because SEO is not one activity with one lag. It is at least four activities with wildly different lags, and averaging them produces a number that is too pessimistic for the fast work and dangerously optimistic for the slow work. A client told “three to six months” who could have seen click-through improvements in three weeks concludes nothing is happening. The same client, told the same range about a competitive head term, concludes at month seven that they were misled.

Why “three to six months” misleads

Consider two pieces of work commissioned on the same day.

The first is a rewrite of title tags on forty pages that currently receive impressions at an average position of nine but a click-through rate under 1.5%. Those pages already rank. Nothing about their authority needs to change. Once Google recrawls them, the new titles are what searchers see, and behaviour changes immediately. Movement is typically visible in two to four weeks.

The second is a campaign to rank for a two-word commercial term where the incumbent results are national brands with a decade of links. No amount of on-page work changes that in six months. The realistic horizon is a year or more, and it depends on work that has not started yet.

Both are SEO. Both were quoted “three to six months”. The first under-promises by a factor of four; the second over-promises by a factor of three. This is the entire reason SEO has a reputation problem with people who fund it.

Timelines by type of work

Read that table as a sequencing instruction rather than a reference. If someone needs to see results by a specific date, the top four rows are the only ones that can deliver. Starting at the bottom because it is where the big volume is means the first six months produce nothing reportable, which is usually when funding stops.

Where the delay actually comes from

The lag is not arbitrary. It is three separate delays stacked on top of each other, and knowing which one you are in tells you whether to wait or intervene.

Discovery lag. Nothing happens until the page is recrawled. On a site crawled several times a day, that is hours. On a small site with few inbound links and a shallow internal structure, it can be weeks. This is the one you can most directly influence — submit the sitemap, link the page internally from something crawled often, and make sure nothing blocks access.

Evaluation lag. Once indexed, the page enters a period where search engines are working out what it deserves. They test it at various positions and watch what happens: whether people click, whether they come back and search again, whether the page satisfies the query relative to alternatives. This needs traffic to generate, which is circular and is why low-volume queries can take longer to settle despite less competition — there simply is not enough interaction to judge from.

Competitive lag. The slowest and least controllable. Your page is being compared against results that have accumulated years of evidence. Overtaking one requires enough countervailing evidence to overcome that, and evidence accumulates at the speed of the outside world linking to and discussing you. No process improvement compresses this materially.

Diagnosing which lag you are in is straightforward. No impressions at all means discovery. Impressions at a volatile position means evaluation. A stable position outside the top ten for months, with a page that is genuinely better than what is above it, means competitive — and that is the only one of the three that needs a different kind of work rather than more patience.

What should be visible at each checkpoint

The useful discipline is knowing what to expect at each point, so that “nothing is happening” becomes a measurable claim rather than a feeling.

Week 2. Fixed pages are being recrawled; indexed count starts to move if it was the problem. Nothing else. Do not read anything into traffic at this point.

Week 4. Rewritten titles show measurable click-through change on the affected pages. Impressions on new pages begin appearing at low positions. Total traffic likely unchanged.

Week 8. Average position on target queries should be improving, even if clicks have not followed. This is the most important early checkpoint: position moves before traffic, because positions 15 to 11 produce almost no clicks but represent real progress. A site that has not improved average position by week eight has a problem worth investigating.

Month 4. The first meaningful traffic change on non-branded queries. Pages that were at 8–20 should be entering the top ten. New pages on low-competition terms should have found a stable position.

Month 6. The trend should be clear enough to extrapolate. Top-ten query count should be visibly higher than at the start. If it is not, the strategy needs changing rather than extending.

Month 12. Competitive terms become contestable if the earlier work built authority. This is where compounding starts to be obvious, and where sites that stayed consistent separate sharply from sites that restarted twice.

Track those checkpoints with the same three numbers each time — indexed pages, top-ten query count, non-branded clicks — as covered in our guide to monitoring SEO performance. Changing which metric you report between reviews is how teams accidentally convince themselves of whatever they hoped.

What legitimately speeds things up

Not shortcuts, but genuine accelerants.

• Working on pages that already rank. The single biggest time saver available. A page at position 12 has completed discovery and evaluation; only the last stretch remains.
• Improving crawl frequency. A clean sitemap, a shallow internal structure and regular publishing all raise how often you are visited, which compresses discovery lag across every subsequent change.
• Choosing queries with weak incumbents. Reading the SERP before committing removes competitive lag entirely on the queries where the top results are dated or mismatched.
• Fixing one thing at a time. Counterintuitive, but changing five variables at once means the result is uninterpretable, and uninterpretable results lead to repeating the wrong work.
• Having something genuinely linkable. Authority accumulates faster when there is a reason for it to, and one strong asset outperforms months of outreach against ordinary pages.

What quietly resets the clock

More important than the accelerants, because these are self-inflicted and common.

Changing URLs without redirects. Discards everything the old URL accumulated and starts the page from zero. A surprising number of site redesigns do this across an entire domain, and the traffic loss is usually blamed on an algorithm update that happened to fall nearby.

Rewriting pages every quarter. Each substantial rewrite triggers re-evaluation. Doing it repeatedly means the page never leaves the evaluation phase, and a page permanently under evaluation never stabilises anywhere good.

Switching target queries mid-flight. Retargeting a page at a different intent at month three throws away three months of accumulated relevance signals for the original.

Publishing volume that dilutes. Adding forty thin pages while working on twelve good ones lowers the average quality of the domain and spreads crawl budget thinner. It also creates the cannibalisation that then takes a quarter to unwind.

The pattern across all four: impatience produces actions that extend the timeline. This is the actual reason most SEO efforts fail — not wrong tactics, but a review cycle faster than the feedback cycle, which guarantees decisions get made on noise.

When to conclude it is not working

Patience is not the same as blindness, and there is a point at which continuing unchanged is the wrong call.

The honest test is leading indicators. If after four months of consistent work your indexed page count has not moved, your impressions have not risen, and your average position on target queries is flat, then the strategy is not slow — it is wrong. Those three metrics move well before traffic does, and their flatness after four months is a genuine signal rather than impatience.

The most common underlying causes, in order: the target queries were never realistic for the site’s authority; the pages are not actually indexed; several pages compete for the same query and none wins; or the content answers a different intent than the query expresses. Each has a different fix, and all four are diagnosable in an afternoon — the process is in our guide to why a site is not ranking.

One more thing worth measuring while you wait, because it moves on a different timescale entirely. AI assistants now answer a large share of the research questions that used to begin as searches, and they select sources by passage relevance and freshness rather than accumulated authority. A page can be cited in an AI answer months before it ranks, which means a project that looks stalled in Search Console may already be visible where the buying research is happening. It is invisible unless tracked deliberately, and it is the one indicator that can move inside the window where everything else is still accumulating.

Frequently asked questions