How Long Before Marketing Actually Starts Working? The Honest Timeline

Quick Takeaways

1. Foundations move fastest. Alpha Ski Tomamu went from 50 to 90 out of 100 on our AEO audit in seven days.
2. Our own site took three months to go from 54 to 91, because the bigger the site, the longer the structural work runs.
3. Revenue lags the score, always. Harvest Bloom Flowers was rebuilt in the last two weeks of July, the campaign went live on 1 August, and the sales followed after that.
4. Two weeks is a data window. Four to six weeks is the first fair judgement. Twelve weeks is where compounding starts to show.
5. The work that produces the visible result happened weeks earlier, and almost none of it was interesting at the time.
6. There is a real difference between "too early to judge" and "this is broken", and the signals are different.
7. Consistency beats intensity. Most campaigns are killed in the exact week they were about to start paying.

The week nothing happened and everything landed

Earlier this month we had a week with no launch in it. No new campaign, no fresh budget tipped into anything to make a report look healthy. Work we had finished weeks before simply turned up on the scoreboard at once.

Harvest Bloom Flowers started getting sales on the Sunday and they kept coming all week, finishing up 580% off a campaign spending $20 a day at a cost per conversion of $2.31. A pool shop ecommerce campaign at $30 a day did $2,201 in online sales. An organic Facebook post I had written for an accountant a fortnight earlier passed 750 likes and delivered around 70 leads, with zero ad spend behind it.

When people see a week like that, the first question is always about the big thing that caused it. What was the campaign, what was the hook, what targeting did you use. There was no big thing. Every number in that week was produced by work that had been done well before there was anything worth screenshotting.

That is the honest answer to how long marketing takes. It takes as long as the invisible part takes, and then the visible part arrives all at once and looks like luck.

What moves quickly: the foundations

Some of this work is genuinely fast, and it is the part most operators skip.

AEO, or Answer Engine Optimisation, is the work of making your website legible and trustworthy to AI tools like ChatGPT, Gemini and Google AI Overviews. We score a site out of 100 across things like schema markup, entity clarity (whether your business is consistently named and described), crawler access, question-shaped content and review signals.

Those scores move in weeks. Alpha Ski Tomamu went from 50 to 90 in seven days. Harvest Bloom went from 57 to 98. Our own site at Infinity1 went from 54 to 91 over three months, and it took longer precisely because there was more of it and we were fixing it in public.

A score is a leading indicator. It tells you the machinery is right. It does not tell you the money has arrived yet, and anyone promising you both in the same fortnight is selling something.

What moves slowly: the money

Harvest Bloom is the clearest case I have. The site was rebuilt in the last two weeks of July. The campaign went live on 1 August. With a few days of the month still to run, August has already produced 84% of the year's revenue for that business.

Read the order of events again, because it is the whole lesson. Rebuild, then launch, then wait, then sales. The structural work happened in July and produced nothing in July. Anyone judging that project on 31 July would have concluded it had failed.

The same pattern shows up in social. That accountant post did not work because I wrote a good Facebook post that morning. It worked because I had spent years living the story behind it and knew exactly how Kiwis in Australia talk to each other. Four posts before it got about four likes each.

The part nobody films

There is a line that has been rattling around my head all month. Hard work beats talent when talent doesn't work hard.

It sounds like something painted on a changing room wall. What I take from it has nothing to do with grinding yourself flat so you can tell everyone how busy you are. It is about consistency, and specifically about doing the work in the stretch where there is no reward for doing it.

Turning up when leads are rolling in is easy. Turning up when the scoreboard has sat still for five minutes is a different skill, because that is when you start wondering whether the audience cares, whether the content is any good, and whether the business down the road knows something you do not.

That is also the precise moment most operators blow the whole thing up. They kill a campaign before it has enough data to be useful, stop posting because the last few went nowhere, or swap the strategy for whatever phrase turned up in their feed that week.

Nobody sees the pages being rewritten. Nobody sees the tracking being checked for the third time because something still looks off. Nobody sees you at two in the morning fixing a pixel because Meta is matching the wrong people and the campaign will never work properly until it stops. They see the result, and then they call it talent.

A realistic week-by-week timeline

Here is what to expect, assuming the foundation work has been done and tracking is clean. Use it to work out whether you are early or in trouble.

Timeframe	What should be happening	What worries me
Week 1	Audit complete, score baseline set, fixes underway. No revenue signal at all.	Nobody can tell you what your score is or what is being fixed.
Week 2	Structural fixes live. Score re-run and moving. Campaign launched, cost per click settling.	Tracking still firing incorrectly, or the pixel matching the wrong people.
Weeks 3 to 4	Click-through rate readable. First enquiries or first sales appearing. Cost per conversion high but falling.	Plenty of clicks, zero enquiries. That is a site or offer problem.
Weeks 5 to 6	Cost per conversion stabilising. Return on ad spend measurable for the first time. Winning creative identifiable.	Return well under 8:1 with clean tracking and a good score.
Weeks 7 to 12	AI tools and search starting to surface your pages. Organic enquiries joining paid ones. Budget worth scaling.	Score has climbed but enquiry volume is flat across the whole quarter.
Month 4 and beyond	Compounding. Content published in month one still earning. Cost per customer trending down.	Nothing has moved since week six. Time for a genuine strategy conversation.

Five checkpoints that tell you the truth

If you want a simple way to judge a campaign without guessing, work through these in order.

1. Check the tracking before you check the results

Confirm the pixel or tag fires on the pages that matter and matches real people. A campaign built on broken tracking will look like a budget problem for months. This is the first thing we check and it is wrong more often than you would believe.

2. Score the destination

If your site sits below 75 out of 100, the campaign is being asked to carry a page that cannot hold its end up. Our policy exists for this reason: we won't run your ads without doing an audit and getting the score above 75.

3. Read the clicks separately from the conversions

Clicks tell you whether the message lands. Conversions tell you whether the page and the offer hold up. When clicks are healthy and conversions are absent, your ads are working and your site is the problem.

4. Give the offer six weeks, not two

Changing the offer every fortnight destroys the only clean signal you have. Judge the offer at six weeks with real numbers behind it, then change one variable.

5. Ask what compounding has been switched on

Published content, structured data, reviews and citations keep working after you stop paying for them. If none of that is being built, you are renting attention every month and starting from zero each time.

So, how long?

For foundations, days to weeks. For the first honest read on a campaign, four to six weeks. For AI search visibility and the sort of compounding that makes marketing feel cheap, think in quarters.

The businesses that get there are rarely the ones with the biggest budgets. They are the ones who kept going through the stretch where nothing appeared to be happening, because they understood that the pages, the tracking and the tedious two in the morning fixes were all quietly loading the spring.

At Infinity1 we score the site first, fix what is holding it back, and then give the work enough runway to compound, usually alongside a Billy Bot handling the enquiries that start arriving once visibility lifts. If you want to see whether you are early or genuinely stuck, let's talk. Bring your numbers and the date you started, and we will give you a straight answer on which one it is.

FAQ

How long does it take to improve an AEO score?

Days to weeks for most sites. Alpha Ski Tomamu moved from 50 to 90 in seven days. Larger sites take longer: ours went from 54 to 91 over three months.

How long before a new campaign produces sales?

Expect a data window of about two weeks, then a first fair judgement at four to six weeks. Harvest Bloom's site was rebuilt in late July, the campaign launched on 1 August, and the strongest month of the year followed.

When should I stop a campaign that is not working?

After six weeks, provided tracking is verified and the site scores above 75. Before that you are usually reacting to noise.

Why do results arrive all at once?

Because the work compounds quietly and becomes visible in a cluster. Pages, audience learning and platform data all mature at slightly different speeds, then line up.

Does consistency mean repeating something that is failing?

No. You still test, still fix things, still play what is in front of you. You keep the game plan while you change the details.