Why We Won't Run Your Ads Until Your AEO Score Hits 75

Quick Takeaways

1. The policy is simple and it has not changed: we won't run your ads without doing an audit and getting the score above 75.
2. AEO means Answer Engine Optimisation, which is whether AI assistants and answer engines can read your business, understand it, and cite it when someone asks.
3. The score measures the destination your ad budget points at, and a weak destination makes a healthy return impossible no matter how good the campaign is.
4. Harvest Bloom went from 57 to 98 before a single ad ran, and August alone produced 84% of their year's revenue.
5. Warner Pool Care went 42 to 96 and runs at 14:1 on thirty dollars a day. Alpha Ski Tomamu went 50 to 90 in seven days.
6. We target 8:1 to 15:1 on ad spend, with 8:1 as the floor, and the floor is where the arithmetic stops working if the site cannot convert.
7. There are real situations where running ads first is the right call, and a business owner who wants leads this month has a fair complaint about waiting.

The rule, and why people bristle at it

Somebody rings up wanting ads. They have a budget ready, they know roughly what they want to say, and they would like it live this week. Then I tell them the rule: we won't run your ads without doing an audit and getting the score above 75.

The response is usually polite and always the same underneath. That sounds like you're finding extra work to sell me.

Fair enough. I would think that too. Plenty of agencies have made a good living bolting a discovery phase onto a job that did not need one, and if that has happened to you, a founder asking you to wait three weeks sounds exactly like the thing that burned you. So here is the real reasoning, including the part where the objection is right.

What an AEO score measures, in plain English

AEO stands for Answer Engine Optimisation. It covers whether AI assistants and answer engines, meaning ChatGPT, Gemini, Perplexity and Google's AI Overviews, can read your website, work out what your business does and recommend you when somebody asks a question in your category.

That sounds like a search problem. It is also a conversion problem, which is the part people miss, because the things we score are the same things a human buyer needs. Does the site say clearly who you are, where you operate and what you sell, in language a machine and a distracted person can both parse in one pass. Does a page answer the question it was landed on inside the first paragraph. Is your business name consistent everywhere it appears. Is there structured data describing your products, prices, reviews and service areas. Are the AI crawlers even allowed in.

We score that out of 100 across ten areas: crawlability, schema, content structure, experience and expertise signals, freshness, entity clarity, reviews, local signals, AI crawler access and question-shaped content. A site in the forties or fifties is usually a good-looking site that has never been asked to explain itself. It photographs well and answers nothing.

Now put paid traffic on top. Someone clicks the ad, arrives, scans for four seconds looking for the one thing they came for, cannot find it, and leaves. You paid for that, and you will pay again twelve seconds later when the next one does the same.

The kicking tee

I grew up on rugby league, so this is how it sits in my head.

You would not hand the kicking tee to a middle forward when you have a fullback landing 85%. The forward might get one over from in front. Wide out on the touchline in the wet, with the game on it, you want the bloke who has kicked ten thousand of them.

Your ad budget is the shot at goal and your website is whoever takes the kick. Most businesses put all their energy into the shot, the angle, the spend, the creative, and almost none into who is actually kicking. Stack the odds first, then kick.

The uncomfortable version of that: a great campaign sending traffic to a site scoring 45 will lose money faster than a mediocre campaign sending traffic to a site scoring 90. Traffic is an amplifier. Amplify a page that does not convert and you have bought a bigger version of the same problem.

What the evidence looks like

Harvest Bloom Flowers is the clearest one I have. A flower farm in New Zealand with a real product and a real grower behind it, and a site that gave away almost none of that. The score went from 57 to 98. We rebuilt the site over the last two weeks of July and the campaign went live on 1 August.

January through July, revenue was NZ$723. August on its own did NZ$3,828.97, which is 84% of the year inside one month. Nothing clever happened on 1 August. The work had already been done and the ads finally had somewhere worth sending people.

Warner Pool Care in Brendale went from 42 to 96 and runs at 14:1 on thirty dollars a day through Google Shopping, a modest budget performing because the destination does its job before the budget is asked to. Alpha Ski Tomamu moved from 50 to 90 in seven days, the fastest turnaround we have done. Our own site went from 54 to 91 over three months, which tells you how the work goes when you are your own client and everything else comes first.

The arithmetic of 8:1

We target a return on ad spend between 8:1 and 15:1, with 8:1 as the floor. Eight is not a magic number. Below it the exercise stops being worth the risk for a small business, because once you take out cost of goods, your own time, platform fees and the inevitable bad month, a 4:1 campaign is a great deal of activity in exchange for a wage you could have earned doing the work yourself.

Follow where the ratio comes from. Spend one thousand dollars and at the floor you need eight thousand back. Every dollar of that has to survive the same journey: the click, the landing, the moment of understanding, the trust, the action. The campaign controls the first step and your website controls every one after it.

With a weak destination you are chasing 8:1 through a leaking conversion rate, and the only lever left is volume. More spend, more clicks, same leak. That is the point where owners conclude ads do not work in their industry.

Harvest Bloom converted at 2.44% in August across 98 orders at a NZ$39 average order value. Ordinary ecommerce numbers, and they were enough, because the site was made ready before the traffic arrived. The campaign only ever had to be competent and pointed somewhere sound.

The objection is reasonable, and here is my honest answer

Someone with payroll on Friday and a quiet pipeline is being told to wait three weeks. That is a real cost to a real business and I will not wave it away, so be specific about both sides.

What the wait costs. Three weeks of leads you will not get. In a seasonal business it can be worse, because three weeks in the wrong part of the calendar is most of the season. There is also the cash flow shape of it: you pay for the audit and fix up front, and the return arrives later.

What the wait buys. Every dollar afterwards works harder, for as long as the site keeps working. The fix does not expire with the campaign. It keeps converting organic traffic, keeps you readable to the AI assistants people increasingly ask first, and lifts every channel you ever run. Advertising before it is done means paying full price for traffic that leaks, then paying again for the same traffic later.

When running ads first is genuinely right. There are four situations where I will say go.

1. A dated deadline you cannot move. A ski season, a Christmas trading window, an event with a fixed date. If the calendar closes before the work lands, run now and fix in parallel, with your eyes open about the numbers being worse than they should be.
2. An existing campaign already converting. If something is running profitably today, switching it off to do foundation work is silly. Keep it alive, do the work around it, then watch what the same campaign does afterwards.
3. An unproven offer. Sometimes the honest question is whether anybody wants the thing at all. A small, deliberately cheap test to find that out is a sensible use of a few hundred dollars before you invest in a site built around an offer nobody has validated.
4. The site is already close. If the audit comes back at 72, we are talking about days of work, and that conversation is different from one that starts at 45.

Outside those four, waiting wins on the arithmetic almost every time, and it is shorter than people expect. Alpha Ski Tomamu moved 40 points in a week.

How to decide, in order

If you want to run this yourself before speaking to anybody, do it in this sequence.

1. Get a score. Any credible AEO or technical audit will do, as long as it is measured the same way twice so you can see movement. Get it before you form an opinion.
2. Read your top landing page as a stranger. Four seconds. Can you tell what is sold, who it is for, where it is available and what to do next. If you have to scroll to answer any of those, an ad click will not survive it either.
3. Check the machines can get in. Confirm AI crawlers are allowed, your structured data is present and valid, and your business name is identical across your site, your Google Business Profile and your socials.
4. Fix the highest-impact gaps first. Question-shaped headings, first-paragraph answers, FAQ and review schema, entity consistency, page speed. That is the work that moves a score fastest.
5. Re-score. Above 75 you are ready for traffic. Below it, the next dollar goes towards subsidising a problem you already know about.
6. Turn the ads on and measure them properly. Set the 8:1 floor as the standard before you start, so you are judging against a number you chose in advance.

What we do with the three weeks

The audit scores the site across those ten areas and comes back with a prioritised list. We fix the highest-impact gaps ourselves: schema, brand consistency, crawler access, page structure, first-paragraph answers, FAQ and review markup. We register the business with the answer engines and tidy the directory footprint, then re-run the score so you can see the movement. Only then does anybody at Infinity1 talk about ad budget. For clients who carry on with us, a Billy Bot sits on the site handling the questions that arrive at ten at night, which is a conversion layer in its own right and an honest record of what buyers keep asking.

The short version

The ad budget is the shot at goal and the website is who takes the kick. Get the score above 75 and a modest budget does the work of a big one, which is how thirty dollars a day at Warner Pool Care becomes 14:1 and how a rebuilt flower farm site turned August into 84% of the year. If you have a deadline you cannot move or a campaign already working, run the ads and we will fix underneath you. Otherwise the three weeks pays for itself, repeatedly.

If you want to know where your site sits before committing another dollar to advertising, get the audit done and let's talk about what the number says.

FAQ

What is a good AEO score? Above 75 is our threshold for advertising. The clients in this post started at 57, 42, 50 and 54, which is about typical for a site that has had normal SEO done to it and nothing more. Getting into the nineties is where a site starts turning up in AI answers for its category with any consistency.

How long does it take to get above 75? It depends on where you start and how the site is built. Alpha Ski Tomamu went 50 to 90 in seven days. A typical fix week moves a site into the seventies.

Does this apply to Google Ads as well as Meta? Yes, and arguably more so on Google, where the intent is already high and the visitor has come to make a decision. Warner Pool Care's 14:1 runs on Google Shopping.

What if I run ads first and the numbers are bad? That is the usual outcome, and the expensive part is what people conclude from it. They decide the channel does not work for their industry, when the fair conclusion is that the destination was never ready for the traffic.