Full website and marketing audit

A crawler can tell you a title tag is missing. It cannot tell you that your two case studies contradict the delivery promise on your homepage, that the phone number in your footer is not the one in your Google Business Profile, that the review widget your designer chose renders as an empty box to every model on earth, or that the campaign you have funded for two years is bidding on the one term you already rank first for.

Those are the findings that change revenue, and every one of them requires a person to look and notice. That looking is what the fee buys — not the hours around it.

Which is the honest answer to why this comes back in three days rather than three weeks. The collection layer is automated and always has been: the crawl, the field performance data, the rendering checks, the index picture, the prompt panel across five engines, the account exports. Machines are good at gathering and we let them do it. What is left is reading the output and deciding what it means, and that is the part you are paying a person for.

Most agencies quote four to six weeks for this scope. A meaningful share of that is queue time and the assembly of data a pipeline should be producing on its own. We would rather not bill you for our own inefficiency, and we would rather you had the plan while the quarter it applies to is still running.

Site audits conventionally sample. Twenty representative URLs, extrapolate, report an average. The logic is sound for measuring page speed and useless for finding defects, because defects cluster on the templates nobody looks at.

In practice the worst findings sit on pages a marketing team stopped opening two years ago: the thank-you page that asks for the information the visitor just submitted, the location pages generated in bulk with the wrong city left in the copy, the tag archive that outnumbers the articles it archives, the video-review pages published empty and still indexable.

None of that shows up in an average. All of it shows up when someone opens each template in turn. So that is what happens, and the report lists which templates were reviewed so you can check we did not skip one.

The visitor fails on trust and friction. They cannot tell whether the reviews are real, cannot find out whether you serve their address, cannot see the submit button on the form. Nothing is broken in a technical sense and the enquiry never arrives.

The crawler fails on access and duplication. Templates get submitted to the index that were never meant to be public, canonical tags point at the wrong version, the same paragraph lightly reworded competes with itself across nine pages, and half the sitemap claims nothing has changed since 2023.

The model fails earlier than either. It cannot retrieve what it cannot reach, and it will not quote a claim that only makes sense after two paragraphs of preamble. Passage-level structure decides whether you are in the answer, and it is invisible to every conventional SEO checklist.

One pass covers all three because the fixes interact. Making a claim self-contained enough for a model to lift usually makes it clearer for the visitor too, and a page that was never indexed is not going to be cited either.

Otterly found roughly 73% of sites were blocked from AI citation outright by crawler-access barriers — robots.txt rules excluding AI user agents, and content that only exists once JavaScript has run.

Sit with that number. Three quarters of the market has a visibility problem no content strategy will fix, because the content is never retrieved in the first place. It is also, by a wide margin, the cheapest thing on any audit's list to solve.

Blocking Google-Extended is the most expensive version of the mistake. It removes you from Gemini and from AI Overviews grounding while buying precisely nothing in classical ranking protection, and it is nearly always switched on by someone acting in good faith on a headline about AI scraping.

This gets checked in the first hour, before anything else, because it decides whether the rest of the findings can matter.

This is not a section we added to round out a brochure. The last two audits we delivered at this price both included the paid and social review, for Ontario renovation firms whose media spend was the largest single line in the marketing budget.

The recurring finding across paid, social and email is not a badly written campaign. It is that the channels are being measured by instruments nobody has checked in years.

Three specifics turn up again and again. Paid is bidding on terms the site already ranks first for, so a click that was free is now $6. The GA4 conversion event counts a page view on the thank-you page, which means it also counts the person who bookmarked it. And three channels each report the same lead as their own, because last-click, platform-attributed and CRM numbers were never reconciled, so the sum of your reported leads exceeds the leads you actually got.

Nobody is lying in any of those cases. The stack was assembled a piece at a time by different people, each of whom did something reasonable. It just adds up to a reporting layer that cannot be used to make a budget decision.

So the tracking gets audited before the channels do. There is no point ranking your campaigns by a number that is wrong.

One thing worth stating plainly, because it cuts both ways: the audit reviews and plans these channels, it does not run them. Running them is retainer work and quoted separately. We do run them — search, paid, social and lifecycle are all executed in-house on the monthly programme — which is exactly why the audit is deliberately sold on its own. An agency that only diagnoses what it also sells has an incentive problem, and the way we handle ours is to make the diagnosis a standalone product you can take to anyone.

The usual quote for this scope is four to six weeks. A fair question is what we are leaving out to do it in three days. The answer is the gathering, and only the gathering.

Every audit has two halves. One is collection: crawling the site, pulling field performance data, checking what renders in raw HTML, reading the index picture out of Search Console, running the prompt panel across five engines, exporting the ad and analytics accounts. That half is mechanical, it is identical on every engagement, and it runs on tooling we built because we were tired of doing it by hand. It finishes overnight.

The other half is judgment: reading all of that, noticing that the thank-you page is indexed, that two case studies contradict the homepage, that the campaign is bidding on a term you already own, and then deciding which three of the forty things matter. That half cannot be automated and is not being skipped. It is what the three days are.

The part of a long timeline nobody mentions is queue. Four weeks is rarely four weeks of work; it is a fortnight of waiting for a slot and then a week of assembling data a pipeline should have produced on its own. We would rather not bill for that, and a plan that arrives while the quarter it describes is still running is worth more than the same plan in November.

One condition on the number, because a guarantee should be keepable: the clock starts the day access lands, not the day you pay. If the ad-account invite takes a week, so does the audit.

A findings list is easy to produce and easy to ignore. Forty things are wrong, all of them are marked important, and the reader closes it having learned only that the situation is bad. The second half of this report exists so that does not happen.

It opens with channel strategy: which channels deserve budget, which should be cut, and what each is for. That section is written as an argument with the evidence attached, not a verdict, because you should be able to disagree with it on the facts.

Then the backlog. Every fix from the review, sequenced by impact against effort, with an effort estimate and a suggested owner — your developer, your agency, your in-house marketer, or us. Then a 90-day action plan drawn from the top of it, ordered so the work that unblocks other work happens first: crawler access before content, entity resolution before citation building, tracking before any budget decision.

Then a 12-month roadmap, quarter by quarter, and a measurement plan saying what to track per channel and what 'working' looks like — so that in a year there is a baseline to judge this against rather than a feeling.

This is the part clients use most. The constraint is almost never knowing what is wrong. It is knowing which three things to do on Monday.

The obvious way to sell this is in tiers: a cheap website audit, a dearer one that adds paid, a dearer one still that adds strategy. We drafted exactly that and threw it away.

The problem is that tiering asks the buyer to choose a scope before anything has told them where the problem is. Someone whose real issue is $4,000 a month of wasted ad spend buys the website tier, because the website is the thing they were worried about, and the audit is then contractually forbidden from mentioning the actual problem.

It also makes the strategy section an upsell, which quietly reframes the base product as a findings list. A findings list with the plan removed is not a cheaper version of this. It is a different, worse thing.

So there is one price and one scope. If your business genuinely has no paid, social or email activity, say so at the intake call — those sections come back short, the review goes deeper on the site, and the price does not change.

The audit ends at the plan. It does not include the implementation, and that is a deliberate structural choice rather than a pricing one.

An agency that has already sold you the remediation has a reason to find remediation-shaped problems. Every finding starts, gently and unconsciously, to point at work it happens to sell. Selling the diagnosis on its own removes that incentive: if the honest answer is that your site is technically sound and the real problem is that nobody can tell what you charge, the report says that, and it costs us the follow-on engagement we might otherwise have written into the recommendations.

Plenty of clients take the document to their existing developer. Some take it to a different agency. That is a fair outcome for something you paid for, and the report is written to be legible to someone who has never spoken to us.

If you do want us to execute, the monthly retainer covers all of it — search, AI visibility, paid, social and lifecycle run by one team — and the audit fee is not charged twice. It is credited against the first month of a retainer started within 60 days.

The free audit on this site is real and it stays free. It is a narrow diagnostic, largely automated: a prompt panel showing how the major engines currently answer questions in your category, plus an entity check across your schema, Google Business Profile and directory records. It is one conversation and it is yours whether or not anything follows.

What it is not is a review of your business. It does not open your templates, read your copy, look at your ad accounts, touch your analytics, or produce anything anyone could implement from. It reports a status. It is designed to show you whether there is a problem worth paying to understand.

The paid audit is that understanding, plus the plan: every channel you run reviewed and argued, and a sequenced answer to what to do next.

Start with the free one if you are not sure. A fair number of people do, and some of them conclude they do not need the paid version. That is a reasonable place for it to end.

Read-only access to Search Console, your analytics, the ad accounts and the email platform, plus the live site. No admin rights, and no credentials over email.

Then a single hour with whoever actually knows the business: what you do, who for, where, since when, which claims on the site you can stand behind, and what you have already tried. That hour changes the report more than any tool does, because half of what looks like a content defect turns out to be a true thing stated badly, and half of what looks like a strategy gap turns out to be something you tested in 2024 and abandoned for a good reason.

If a claim on your site cannot be verified, the audit says so rather than repairing the wording around it. That has been an uncomfortable conversation more than once, and it is still the right output.