An audit is a diagnosis, not a treatment
An SEO audit tells you what's wrong and in what order to fix it. It does not fix anything. That distinction sounds obvious right up to the moment you're four weeks in with a 60-page PDF, no engineering time booked, and the same traffic you started with.
So the useful question isn't "should I get an audit?" It's "what decision will this audit let me make that I can't make today?" If you can name the decision, buy the audit. If you can't, you're buying reassurance, and reassurance is cheaper elsewhere.
What goes inside a good one — the four sections, the prioritised fix list, the price bands — is covered in what an SEO audit is. This page is only about timing.
Trigger one: you're replatforming or migrating
This is the trigger where an audit most reliably pays for itself, because migrations are the single most common way a healthy site loses half its organic traffic in a fortnight. WordPress to Shopify, a domain change after a rebrand, a subfolder restructure, a headless rebuild — the risk is identical and it's mostly clerical.
Before you move
- A full crawl and URL inventory of the current site, with traffic and links attached to each URL, so you know which pages actually matter.
- A one-to-one redirect map. Not a bulk redirect to the homepage, which Google will largely treat as a soft 404.
- A template inventory: which page types carry the internal links, the schema and the copy that earned the rankings.
- A frozen baseline of rankings, impressions and organic leads, dated, so you can prove afterwards whether the move cost you anything.
Two to four weeks after you move
- Verify every redirect resolves in one hop, with no chains and no loops.
- Check indexation of the new URLs in Search Console against the old set, page type by page type.
- Confirm canonicals, hreflang and schema survived the rebuild — they routinely don't.
- Compare internal link counts per template. Losing a footer or a related-posts block quietly strips thousands of internal links.
Trigger two: organic traffic dropped and nobody can say why
A drop is a legitimate trigger — but do the free triage first, in this order. Most drops resolve inside half an hour and don't need anyone's invoice.
- Check the tracking before the site. A broken GA4 tag or a consent-banner change looks exactly like a traffic collapse. Cross-check against Search Console clicks, which are measured on Google's side.
- Look at the exact date the line breaks. A cliff on one day is technical or a manual action. A slope over two weeks is usually an algorithm update or seasonal demand.
- Check Search Console for manual actions and indexing. A
noindexshipped by mistake, a robots.txt block, or a server that started returning 5xx to Googlebot are all findable in minutes. - Segment by page type and query. A drop concentrated in one template is technical. A drop spread evenly across the site is usually a core update or lost demand.
- Check whether demand fell. If the query itself is being searched less, or now resolves without a click, your ranking may be fine and your traffic still gone.
Trigger three: you're hiring an agency, or firing one
An independent audit before you sign a retainer does two things. It gives you a scope that isn't written by the party being paid to deliver it, and it gives you a dated baseline — which is the only thing that makes any future claim of improvement checkable.
On the way out it does something different: it tells you what you actually own. Whether the content is thin or solid, whether the links are earned or bought, whether there's a penalty risk sitting in the backlink profile that becomes your problem the day the agency leaves.
One rule, and it isn't negotiable: don't buy the audit from the agency pitching you the retainer. A free audit is a sales document, and it will find precisely the problems that agency sells the fix for. Pay a third party a few tens of thousands of rupees, or pay an agency you've already ruled out for the retainer.
Trigger four: you stopped getting cited in AI answers
This one is newer and needs stating carefully. There's no public API for AI Overview or assistant citations, no rank tracker with authoritative coverage, and the selection mechanics aren't documented by anyone who actually knows them. Any consultant quoting you a precise "AI visibility score" is quoting a proprietary sample, not a fact.
What is checkable is whether you've made yourself citable, and that's a real audit scope:
- Crawler access. Whether your robots.txt and CDN rules allow the AI crawlers you want to be read by, and block the ones you don't. Plenty of sites blocked them by accident during a bot-mitigation config.
- Extractability. Whether your pages answer the question in a self-contained paragraph near the top, or bury it under 400 words of preamble.
- Entity footprint. Whether your brand, people and products are described consistently across your own site and the places that describe you elsewhere.
- Query reality. Whether your money queries now resolve with an AI answer at all, and whether they used to send clicks. Sometimes the traffic didn't move to a competitor — it stopped existing.
Three times an audit is just procrastination
We turn down audit requests for all three of these, because taking the money would be the easier and worse decision.
- You already have last quarter's fix list and haven't done it. A second opinion on a list you're ignoring costs money and buys three more weeks of not starting. Ship the first list.
- The site is four months old with 12 pages. There's nothing to diagnose. You need pages people search for, and the answer is already obvious from a keyword list.
- Nobody has the engineering time to act on it. An audit with no dev capacity behind it is a document. Book the sprint first; buy the audit to fill it.
- Bonus, and the most common of all: it's an annual habit. Nothing structural changed and the calendar said July. Audit against events, not against the financial year.