People ask for my SEO audit template expecting a spreadsheet with two hundred rows and a tab for every problem type. What I actually use is four pages in a document, and two of those pages are written for the client rather than for me. The giant spreadsheet looks thorough and gets read once, because nobody can turn it into a decision.
The template exists to answer one question: is this site a fix or a rebuild? Every row works toward that answer. This is the structure I fill in, and it starts with the free website auditor scan rather than with a blank page.
Start with the scan, not the template
Run the scan first and let the template record what it found. The scan returns status, HTTPS, redirects, canonical, indexability, title, description, headings, links, images, schema and response time for one page. I copy each result into page one and mark it pass, warn or fail. A fail is something that can stop a page being indexed or found. A warn is something that costs clicks or trust. A pass goes into the file so the next audit can show that it stayed fixed.
The problems that cost the most are usually the ones you cannot see by looking at the page. A noindex tag left over from staging, a canonical pointing at an old domain, a robots rule blocking the stylesheet that renders the layout. The scan shows those because it reads the head and the headers, not because it is clever.
Page one: the signal sheet
Four columns: page, what I checked, the finding, the evidence. The evidence column is the one people skip and the one that matters. "Canonical is wrong" is an opinion. "Canonical points at http://staging.example.com instead of the live URL" is a finding a developer can act on without asking a question.
Keep the evidence small. The line from the header, the tag as it appears in the source, the status code next to the final URL. A screenshot of the whole page proves nothing and makes the file heavy. When the audit is later questioned, the evidence line ends the argument in one read.
Page two: the demand check
This page is empty if the client cannot share Search Console. Write that down and say the audit covers the page, not the demand. If the data is available, the columns are page, impressions, clicks, average position and the query the page seems to serve. The point is not to rank pages by traffic. The point is to see whether the page that fails the scan is also a page that gets served, because that changes the priority.
A page with no impressions and a broken canonical is a candidate for deletion, not repair. A page with real impressions and a thin answer is the cheapest win on the site. Same finding, different order.
Page three: the fix list
One row per fix, with five columns: what, where, why it matters, effort, owner. Order the rows by what unblocks indexing first, then what wins clicks, then everything else. The order is what turns the audit into a quote. Without it, the client reads twelve problems, feels overwhelmed and buys nothing.
Write the effort honestly. "Small" for a title change, "medium" for a redirect map, "needs the host" for a server header. The owner column is usually the client, not me, and saying so up front prevents a month of polite emails.
Page four: the client note
One page, in the client's language, with three parts: what is fine, what is costing money, and what I would do first. No jargon, no scores out of a hundred, no "opportunities" list that pads the report. If the site is basically healthy, say that and give the two small fixes. An audit that invents work to justify its fee is the reason people distrust audits.
The scoring rule
I do not score sites out of 100. A score implies precision that an audit does not have, and it starts an argument about the number instead of the fix. Instead I count fails: none means the page is healthy, one to three means a short fix list, four or more means the plumbing comes before any content work. That is a rule of thumb, not a measurement, and I say so when I hand it over.
What the template leaves out
Backlinks, rank tracking, competitor traffic and any number that a third party tool reports. Those tools are useful for a decision about strategy, not for an audit of a page. A template that mixes the two produces a document nobody can finish. Keep the audit to what the site controls.
The blank template
- Page and URL tested. One line, with the date.
- Scan result. Pass, warn or fail for each of the checks you ran.
- Evidence. The tag, header or line that proves the finding.
- Demand. Impressions, clicks and the query, if the client shares the data.
- Fix list. What, where, why, effort and owner, in priority order.
- Client note. What is fine, what is costing, what to do first.
- Decision. Fix or rebuild, with the reason in one sentence.
The template takes about an hour per page to fill in properly and about twenty minutes once the site is clean. If you want the scan that fills page one, the website auditor is free and takes about a minute. The 40-check walkthrough explains the order of the checks, and the Fix Sprint is the version where I fill in all four pages for you. The tools hub has the other checkers the template calls for.