When someone asks us for an “SEO audit,” there are two ways to do it. One takes three weeks, produces a sixty-page PDF, and almost nobody reads it to the end. The other takes forty minutes and ends with a short list of what actually needs to change. This is the second one, step by step, with the stopwatch running.
It doesn’t replace in-depth work. It helps you decide where to start, which is where the most money gets lost.
Minutes 0 to 5: what Google already knows about the site
The first move isn’t opening the site, it’s opening Search Console. That’s where you find the only thing that isn’t opinion: which pages show up, for which searches, and how many people click through.
We look at three things, over 28 days:
- how many distinct pages have had at least one impression;
- how many clicks they add up to;
- whether those clicks are spread out or concentrated in two or three URLs.
To show this isn’t just theory, we ran it today on graphemics.net: over the last 28 days, 125 pages had impressions, adding up to 7,178 impressions and 53 clicks. And the page that brought in the most clicks, an image-prompt guide, took 33 of those 53, with 1,108 impressions at an average position of 7.2. In other words, more than half the traffic of a site with over five hundred articles depends on a single page. In five minutes, that already changes the conversation.
Minutes 5 to 12: the queries that almost work
Inside Search Console, we look for queries that sit between position 4 and 20. These are searches where Google already considers the site relevant but doesn’t place it where people click. It’s the lowest-hanging fruit there is: you don’t need to build authority, you need to improve a page that’s already in the race.
We also look for the opposite pattern, which is the most revealing: lots of impressions, a good position, and zero clicks. On our own site, the search “digital marketing” racked up 538 impressions at an average position of 4.5 without a single click in the period. A position like that with a result like that is almost never a content problem: it’s a title issue, a description issue, or the English search simply doesn’t match a page written in Spanish.
Minutes 12 to 20: the title and description of what already ranks
With the list from the previous step, we open each result the way a searcher would see it. We ask three questions:
- Does the title say what the person typed, or what we like to say?
- Does it get cut off in the search results? Past about 60 characters, Google tends to truncate it.
- Does the description give a reason to click, or does it just repeat the title in different words?
It’s the cheapest change and the fastest one. A title can be rewritten in five minutes, and its effect is usually visible in weeks, not months, although the exact timing depends on when Google next crawls that URL.
Minutes 20 to 27: the technical issues blocking everything else
Here we’re not looking for perfection, we’re looking for whatever keeps the previous steps from working. We check, in this order:
- robots.txt and sitemap: that they exist, that they respond, and that they aren’t blocking something important by mistake. It takes a minute and you can see it from the browser.
- Indexing: how many of the important pages are indexed and which ones show up as “crawled, currently not indexed.”
- Mobile speed of the page that brings in the most traffic, not the homepage out of habit.
- Language versions: that each translation points to its original and doesn’t compete with it.
Technical work stays at this stage on purpose. A site with a thousand minor errors and good titles usually performs better than one that’s flawless under the hood and silent on the outside.
Minutes 27 to 34: which pages exist and what they’re for
Now we walk through the site like a person would. We count how many pages are meant to sell something and how many are articles, and we check whether someone landing on an article has a clear path to the service.
The most common mistake we find is the buried service page: it exists, but no other page links to it and it doesn’t show up for any commercial search. Meanwhile, twenty informational articles get all the internal links. Fixing it is pure architecture, without writing a single new word.
In this part we also look for the opposite problem: two or more pages answering the same search and stepping on each other. The fix is deciding which one is the main page and linking the others to it.
Minutes 34 to 40: the list of five
The output of the audit isn’t a diagnosis, it’s a list of five actions ordered by speed and return. Each one comes with three details: which page, what change, and which number we expect to move.
A typical list looks like this:
- Rewrite the titles of the three pages with the most impressions and the fewest clicks.
- Move a page from position 8 to page one with content and internal links.
- Link the service page from the five most-visited articles.
- Merge two pages that compete for the same search.
- Fix the one technical block that does affect an important page.
Five actions, not fifty. If the site justifies a longer analysis, that list tells us where to dig.
The order matters more than the tools
You can do this review with paid tools or with just Search Console and a browser; the result barely changes. What does change the result is the order: start with Google’s data, move on to what almost works, and leave the technical work for last. People who start with a two-thousand-line technical crawl usually end up fixing things that never moved a single number, and reach the pages that actually mattered with no time or energy left.
What this audit doesn’t do
It’s worth being clear, because forty minutes has its limits. It doesn’t audit external links in depth, it doesn’t measure the competition keyword by keyword, and it doesn’t replace understanding the business: we can’t know whether a page “should” rank until we know what’s being sold and to whom. And it doesn’t promise results. Titles show their effect in weeks; a new service page usually takes months to settle in.
What it does do is help you avoid the costly mistake: spending three months writing new articles when the money was in four titles and one internal link.
How to do it with your site
If you have Search Console connected, you can do the first two steps yourself today in fifteen minutes. If you’d like us to do it, send us your site’s address and we’ll send back the list of five, with the numbers from your own Search Console right in front of you.
Search Console figures for graphemics.net, last 28 days, pulled on October 12, 2026. The timings per step and the time frames for results are approximate and vary by site.
Versión en español: Cómo audito un sitio en cuarenta minutos
