This site has 506 articles and 35 published pages. In the last 90 days, 120 URLs showed up at least once in a Google result and 14 got at least one click. In total: 14.911 impressions and 53 clicks.
The usual industry reaction to those numbers is to publish more. We’re doing the opposite: thirty days without publishing anything new, just fixing what already exists. And we’ll share the before-and-after here, like we did with the title experiment.
The inventory, plain and simple
- 541 URLs published across articles and pages.
- 120 with some impressions in 90 days. The other 421 don’t exist to Google.
- 14 with at least one click.
- The homepage alone: 9.215 impressions, average position 4,3 and 4 clicks.
- One article —a list of prompts to generate images— pulls 595 impressions and 26 clicks, half the site’s total clicks.
Those last two points are the full diagnosis. There’s a page that shows up a lot and doesn’t get clicked, and an article that appears less and gets clicked a lot. The issue isn’t volume; it’s alignment between what people search and what we offer in the result.
The four URLs competing with each other
Under the homepage slug lives a family of service pages. In 90 days they add up to 1.403, 860, 533 y 349 impresiones, with average positions of 30,1, 30,6, 40,4 and 70,0. And zero clicks across all four.
These are real service pages we actually sell, showing up for commercial queries, and they’re on pages three and four. None of them needs a new article to keep it company; they need someone to organize them, choose the best one for each topic, and consolidate into it what’s currently spread across four.
What we’re doing in thirty days, in order
- Week 1 — The homepage. It has 9.215 impressions at position 4,3 and four clicks. Anything that lifts that CTR moves the whole site. Start with the snippet: title, description and —we learned this the hard way— the first paragraph of the body, because when Google rewrites the summary, it pulls it from there.
- Week 2 — The four cannibalizing pages. Pick one per topic, merge the useful content from the others, redirect the ones we retire, and fix the internal links that pointed to the old ones.
- Week 3 — The 14 that actually bring traffic. Expand, refresh, and interlink the URLs that already have clicks. It’s the fastest-return work and the one almost no one does, because it doesn’t feel productive.
- Week 4 — The 421 invisible ones. Decide piece by piece: combine those on the same topic, update the ones still relevant, and retire with redirects the ones that aren’t. A site with 150 live pages outperforms one with 541 pages where 421 are filler.
Why this works better than publishing
We’ve already tested it in-house. Two weeks ago we changed only the title and the description on 15 pages, without touching a word of the content. Impressions for that group went from 69 to 183 in fourteen days. A page that had one impression hit a hundred. Two that didn’t show up started showing up.
That lift came from an afternoon’s work. No new piece on this site has done anything close in its first month.
The reason is boring and it’s the right one: a page that’s already indexed starts with history. Google already knows it exists, has shown it, and has signals about it. Changing its title moves a piece that’s already on the board. Publishing a new one is starting from zero and waiting months.
How to know if your site needs the same
Three Search Console numbers, twenty minutes:
- How many URLs you have published versus how many had impressions in 90 days. If the second figure doesn’t reach 30% of the first, you have an inventory problem, not a production problem.
- How many have at least one click. That’s the part of the site that works; the rest is weight.
- Where your impressions without clicks are. If they’re between positions 1 and 10, the problem is the result. If they’re 20 and beyond, the problem is the page.
Both problems can be fixed without writing a new article, and the second almost always gets fixed by consolidating what you already wrote.
We’ll publish the 30-day check-in with these same numbers side by side, whatever the outcome. If you want yours sooner, write to us: we’ll tell you how many of your URLs are alive and which ones we’d tackle first.
Inventory tallied by the graphemics.net content API and Google Search Console data from July 2 to September 29, 2026, downloaded on September 30. Search Console figures stabilize over a few days and may vary slightly in later revisions.
Versión en español: Treinta días sin publicar nada nuevo, solo arreglando lo viejo
Versión en español: Treinta días sin publicar nada nuevo, solo arreglando lo viejo
