Case: how I decided what to sell with data, not intuition

What needed solving

I was going to advertise as an SEO auditor. It made sense: an audit ends in recommendations, and every recommendation is a service. But I hadn't measured anything.

The risk was the usual one: picking the biggest search without looking at who types it or who already wins it, and finding out six months later with a whole site aimed at the wrong place.

What was done

I measured Google Ads volumes in Colombia, Spain and Mexico, checked who occupied the top 10 for each candidate and calculated difficulty from that real top 10. Then I grouped by intent, not by how similar the words looked.

I try to cross-check every volume with a second tool and declare the difference. In this research I could only do it for «auditoria seo»: 70 searches per month in Google Ads and 60 in Mangools. The second source got blocked by its usage limit, and I wrote that down instead of presenting the other figures as cross-checked.

On September 17 I repeated the exercise in English for the future version of the site. There, calculated difficulty changed a decision already made: «semantic seo» looked like the ideal cluster by volume and fell apart when I measured who occupies its top 10.

How it turned out later

This site's home page positions as a technical SEO consultant, the audit lives on its own page as a deliverable, and the AI assistants topic has its own service. That was the decision. What follows is what Google says four days after indexing the site.

Ten impressions are not a trend and I won't dress them up as one. The home page shows in position 1 only for searches of my name; the blog pages are starting to surface in positions 77 to 95 for «core web vitals», which is exactly what you'd expect from a four-day-old site.

What I learned, mistakes included

Until September 15 this page said «I cross-check every volume with a second tool». It was an exaggeration: I had only done it for one keyword. I corrected it and left the correction in writing.

Commercial tools gave me difficulty 0 for «consultor seo» and 0 for «single page application seo». Both had a top 10 with real authority. Since then difficulty gets calculated, not looked up.

And I learned that volume doesn't decide alone. «semantic seo» had 880 searches and my LinkedIn headline pointed there; the top 10 said no, and the top 10 rules.

If you want to check it yourself

Search «consultor seo» on Google from Colombia and look at who occupies the first ten results: personal sites, not agencies. Then search «agencia seo bogota». The difference in intent is what decided this site.

• The tool I used to calculate difficulty is Cayu.

• The first pillar in English, SPA SEO, rests on the React app case.

• How this site was put into production after deciding: the domain and Worker case.

What I can't claim

Google Ads volumes come in rounded bands: «10» means «between 0 and about 15». And except for «auditoria seo», every figure comes from a single instrument.

I don't claim positions for any commercial keyword: with ten impressions, any average position is noise.

Next measurement

The dates are written before the data exists. If I haven't updated this page by then, call me on it.