Real SEO case study: how we tripled traffic for a Catalan SME in 6 months

This is a real SEO case study with Google Search Console data, mistakes included and decisions that weren't always optimal. The client: an HR consulting firm in Eixample, Barcelona. Three years with the same website, good offline reputation and practically invisible on Google. Here I'll explain exactly what we did, in what order and why.

The starting point: three years without touching anything

When the client contacted us, their website was three years old, had a correct design and had never had any SEO work done. The schedule stayed full thanks to personal referrals, but whenever someone searched for "HR consulting Barcelona" on Google, they appeared on page 4. No organic leads in months.

I recognize this pattern immediately because I see it constantly: a physiotherapy clinic in Tarragona with 200 visits/month that depends 100% on Google Ads, a fashion store in Girona that doesn't even appear for its own brand, an e-commerce accessories store in Sabadell that sells well on Instagram but captures nothing organic. The underlying problem is always the same: technically fragile website, content without search intent and domain with no external authority. Three factors that reinforce each other.

Phase 1 — Technical audit with priority order

The first thing we do on any project is a technical SEO audit with Screaming Frog, Google Search Console and PageSpeed Insights. But what makes the difference isn't doing the audit, it's knowing how to prioritize: a report of 200 errors with no action order is wet paper.

The problems, ordered by real impact

1. Core Web Vitals on mobile — critical: LCP of 6.2 seconds (acceptable threshold is 2.5 s). The cause was a homepage slider with uncompressed images blocking render. We fixed it in 48 hours by removing the slider and applying lazy loading. LCP dropped to 2.1 s.
2. URLs blocked by error — critical: In Search Console, under "Indexing" → "Pages" → "Reason for non-indexing", we found 22 URLs with noindex tag added by accident during an old migration. Google was completely ignoring them. Fixed in one day.
3. Canonicals pointing to themselves incorrectly: 14 pages with misconfigured canonicals generating contradictory signals to the crawler. In Screaming Frog, filter "Canonicals" → "Non-indexable Canonical" detects it in thirty seconds.
4. Duplicate or empty titles: 38 pages without unique title. Filter "Page Titles" → "Duplicate" in Screaming Frog. Each page should have a title that describes exactly its content, not a variant of the company name.
5. Sitemap not declared in Search Console: The sitemap.xml file existed but had never been submitted. Small detail, direct impact on crawl speed for new pages.
6. Google Business Profile abandoned — highly undervalued: Without secondary categories, reviews unanswered for eight months, incorrect hours and zero posts in a year. For a local business, the GBP is the second most important digital presence you have — and the fastest to improve.

We fixed the first four points in the first two weeks. Without solid technical foundation, new content takes 30-50% longer to be indexed and ranked. There's no point publishing articles if Google can't read the website properly.

Phase 2 — Content that answers real questions

Once the website was working correctly, we designed the content architecture using thematic clusters: a pillar page for each main service and blog articles answering specific questions from potential clients. The logic is that Google ranks domains that cover a topic in depth better than those with isolated pages with no connection between them.

Why we didn't go for the highest volume keywords

The classic trap is attacking "human resources" or "labor consulting" — terms dominated by national portals and large firms with ten years of accumulated authority. The criteria we use is different: we look for SERPs where the top 5 is made up of websites from similar SMEs, not directories or Wikipedia. If the competition is manageable, the keyword is viable.

We published between 2 and 3 new articles per month. Each piece answered a real question extracted from two sources: Google's "People Also Ask" searches and questions the client regularly received by email or phone. This second source is pure gold and almost nobody uses it systematically.

A concrete example: the article "When is it mandatory to have an equality plan in the company?" had modest volume, but the intent was very clear and competition was low. Within three months it was in the Top 3 and brought qualified leads every week. It wasn't the most "impressive" content on the blog, but it answered exactly what someone was searching for just before needing consulting.

Phase 3 — Local link building without buying anything

This is where many projects get stuck. Either link building is completely ignored ("we'll rank with content"), or dubious link packages are bought that do more harm than good in the medium term. Neither option works sustainably.

Our approach: thematically relevant backlinks, from domains with real activity and verifiable traffic. In six months we achieved between 18 and 22 quality new links through three channels:

• Catalan sector directories with their own traffic: Not all directories are worth it. We prioritize those that appear in SERPs for sector searches, not those that exist only to sell links.
• Opinion articles in local business media: Two articles signed by the consulting firm's director in publications from Barcelona and Girona. Natural link, real authority and brand visibility as added benefit. The director was already writing articles internally — we simply channeled it outward.
• Content collaborations with complementary companies: Mutual mentions with a tax advisory firm in Sabadell and a labor law office in Terrassa. It's not link exchange in the penalizable sense: it's cross content with real value for each website's reader.

A single backlink from a sector reference medium is worth more than fifty from a generic directory. The quality of the context where the link appears is what Google values, not the raw number.

Results at 6 months

What surprised the client most wasn't the traffic volume, but the quality of contacts. Leads arriving via organic already came "educated": they had read two or three articles, understood the services and knew roughly what they needed. The sales cycle was noticeably shorter than with paid leads.

We see the same pattern in other sectors: a physiotherapy clinic in Tarragona we worked with went from depending exclusively on Google Ads to generating 60% of new patients through organic in eight months. A restaurant in Gràcia that, working only on GBP and three well-targeted local content pages, doubled online reservations in four months without spending a euro on advertising.

SEO isn't magic and it isn't fast. But when it works, the return is cumulative: each month builds on the previous one, something no paid campaign can do.

What we do differently and why it matters

Most SEO agencies work with the same process for everyone: automated audit, long report, generic content plan. The problem isn't that it's bad, it's that it doesn't account for the real business context or the specific competition in the Catalan market.

What we do differently:

• We prioritize by impact, not by report length. The first fifteen days always go to solving what blocks Google, not producing documentation.
• We choose keywords by looking at real SERPs, not just tools. A keyword with 200 searches/month dominated by SMEs can be more profitable than one with 2,000 dominated by Amazon or national portals.
• We integrate what the client already does well offline. Reputation, network of contacts, sector expertise — all of this can be converted into digital authority if you know how to channel it.
• We measure in leads, not positions. Being in the Top 3 for a keyword that doesn't convert is useless.

If you want us to review your website and explain what the three first concrete steps would be for your case, write to us for a free initial review. No templates, no guaranteed position promises.