Cases

Real results, and how we got there.

Anonymous cases: we don't publish client names without their written permission. What we do show is what we found, what we did and what changed, separating what our work drove from what the business would have grown anyway.

organic revenue in one month
+31.6%
technical errors after a relaunch
-93%
in its category in AI answers
#1

Index

Six cases, from diagnosis to result.

Each case separates what the work drove from what grew because of the business, and also includes what didn't improve.

01Prepared-meal delivery ecommerce

One month, eight hours and +31.6% in organic revenue.

A store with a great product and stalled traffic. The problem wasn't the content: it was technical errors nobody had measured.

  • SEO
  • CRO
Organic revenue
+31.6%
Organic orders
+20.9%
Clicks from Google
+41%
Of work that month
8h

Measured against the baseline, before we touched the site. Organic CTR went from 6.4% to 9.7%.

What we found

  • Template 302 redirect on 239 pages
  • The best-selling product returned a 404
  • The checkout was indexed on Google
  • 87% of organic traffic depended on brand searches

What we did

  • Fixed the template redirect on all 239 pages
  • Restored the best-selling product page
  • Removed the checkout from the index
  • Set a baseline to measure month over month

02Nature boutique hotel

A new site without losing what already ranked.

The hotel was launching a new brand and needed a new site. The risk was losing rankings that took years to earn. We rebuilt the site keeping its URLs and measured before and after against a frozen baseline.

  • SEO
  • UX
  • UI
  • AEO/GEO
Technical errors (from 121 to 8)
-93%
Organic traffic to room pages
+15.7%
Mobile abandonment on the first screen (was 69%)
13%
Clicks on the booking button from the home page
+16%

Compared against the previous site's baseline. To be transparent: LCP got worse in the first weeks and became a technical priority.

What we found

  • Accumulated errors: broken links, duplicate titles and H1s
  • On mobile, most visitors left each room page on the first screen
  • Adding a new tour required a developer
  • The new brand wasn't reflected on the website

What we did

  • New WordPress site that keeps the ranking URLs, with only 3 redirects
  • Room page redesign guided by recordings and heatmaps
  • Bilingual templates so the hotel team can publish experiences without code
  • AI visibility tracking: 31% share of voice, ahead of the competitors measured

03Car rental · Tourism

The number one brand in its category in AI answers.

A direct competitor was already investing in ChatGPT visibility, and the company didn't know how it showed up there. We fixed the technical foundation and structured data, and turned AI into a measurable channel.

  • AEO/GEO
  • SEO
  • Local SEO
In its category in AI answers
#1
Average Google position (was 15.5)
9.4
Conversions in one month from AI assistants
71
Main keywords at #1 in its core market
9/11

Same month versus the previous year. For the full picture: organic clicks fell 14.3%. We showed with data that the click was being lost, not the ranking.

What we found

  • Most locations had no local business structured data
  • No fleet page had schema
  • Brand clicks were falling and no one knew whether AI was the cause
  • A competitor was already investing in AI visibility

What we did

  • Schema for every location and the fleet, in every site language, validated
  • Pages rewritten as citable questions and answers
  • AI visibility measured with two tools in parallel
  • AI assistant channel tracked in GA4

04Vehicle leasing · B2B

The only channel that keeps growing is organic.

Traffic depended on paid ads and organic wasn't taking off for key searches. Many pages were competing with each other. We reorganized the architecture and built model landing pages designed for quote requests.

  • SEO
  • Content
  • Analytics
Organic conversions
+17.9%
Organic conversion rate, the best of all channels
6.7%
Average position (was 17.8)
12.2
For the industry's three main searches
#1

Same month versus the previous year, in GA4 and Search Console.

What we found

  • 29 of its own URLs competing for the same search
  • One anchor repeated on 687 pages
  • A mirror domain ranking above the home page

What we did

  • Cannibalization plan with 301s, noindex and retargeting
  • Model landing pages with spec sheets
  • Links in industry business publications
  • Reporting that separates the effect of paid ads from the effect of SEO

05Industrial manufacturer · B2B

Pages that grow even while the site declines.

A manufacturer's service pages had large gaps against competitors and barely appeared in AI answers. We rewrote four priority pages with FAQs and structured data, and measured with equal time windows.

  • SEO
  • AEO/GEO
  • Content
Clicks on the optimized pages
+40%
Impressions while the whole site declined
+30%
Citations in AI Overviews
2x
Positions on a key page
+9

Same number of days before and after publishing, in Search Console. The optimized pages grew while the site as a whole declined.

What we found

  • A duplicate, hidden H1 across the site
  • Large keyword gaps against competitors

What we did

  • Page-by-page action plan aligned with the paid campaigns
  • FAQs with schema on four priority pages
  • Fixed resources loading without HTTPS

06The same ecommerce · Operations

From counting by hand to a plugin that matches exactly.

The week's production was counted by hand, order by order. We built a WordPress plugin that groups orders by delivery date.

  • WordPress development
  • Automation
Match with the official export
100%
Vulnerabilities in the review
0

Security review before going to production.

What we found

  • Manual production count every week
  • Errors when matching delivery dates
  • Hours of operations work repeated every week

What we did

  • Custom WooCommerce plugin that groups by delivery date
  • Validation against the official order export
  • Security review before production

Next step

Is your business the next case?

We start with a baseline so every result can be measured.