Pedro Aureliano
SEO especialist
New to Contra
Pedro is ready for their next project!
8x the commercial pipeline in 90 days, across four continents.
THE SITUATION
Sax Trading LLC exports food commodities, operating out of Brazil
and Florida. The product was sound and the market was there. The
company had no digital infrastructure of any kind: no site that
converted, no funnel, no CRM, no attribution, no structured
commercial process.
Deals happened through relationships and inbound luck. Nothing was
repeatable, and nothing was measurable.
This is the hardest kind of engagement: there is no baseline to
optimize. Everything has to be built, and it has to be built in the
right order.
WHAT I BUILT
Digital foundation
Institutional site and conversion-focused landing pages, structured
for an international B2B buyer rather than a consumer. Multilingual,
built around the actual purchasing logic of commodity importers.
Multi-market paid media
Simultaneous campaigns across four continents, running on Google Ads
and social platforms. Separate structures per market, because an
importer in Europe does not search the way one in Asia does.
Segmentation, creative and bidding treated as market-specific, not
translated.
Sales engine in GoHighLevel
The piece that made the rest work. Every lead from every market
entering a single pipeline, with source attribution preserved from
first click. Automated follow-up sequences, stage-based tasks,
commercial scripts and a view where the sales team sees context
instead of starting from zero on every conversation.
Content and authority
Social presence structured to reinforce credibility in a category
where trust decides the deal. Strategy driven by positioning, not
by posting calendar.
THE RESULT, AFTER 90 DAYS
Commercial pipeline went from R$800,000 to R$6,600,000 in a single
month, measured inside the CRM that had just been built.
That is 8x, and roughly US$160K to US$1.3M at the exchange rates of
the period.
WHY IT WORKED
The order. Most agencies would have started with ads, because ads
show activity fastest. Running paid traffic into a company with no
funnel and no CRM would have produced clicks and lost every one of
them.
The site and the pipeline came first. The campaigns came second.
By the time budget started moving, every lead had somewhere to land,
someone assigned to it, and a trail back to the campaign that
produced it.
WHAT THIS CASE DEMONSTRATES
Building commercial infrastructure for a company operating in
multiple markets at once, with attribution intact across all of
them. Not a channel. The whole system.
0
18
Reversing an 82% organic collapse
THE SITUATION
A B2B technical e-commerce on Magento lost roughly 82% of its
organic traffic after a migration. Nobody could say exactly why.
WHAT I FOUND
21+ critical URLs sitting in 404 with no redirects. Missing meta
tags across key templates. A generic homepage with no dynamic
product surfacing. Mobile performance below 70.
WHAT I DID
21 permanent redirects mapped and deployed
Complete JSON-LD implementation
New homepage with dynamic Magento product widgets
A service page rebuilt for conversion
Performance work taking mobile to 81
All existing automation preserved and untouched
THE RESULT
Active users +69.7%
New users +74.3%
Key product page +748%
Recovered variant +2,159%
PageSpeed, all scores 95 to 100
ChatGPT citations 86 in one month, up 38.7%
Client's reaction, verbatim: "Ficou lindo demais."
1
23
From paid-dependent to organic-dominant in seven months
THE SITUATION
A B2B headhunting firm was running Google Ads with a CPA that
looked healthy and a pipeline that wasn't. Something was wrong
with the data, not the budget.
THE DIAGNOSIS
Cross-referencing search terms with SimilarWeb, SEMrush and
AnswerThePublic revealed the algorithm had learned to optimize
for job candidates, not hiring companies. 85.9% of the conversion
history was contaminated. The low CPA was real, and worthless.
THE REBUILD
New campaign from zero
Negative keyword list with 1,100+ terms
Mandatory company registration field
A CRM funnel that fires the conversion signal to Google and Meta only when a deal is marked Won
That last piece is what changed everything. The algorithm started
optimizing for closed contracts instead of form fills.
In parallel, the full Organum method ran on the site.
THE RESULT
Marketing ROI 1,184%
ROAS 31.9x
B2B cost per lead $6.58, down from $16.94
Close rate 62.5%
And the structural win: organic search overtook paid traffic for
the first time in the account's history. The business became less
dependent on ad spend every month.
The account also began appearing in ChatGPT citations within its
category, with sessions arriving from chatgpt.com (http://chatgpt.com) in Analytics.
1
31
98 vs 43: rebuilding a real estate portal to outrun the market
THE SITUATION
Real estate portals are heavy by nature: map rendering, image
galleries, live listing feeds, filter logic. The industry treats
slow as inevitable.
I tested that assumption against six portals, including the three
largest in Brazil, plus Zillow and Craigslist.
THE RESULT
Mobile PageSpeed Insights, measured September and October 2026:
Performance 98 43
Largest Contentful 2.0s 12.3s
Total Blocking Time 0ms 1,444ms
Cumulative Shift 0 0.011
Sum of all four Lighthouse scores: 398 out of 400. The closest
competitor scored 347.
It wins outright on seven metrics and ties at the top on three.
It does not lose a single one.
AGENTIC BROWSING
I also tested whether AI agents can read each portal. Brokers
passes 3 of 3. The three largest Brazilian portals fail. Zillow
passes. This matters more every month, as more purchase research
starts inside an AI assistant rather than a search box.
HOW IT WAS PROVEN
Claims like this get challenged, so the build was verified by
mutation testing: 101 test batteries, 52 mutants, 173 of 173
files matching by md5 checksum. The numbers are reproducible,
not asserted.
THE HONEST PART
A fair objection is that larger portals do more. True. But every
metric above is measured on the same device profile, same network
throttling, same tool. Complexity explains some of the gap. It
does not explain 10 seconds of LCP.
0
32