01 · Context
Every marketing change waited in the product pipeline
KTO was a licensed betting platform in Brazil, and everything a visitor saw before logging in lived inside the product's codebase. A new landing page, a banner swap or an SEO fix had to enter the same pipeline as product features and wait its turn. By the time a campaign page shipped, the moment it was made for had usually passed, so most urgent marketing ideas never got built at all.
Organic search was the third-largest source of active users, about 15% according to Google Analytics, behind direct traffic and affiliates. The surface that received those visitors was the one Marketing could not touch.
02 · The old site
The logged-out site was the logged-in app with parts hidden
Technically, there was no logged-out product. The site was a single app built for logged-in players that hid elements when nobody was signed in. Every page carried the structure, navigation and weight of the product, and none of it was designed for someone arriving from a search result.
The initiative to change that existed before I joined, and I was hired to design it. The marketing director who sponsored it left soon after the project began, and the team building it carried it forward on its own: a fullstack developer, the Head of SEO, the Digital Marketing Lead and me. A PM joined near launch to align the rest of the company, set expectations and plan the release.
03 · Two audiences
Visitors and players were different audiences
Before designing anything, I compared how logged-out visitors and players behaved, using Google Analytics, Mixpanel, the results of paid traffic tests, interviews with stakeholders and an audit of the existing architecture. One number summed it up: 97.6% of site searches came from logged-in users (Mixpanel). Players searched for a game they already knew. Visitors arrived from Google on a specific page and needed that page to answer their question.
That ruled out the easy fix of restyling the existing pages. A visitor needed content, context and a clear path to registration, while a player needed speed and their account. Serving both from the same structure meant serving neither well.
04 · A separate site
A separate site Marketing could run without engineering
The logged-out area became its own product, a static WordPress site with its own stack, so Marketing and SEO could publish without writing code. My job was to make that freedom safe. I designed components that held up under any content Marketing put in them: long and short titles, campaigns with and without images, pages with one block or twenty.
I also split the site by intent. A single homepage trying to rank for every search would have ranked for none, so each vertical got its own content pages, with promotions in a separate layer Marketing could change without touching the rest. I defined the toolkit Marketing used to test variations of those pages without engineering.
05 · The SEO header
The SEO header settled a fight between the campaign hero and the H1
Marketing wanted a campaign carousel at the top of the page, and SEO needed a clear H1 that described the page for search engines. Either one alone hurt the other: a carousel first buried the heading, and a heading first pushed the campaign down.
I designed an SEO header block that sat above the carousel and belonged to Marketing to edit. It gave each page a real H1 and a short introduction for search, while the campaign kept its place right below. Both teams got what they needed without duplicate headings or a compromise neither liked.
06 · Bottom navigation
I chose not to fight over the bottom navigation
The product team asked us to bring the logged-in bottom navigation into the logged-out site. We argued against it at first, because it tied the two codebases together and gave visitors a menu built for players. Eventually I chose to concede. The design and development cost was minimal, and the discussion was consuming more effort than the decision was worth.
07 · Results
Search got better, and Marketing stopped waiting
The site launched on 26 November 2025. Comparing the three months before and after launch in Google Search Console, the organic click-through rate rose from 15.9% to 24.2%, and the average search position improved from 6.2 to 5.6. In a 15-day window about a month before the 2026 World Cup, Mixpanel tracked 49,984 visitors starting a registration; 39.5% of them created an account, and 52.1% of those accounts completed identity verification.
The bigger change was who owned the site. The logged-out home became Marketing's product. They ran tests and redesigns, decided what the top banners offered, managed the content of every page and launched campaigns, with the graphic design team only producing the assets. During the World Cup, the Head of SEO asked me for a landing page of the biggest prize winners, and it went from request to live in one week. Through the old product pipeline, with two-week sprints and no guarantee a request made the next one, a small change took up to a month and a larger one one to two months, by my estimate. After launch, the Casino team asked for its own logged-out revamp.
08 · What I'd change
What I'd do differently
My own biggest mistake was in the promotions block. I designed its first version assuming there would always be an active campaign, and the layout broke the first time there wasn't one. I redesigned the components to handle the empty state, and since then every component I design starts from the question of what happens when there is nothing to show.
Organic traffic also dropped after launch, with a Google algorithm update and the re-indexing that follows any migration hitting at the same time. The Head of SEO and I took a second lesson from it: map every redirect before launch, monitor indexation from the first day, and have instrumentation ready before the switch, not after. Each of those would have made the weeks after launch calmer and easier to read.