Globo · Globoplay · 2020
Globoplay Highlight
The home's prime component, rebuilt from a monolith into a system — then the experiment that lifted engagement and the variant that drove sales.
The stage · 2020
Brazil's largest streaming product, becoming all-in-one.
Lockdown drove a streaming surge and cord-cutting hit its peak. In that year Globoplay was folding everything into one product: the video-on-demand catalog, simulcast of 17 live TV channels, live news and sports — and it carried the group's largest investment of the year.
Streaming surge and cord-cutting, 2020. BBC · TechCrunch.
The prime slot
The most contested space in the product.



At the top of the home sat the Premium Highlight. First in impressions, first in click-through rate, the origin of more than half the clicks on the page. Every channel director wanted their content there.
#1
Impressions on the home#1
Click-through rate+50%
Of all clicks on the homeThe problem
One block of code, asked to become many things.
The Highlight was a monolith with business rules nested inside it. The new plan asked it to behave as many products at once. It had to become parameters.
5Content types
- Title
- Video
- Live
- Simulcast
- Sports
2User profile
- Subscriber
- Free
2Layout
- Premium
- Rail
Real variant screens, iOS.
The scale
5 × 2 × 2 =
It had to become modular.
Rebuilt to be modular and easy to scale — ready for the next scenario without touching the whole.
The rebuild
Rebuilt from the atoms up.
I broke the monolith into an atomic system: the smallest pieces compose into groups, groups compose into the finished component. Change a part, not the whole.
Component vs product
Separate the interface from the business rules.
One component, two readings. The interface declares what each slot expects — an image, a string, a crest, a button. The product layer wires each slot to a data source. Same shell, any content pipe behind it.
- Image / video URL
- Image URL
- Image URL + number
- Strings
- Button element
- Broadcast API
- Channels API
- Sports Data System API
- Electronic Program Guide API
- Broadcast API
The bet
Lead discovery from the one place people look.
Users barely scrolled the home. Their clicks concentrated on that first component. The hypothesis: bring more content into the prime area and let it lead discovery, lifting engagement without asking anyone to scroll.
Research
No one else put a carousel on the home.
With no time for a usability test before launch, I did two things: benchmarked how other streaming products surfaced content on their home — none used a carousel there, which made it a bold bet — and leaned on published carousel research to set guidelines.
- Netflix
- Prime Video
- Disney+
- HBO Max
- YouTube
- Star+
- Paramount+
- Apple TV+
Home-screen benchmark, 2020. Carousel guidelines drawn from Nielsen Norman Group research.
The carousel
An Instagram-story model for the prime slot.
Full-bleed media with story-style progress bars that advance — the signature interaction, one model across every platform.
One model, every platform
Consistency meant a reason for every difference, not making everything identical.
The experiment
Data is king.
A clean A/B test, 50/50 to all users. The standard Highlight stayed as a fallback so nothing could break the page. With millions of monthly users, one month reached significance.

Control · standard highlight
Variant · carousel
A/B test · 50/50 · fallback-safe · Aug 2 – Sep 2, 2020.
Results
Impressions held where research said they'd fall.
Carousel research warns that content past the first slide gets buried. Ours didn't — most users reached the third slide, on every platform.
Three clickable items now stood where one used to. The hypothesis held.
+75%
Click-through rate — expected: three clickable items against one+2.7%
Average playtime+5%
Titles per user — the engagement indexFrom engagement to revenue
A sales variant, shipped fast because it was modular.
Componentization plus the carousel win let us ship a sales variant without rebuilding anything, launched with the Disney+ bundle.
15.8k
sales in the first month11%
of total sales originated on the home21%
that home relevance, up vs. the previous monthNov 1 – 30, 2020.
The year, in context
Globoplay's strongest year.
2020 made Globoplay the most-watched streaming service in Brazil. The numbers below reflect the work of every team, well beyond this project. I include them as context, not as my result.
+145%
Subscribers+336%
TV-series playtime+403%
Movies playtimeGloboplay full-year results, 2020 — company-wide, all teams.