Every year, LinkedIn used to publish a list of Brazil's most desired startups, built from four signals: headcount growth, brand engagement, interest in job openings, and the ability to attract people away from other large companies. Rereading the 2019 edition today is a useful exercise, because the ranking measured desire, not solidity — and time has separated the two.
What the 2019 list said
Fintechs dominated. Nubank ranked first for the second year running, C6 Bank came in second, and Neon, Creditas, Warren, Koin, Zoop, Contabilizei, and Atlas Quantum were scattered across the rest. Next came logistics companies, such as Loggi, Cobli, and CargoX, and real estate and construction companies, with Loft in third, QuintoAndar in fifth, and EmCasa in eighth. There was also mobility, human resources, airline miles, food, and education.
What time did to them
Their fates turned out very differently, and that's the lesson:
- Some confirmed their promise. Nubank went public on the New York Stock Exchange in December 2021 and became one of the largest financial institutions in Latin America.
- Others shrank. When cheap money ran out in 2022, several of the companies that had been hiring aggressively went through layoff rounds and traded growth for margin.
- And some disappeared. Atlas Quantum, which appeared on the list, collapsed amid customer problems and serious accusations. Shared mobility companies shut down operations in Brazil.
None of that was visible in the ranking's four criteria. Fast hiring and interest in job openings measure attractiveness as an employer at a moment of abundant cash. They don't measure whether revenue covers cost.
Why this matters when choosing a vendor
The same trap shows up when a company is about to hire software, a platform, or a service: the list of awards, the recent funding round, and the growing team are presented as proof of stability, and they aren't. They're indicators of expectation, and expectation can shift meaning fast.
The questions that actually say something about whether a company will still exist three years from now are different ones: does it charge for what it delivers or is it subsidizing the price, how many paying customers does it have, what happens to your data if it shuts down, and can you export everything today if you want to leave. It's the same reasoning as the Guiabolso case, an app with millions of users that was bought and shut down.
Choosing technology by reputation instead of by testing is the mistake we described in survivorship bias, and that's why any AI implementation project should start by measuring your task, not the vendor's list of awards.
Frequently asked questions
How did LinkedIn choose the most sought-after startups?
By four criteria: headcount growth, engagement with the brand on the platform, user interest in job postings, and ability to attract professionals coming from large companies. These are indicators of employer appeal, not of financial health.
Did the startups from the 2019 ranking actually pan out?
Their paths diverged sharply. Nubank, the top-ranked company, went public in New York in December 2021 and became one of the region's largest financial institutions. Others shrank and laid off staff once cheap money dried up in 2022, and some vanished altogether — Atlas Quantum among them, which collapsed.
Can you judge a tech vendor by its awards and funding rounds?
No. Awards, fundraising, and headcount growth measure expectations, and expectations can flip direction fast. The questions worth asking are whether the company charges for what it delivers or is subsidizing the price, how many paying customers it has, what happens to your data if it shuts down, and whether you can export everything today.