Nov 2 · Remi · I filed Remi's trademark applications with IMPI, Mexico's trademark office. ›
← Blog

Who Carlos Gotlib is: three companies, one acquisition, and why I'm building in public now

I founded three companies, bought one, and have invested in startups since 2016. Almost everything I know I learned from what went wrong. This is the story in order, told as decisions, and why this time I'm telling it as it happens.

Carlos Gotlib6 min read
Carlos Gotlib in a dark sweater next to the title "Who am I and why now?"
In short
  • SpeakOn (2015) and WeHelp (2019) both hit #1 on the App Store in Mexico; neither survived as a business.
  • I've invested in startups since 2016, mostly Y Combinator companies: almost half have shut down.
  • Kimik sold a million dollars in six months and TurboCar 2.5 million; both shut down.
  • The mistake I keep making is starting fast and doing too many things at once. This time I'm choosing slowly and sharing it as it happens.

I'm Carlos Gotlib. If you got here from a video or a random post, this is the context you're missing.

It's not a résumé. It's the list of decisions that brought me here, what went wrong in each one, and what I took away. Each company has its own full post; here's the short version, in order.

2015: SpeakOn, the #1 app that couldn't handle its users

SpeakOn was a messaging app that translated in real time across more than 60 languages. You wrote in Spanish and the other person read it in theirs.

In June 2015 it hit #1 among free apps on the App Store in Mexico, above WhatsApp and Facebook. 200 thousand users in 48 hours and 350 thousand sign-ups in total.

And that's exactly what killed it. The app couldn't handle the load, and fixing it meant rebuilding it from scratch.

The decision that killed it came before the success: I hired developers through Indeed and at UNAM, I didn't look for a technical cofounder because I didn't think I needed one, and nobody on my side understood how it was being built. After that I learned architecture: today I can read code even if I don't write it.

What I took away: if you don't understand how your product is built, you don't control your company.

The full story: SpeakOn: the app that hit #1 on the App Store and why it died. And on the development side: What I learned from outsourcing development of my first app.

2016: I start investing

I applied to Y Combinator with SpeakOn. We made it to the first interview and weren't accepted for lack of "technical skills." Something I didn't expect came out of that: I started connecting with the YC community and investing in its companies. I joined their seed rounds because I like helping other founders: I know what it's like to be there.

Since then I've invested in 32 startups, 27 of them from Y Combinator, with checks of around 25 thousand dollars. Today 17 are still alive, 14 shut down, and one had an exit: Plerk, acquired by minu. That exit was paid in minu stock, not cash.

What I took away: almost half don't make it, even with the best accelerator in the world behind them. Watching other founders win and lose from the outside is half of my judgment.

The details: 32 startup investments: what the 14 that shut down taught me. The full portfolio is in Investments.

2019: WeHelp, which later became Donna

WeHelp was a personal safety app: an SOS button that alerted your contacts, and later a call center and monitored trips. In January 2019 it hit #1 among free apps on the App Store. Again.

It grew to 700 thousand registered users. In 2022 we added physical security and a call center. Most of the people using it were women, so in 2023 we focused only on them and renamed it Donna.

Donna had a 24/7 call center and plans from 99 to 149 pesos a month. The answer we heard most was that safety should be the government's job. We couldn't keep the call center going.

What I took away: people using something doesn't mean they'll pay for it. Users aren't a business.

The story: Donna: 700 thousand users and it never made money and Why people don't pay for safety apps.

2019: Kimik, the product the market didn't want

In November 2019 I started Kimik: buying ads on 2,500 outdoor screens in minutes, the way you buy an ad on Facebook. We used geolocation to measure who walked past the screen and whether, in the next 7 to 30 days, they visited the store or downloaded the app.

We sold a million dollars in the first six months. And we shut down in 2021.

The product worked. The problem was the market: a few companies control the screens in Mexico, and they make more selling by the month than programmatically. They had no reason to change.

What I took away: a product that works won't survive a market that doesn't want it. Before building, look at who controls distribution and whether your existing is good for them.

The story: Kimik: a million dollars in 6 months and we still shut down and Why programmatic outdoor advertising isn't taking off in Mexico.

2024: TurboCar, my first acquisition

In July 2024 I bought, for 160 thousand dollars, a shop in Miami that bought wrecked cars at Copart and IAAI auctions, repaired them, and sold them.

I was tired of tech and wanted something I could touch. It was a really bad purchase: the shop was a mess and a lot of employees left when I arrived. I switched it to luxury salvage, and profit per vehicle went from 1–2 thousand dollars to 10–15 thousand. We sold around 2.5 million dollars.

I still lost money. More and more cars were being marked as destruction instead of salvage; I think because of some regulation or insurance-company lobbying, though that's my opinion. There were fewer repairable cars, more people bidding, and we could no longer buy at a price that made sense. I shut it down in March 2026. It was one of my worst decisions.

What I took away: a rule. I don't buy a business in the United States with less than a million dollars a year in EBITDA. Below that, you're buying a job, not a business.

The story: TurboCar: my first acquisition and why I shut it down.

The pattern

If I put it all together, there's one mistake that keeps repeating, and it isn't technical.

I was the guy who did everything really fast. An idea would come to me and I'd want to do it that same week. I have a ton of ideas, and that worked against me: too many things at once, my attention split, nothing with the depth it needed.

So this time I did the opposite. I spent four or five months choosing what to build before starting anything. It felt unnatural, and that was exactly the point.

And there's something else I learned: there's nothing worse than a company that doesn't go under but doesn't grow much either. I'd rather have a company that doesn't sell than one that sells just enough to keep me hooked but never fully takes off. When it doesn't sell, it forces you to shut it down and move on.

What I'm building today

What I'm building today:

  • Remi: the company I'm building today, which I chose in public out of 112 ideas. A monthly membership in Mexico for people with diabetes, high blood pressure, or weight issues.

Where it stands this week is in What Carlos Gotlib is doing. The full plan, in Remi.

Why in public

Everything above I wrote after the fact. I already knew how each story ended, and that makes it easy to make sense of everything.

This time I'm telling it as it happens: every decision with its date, every number, every test that failed. Without knowing yet if it'll work out.

And telling it this way forces me to do what's been hardest for me: say what I'm going to do and when, and then show whether I did it. If I get scattered again, it'll be in writing.

How it works, and what I share and don't, is in What you'll see here.

The decisionThis time, choose slowly and share it as it happens, not after.
The number3 companies founded, 1 acquired, and #1 on the App Store in Mexico twice.
The mistakeDoing too many things at once and, because of that, not giving enough time to validating whether the problem was real and whether there was product-market fit.
What went rightRealizing that I know how to build a product, put it in people's hands, and get users and customers.

Frequently asked questions

Who is Carlos Gotlib?

Founder and startup investor. He founded SpeakOn (2015), WeHelp, later Donna (2019), and Kimik (2019); bought TurboCar in 2024; and since 2016 has invested in 32 startups, 27 of them from Y Combinator. Today he's building Remi in public.

Was Carlos Gotlib in Y Combinator?

Not as a founder. He applied with SpeakOn and was rejected after the first interview. That's how he started investing in YC companies.

Which of Carlos Gotlib's apps hit #1 on the App Store?

Two: SpeakOn, in June 2015, with 200 thousand users in 48 hours, and WeHelp (later Donna), in January 2019, which grew to 700 thousand registered users.

What did Carlos Gotlib learn from his companies?

That his recurring mistake is starting fast and doing too many things at once. That's why he spent four or five months choosing what to build before starting, and now he's building it in public.

Why does Carlos Gotlib build in public?

Because telling it as it happens forces him to say what he'll do and when, and then show whether he did it, without knowing yet if it will work out.

Learn how to build a company by watching me build mine.

Every week: the decision, the number, the mistake and the win.