What I learned from outsourcing the development of my first app
Outsourcing wasn't SpeakOn's mistake: not understanding how it was built was. Today, with AI, I could build something like it without someone technical on my side; for hard tech or greater complexity, I'd still need that person.

- If you don't understand how your product is built, you don't control your company.
- An app designed for a thousand users may not work for a hundred thousand; you have to design it for the day it takes off.
- Today I could build something like SpeakOn with AI without someone technical on my side; for hard tech or greater complexity, I'd still need that person.
Outsourcing app development isn't the mistake. In 2015 my mistake was doing it without understanding how it was being built, or having someone on my side who did. I learned that with SpeakOn, my first app: it hit #1 on the App Store in Mexico with 200,000 users in 48 hours and went down because its architecture couldn't handle the load.
If you're about to hire an outside team to build your app, this post is the checklist I wish I'd had in 2015.
How I put together the SpeakOn team
SpeakOn was a chat app that translated in real time across 60 languages. The idea came from my conversations with Alibaba suppliers, where we couldn't understand each other.
I didn't code. I posted openings on Indeed and other places, and I also hired people from UNAM who were supposedly very good. I ended up with four developers: two backend, one Android, and one iOS. The backend was built in Go (Golang) with MongoDB.
I didn't look for a technical cofounder. I didn't think I needed one.
What happened when it worked
We launched with a party at Hotel Habitat in Mexico City, with celebrities and press. At the office we had a screen showing downloads and sign-ups climbing at roughly one per second.
Then people kept downloading, but they couldn't sign up anymore. After that, everything went down, including the website.
The Go community was small and closed, and I couldn't find anyone to help us. I brought in two of the best Go engineers in the US. Their diagnosis was short: "This has to be rebuilt. It can't be fixed."
They explained it like this: it's like a house. If you didn't design it for a pool on the top floor, adding one later is almost impossible.
We tried to rebuild it. It took too long and we couldn't keep the users. The full story is in SpeakOn: the app that hit #1 on the App Store in Mexico and why it died.

Screenshot of SpeakOn at #1 among free apps on the App Store in Mexico, above WhatsApp and Facebook. SpeakOn archive.
The signal I didn't know how to read
When I applied to Y Combinator with SpeakOn, we made it to the first interview and didn't get in. The reason: a lack of "technical skills" on the team.
They were spelling it out for me. The problem wasn't code someone wrote badly. It was that nobody at the decision-making table understood how the product was built.
Is outsourcing app development a bad idea?
No. Plenty of good companies started with an outside team. The risk isn't in who writes the code, but in who reviews the decisions.
When you hire developers and you don't know what to ask for, they build what they know how to build, with whatever architecture is most comfortable for them. Nobody does anything wrong on purpose. It's just that nobody is looking at your product's future through your eyes.
The four mistakes I made
| Mistake | What happened | What would have prevented it |
|---|---|---|
| Having no one technical on my side | No cofounder or advisor was reviewing the team's decisions. | A technical cofounder or an advisor who answered to me. |
| Building for the launch, not for success | The architecture worked for a small app; the launch was for a big one. | Defining from day one how many users it had to handle. |
| Not testing under real load | We never simulated thousands of sign-ups at once. | Load testing before the party. |
| Launching to everyone at once | The party, press, and celebrities brought everyone in on the same day. | A waitlist or invites. |
None of the four is a coding mistake. They're founder decisions.
How do you choose who to hire to build your app?
Here's what I'd check today before signing with an agency, a freelancer, or an in-house team:
- That they've already built something that handled volume. Ask how many concurrent users the last thing they built supports and how they know.
- That they can explain the architecture in plain words. If they can't, they won't tell you when something's wrong either.
- That there are people available for the technology. With Go in 2015, I struggled to find anyone to help me. If you pick something with a small community, have a reason.
- That they accept outside review. A good team doesn't get offended if someone else reviews their work.
- That the code and accounts are yours from day one. The repository, the servers, and the app store accounts in your name.
Do I need to know how to code to start a tech startup?
No. And today I wouldn't say I need someone technical on my side to build something like SpeakOn either. With AI, I could do it myself. The world changed recently.
For hard tech or more complex products, I would still need someone technical. Understanding how my product is built still matters; what changed is how I can build it.
After SpeakOn, I decided to learn code and architecture. Today I understand architecture, infrastructure, services, and microservices, and I can read code even if I don't write it.
I didn't become a programmer. But now I know what to ask a technical team, like how many users the system can handle and what breaks first, and I understand the answer.
Why do apps crash when they go viral?
Because their architecture was designed for few users. While the app is small, almost any design works. With thousands of sign-ups at once, the parts that weren't built to grow get overwhelmed.
And it happens right when the most people are watching.
What I'd do today if I outsourced an app
- Separate what I can build with AI from what needs more technical expertise. Today I could build something like SpeakOn with AI, without someone technical on my side. For hard tech or greater complexity, I would still need that person.
- Write down the target number of users in the contract or the scope document.
- Run load tests before any campaign.
- Launch in stages, even if it feels less spectacular.
- Learn enough to ask the right questions. You don't need to code; you need to know where what you're selling can break.
What I took away
If you don't understand how your product is built, you don't control your company.
That line has stayed with me through everything I've done since. I tell it along with the rest of my companies in who I am and why I now build in public.
Frequently asked questions
Is outsourcing app development a bad idea?
No. At SpeakOn in 2015, the problem was not understanding how it was built or having technical oversight on my side. Today, with AI, I could build something like it without that person; for hard tech or greater complexity, I'd still need them.
Why do apps crash when they go viral?
Because their architecture was designed for few users. With thousands of simultaneous sign-ups, the backend gets overwhelmed, and fixing it can mean rebuilding the app.
How do you keep an app from crashing at launch?
With load testing before launch, an architecture built to scale, and a staged rollout, like a waitlist.
Do I need to know how to code to start a tech startup?
No. Today I could build something like SpeakOn with AI, without someone technical on my side. For hard tech or greater complexity, I'd still need that person. Understanding how the product is built still matters.
Learn how to build a company by watching me build mine.
Every week: the decision, the number, the mistake and the win.