Ask HN: Why was no company able to replicate the Stripe API?

2 points

6 years ago

Stripe API is great. How they do it is available online. But why was no company able to replicate it? Thinking about everything I read about them, my take is that the APIs are only the tip of the iceberg.

- The first released API took 1 year and was the 4th rewrite. Even in a pre-product and pre-revenue startup. A 5 people company.

- Stripe hired the first PM at 150 people, 5 years after founding. Engineers have a strong product mindset. They hire a lot of engineers with product experience. [2] - Today they use the process created at Amazon of Working backward from the press release. By the time they start building the API: docs, FAQs, UI mockups, and a lot more is already built.

- In the North Star podcast patio11 mentions an example: if 8 people work on a project for 6 months, and the feedback is bad, they throw it away. [3] I saw teams cracking for having their work slowed to launch or launching with low usage. I think explicitly throwing it away would be 10x harder.

- The design process for the API is equal to any product with great UX. Many iterations of mockups, user interviews, usability tests, rewrites, and so on. This means much more design vs building time.

So, I think it's not about documentation or API design, it's more about culture and trade-offs. No company is willing to make the required trade-offs. Like:

- Throwing months of work away

- Postponing launch dates to improve nail quality

- Putting 10x more time into your API development time compared to today

These things cost a lot of money, without obvious ROI, and depends a lot on culture. Not easy to start doing after you have a not so good API in the market and a different culture in place.

What is your take?

[1] https://youtu.be/IEe-5VOv0Js?t=441

[2] https://www.quora.com/Does-Stripe-have-product-managers-or-do-engineers-manage-the-products-themselves

[3] https://youtu.be/YrTrWf6EjqA?t=3161

edit1: add line breaks

3 comments