Ask HN: How do you develop MVP if your main feature depends on an external API? | Hacker News Reader

Ask HN: How do you develop MVP if your main feature depends on an external API?

6 pointspmtarantino17 comments

I developed a web application which I find very useful. It's for a niche and I started because it resolved a problem I had.

I shared it with a few friends (who are part of this niche too) and they like it. My problem is that the main feature (the only one feature, in fact) depends on a few external API. And I am afraid if I release and start asking money for it, it stops to work for external reasons (API shuts down, API limit rates), etc.

So my question is, how do you handle this?

PS: The APIs I am using are free and have not a paid plan with more limits.

View on HN

Never "bet the farm" building on top of an API / service / platform that you don't control. It's almost inevitable that it will come back to bite you. If it's just a side project, or just something you want to try and generate some passive revenue from, and you won't lose your shirt if you get shut down, then sure, go for it. But if this what you intend to be the next billion dollar thing, and you're hamstrung by being tied to someone else's API, I'd consider going in a different direction.

Just search HN for all the horror stories over the past few years where some promising service got screwed over by changes from upstream providers (Twitter, Craigslist, etc., etc., etc.).

This. I'm mainly giving advice in a "side-project, passive revenue" mindset because I assume doing anything else on top of someone else's api is nothing short of suicidal. Keep it low-key and manage your expectations and you'll be fine.

Yes, it will be for sure a side-project, nothing like a new Tumblr or anything. Thanks :)

Many people here say that it's suicidal to bet it all on a single API.

Yet, there's also a LOT of businesses that have thrived and they rely on API's. Examples being HootSuite (Twitter/FB api), SocialBro, BufferApp, SproutSocial, Klout, and a lot of social media analytics platforms. It's certainly a risk, but it's better to go forward with your plan, than not even try because you're afraid of being too over-reliant.

To mitigate the risk, is there anyway you can reproduce what the API does? Even if it's slightly shady, such as scraping.

Yes, I could scrap some webpages, but is it okay to make a business out of it?

No right answer. But many other businesses have done it. Look at SEOMoz, who've built tools centered around scraping Google results and social media pages.

Go for it. The worst that can happen is that EBay says "You can't do this anymore" and you go back to just being a service for you and a few friends. Its probably not gonna be a very sustainable business either way, but there's nothing stopping you from trying, and at least succeeding for a short time.

Yes, that was my thought. But what I say to people who paid for it? Should I return its money?

Charge monthly. That's better for a hosted service anyway, as it matches up better with your expenses. If it stops working, stop charging them. Refund months they've already paid for that you won't be able to fulfill. Make sure they know up front that this service depends on EBay not being a jerk. That way you'll still get paid for however long you DO operate it, and not have to feel guilty about keeping the money when it stops working.

Yes, if I charge the uses, it has to be monthly, but even that, it's scary for me :) I never charged anyone online.

Sounds like you're using ecommerce APIs. Check out https://www.semantics3.com. It might help you take some risk out.

Thank you, I will check it :)

Cache everything you can. Do anything in your power to reduce api calls. Of course this only works if you are primarily reading data as opposed to writing.

Try to talk to someone at the API provider. Get their blessing.

a) are there alternatives you can build on

b) can you replicate those APIs yourself over time and reduce then remove the dependency?

I am using eBay API, among others, so no, I can't (for both questions)

Well you need to decide then really. If you're not infringing any terms (or the spirit of them) then the opportunity can easily outweigh the risk.

Reply on news.ycombinator.com