Re:Lessons Learned from Payments Startups
by tommiegannert on Wednesday February 07, 2024 @11:24AM
(#39290481)
This is a nicely varied list of lessons.
> Often, I will prefer the database to logging
Just don't start writing exception logs to the database, and not tell anyone that you've created a flood gate that will make you run out of production data storage at the worst possible time.
Re:Lessons Learned from Payments Startups
by zeroxfe on Wednesday February 07, 2024 @12:03PM
(#39291024)
I wrote a [1] post-mortem when I shut down my payment startup a few years ago. The vast majority of our challenges were non-technical: navigating regulation, high legal costs, complex insurance requirements, etc. on an uneven playing field with competitors with much deeper pockets than even a well-funded startup.
Re:Lessons Learned from Payments Startups
by potamic on Thursday February 08, 2024 @01:58PM
(#39306008)
The payments platform you talk about is a piece of tech every financial company ends up building. What did you do with all that code after shutting down? Did you consider either selling or open sourcing it?
Re:Lessons Learned from Payments Startups
by umeshunni on Wednesday February 07, 2024 @12:26PM
(#39291334)
These are great lessons from an engineering perspective. From a product perspective, I've learned 3 hard lessons:
1. Fraudsters will find your product and abuse it before real users do. If you see a sudden spike in usage, it's probably fraud.
2. Real users are very emotional when it comes to their money, even if it's a "low volume transaction" like $20. Anything you build that introduces friction between them and their money will get your support team yelled at.
3. Regulators and banks are suspicious of you, for lessons they have learned from 100+ years of balancing lessons 1 & 2 above.
Balancing 1, 2 & 3 is an art and the difference between building a successful product and failure.
All the cute UI and funnel analysis you do breaks when it hits fraud, user emotions and regulation.
Re:Lessons Learned from Payments Startups
by user_7832 on Wednesday February 07, 2024 @03:48PM
(#39294119)
I'm curious, would you say using a third party to handle payments (stripe/paypal/shopify etc) would help avoid some of these issues? I'm not sure I understand point 2, especially if there's a "regular" checkout page by one of these companies.
Re:Lessons Learned from Payments Startups
by PeterisP on Thursday February 08, 2024 @12:32PM
(#39304862)
It would help avoid these issues, but if you're a payments startup then you probably want to actually keep some control over your core business instead of just being a reseller intermediary for an actual payments company which holds your customer's money, has all the leverage over what features and policies "your" product will have and gets most of the profit.
Re:Lessons Learned from Payments Startups
by monknomo on Thursday February 08, 2024 @12:49PM
(#39305060)
In some ways it makes it worse, because your users will not understand that you aren't screwing up their transaction, it's Paypal. They'll just yell at you
Re:Lessons Learned from Payments Startups
by threeseed on Thursday February 08, 2024 @02:46PM
(#39306668)
None of this is unique to payments or the financial sector.
And it is bizarre to me that you could run a payments system without high-availability or immutability at its core.
Also putting logs in a database alongside your core data is such a bad idea. There are plenty of solid, proven systems for log management that allow you to correlate against other databases. And it avoids the issue of you wasting limited write I/O on a non-priority task.
Re:Lessons Learned from Payments Startups
by anonzzzies on Thursday February 08, 2024 @09:29PM
(#39310724)
In how many instances though? This sounds like premature optimising. We are not all Stripe; many companies making millions, but not billions, would have an easier life without doing more work or paying a fortune to niche saas companies for these type of things.
Re:Lessons Learned from Payments Startups
by pgr0ss on Thursday February 08, 2024 @05:42PM
(#39308811)
Hi, I’m the author of the blog post.
> And it is bizarre to me that you could run a payments system without high-availability or immutability at its core.
I’m not sure what you’re referring to here as my post didn’t really go into these topics.
> Also putting logs in a database alongside your core data is such a bad idea.
I’m not suggesting putting all logs into the database. I’m suggesting that for a certain type of info (that is often logged), I find it more useful to put it in the database instead.
> Often, I will prefer the database to logging
Just don't start writing exception logs to the database, and not tell anyone that you've created a flood gate that will make you run out of production data storage at the worst possible time.
[1] https://www.linkedin.com/pulse/quid-sudo-shutdown-h-now-mohi...
I am surprised you didnt mention fraud - is it because micro-payments are not "fraud worthy" or because it is a manageable problem ?
But also, yes, micropayments are far less "fraud worthy" from an attackers perspective.
1. Fraudsters will find your product and abuse it before real users do. If you see a sudden spike in usage, it's probably fraud.
2. Real users are very emotional when it comes to their money, even if it's a "low volume transaction" like $20. Anything you build that introduces friction between them and their money will get your support team yelled at.
3. Regulators and banks are suspicious of you, for lessons they have learned from 100+ years of balancing lessons 1 & 2 above.
Balancing 1, 2 & 3 is an art and the difference between building a successful product and failure.
All the cute UI and funnel analysis you do breaks when it hits fraud, user emotions and regulation.
And it is bizarre to me that you could run a payments system without high-availability or immutability at its core.
Also putting logs in a database alongside your core data is such a bad idea. There are plenty of solid, proven systems for log management that allow you to correlate against other databases. And it avoids the issue of you wasting limited write I/O on a non-priority task.
Is it really an issue, though, in most cases?
https://news.ycombinator.com/item?id=39276069
> And it is bizarre to me that you could run a payments system without high-availability or immutability at its core.
I’m not sure what you’re referring to here as my post didn’t really go into these topics.
> Also putting logs in a database alongside your core data is such a bad idea.
I’m not suggesting putting all logs into the database. I’m suggesting that for a certain type of info (that is often logged), I find it more useful to put it in the database instead.