I’ve analyzed quite a few data sets from startups over the past 10 years and I have some core insights that are too brief to blog about, but useful just the same:
First, most startups don’t get anywhere until about year two or year three, and even once they’re solving real problems for happy clients things still feel “broken” because corners have to be cut somewhere. One startup manually charged credit cards with a machine.
Another had a new DB per client, and not for performance reasons. The founder was a business type that didn’t really know better. It ended up saving them because the codebase was so riddled with security bugs the DB was frequently owned.
Another company had a super viral product with a horribly brutal churn rate. My advice was to close up that product line and pivot to something else. To my surprise they actually listened to me and were ultimately acquired for nine digits.
Out of all the companies I’ve analyzed data for only a handful took any precautions to obfuscate or encrypt potentially sensitive client data. I know it’s shitty, but it’s the reality. They’re not incentivized to mitigate risk.
The thing that seems to correlate with success is primarily determination. The ones that fight hard to fix the problems succeed. The ones that fuck around with new unproven tech over just building features or solving other problems usually fail to massively succeed.
Some other common patterns I’ve noticed: Winners tend to use MySQL or Postgres over other stores. They also (surprisingly) tend to just shell out in their codebase more.
Lastly, complex deployment seems to be an anti-pattern. Most of the successful startups I've helped just use shell scripts or simple-ish deployment like chef or similar.
~fin~
~fin~
