We're Collectors of Ops Software Stories

We're Collectors of Ops Software Stories

Angus Mitchell's photo
Angus Mitchell
·Mar 31, 2022·

3 min read

Subscribe to my newsletter and never miss my upcoming articles

(If you haven't read the high level overview, Simple x "Never" = Insane, you might want to start there.)

Some of these stories are insane, and it's tempting to make of fun of these companies for being dysfunctional. But they're not dysfunctional. They're just people with tight deadlines and insufficient tooling.

We needed a tool to cancel loans. It was one endpoint and a few input fields. It took from December to end of February.

— PM at Series C payments company

 

Pricing is controlled by “configurations” which are JSON files. They can apply to the city or state level. It’s confusing for people to learn. It took me 2 months to get really comfortable, and I still make mistakes. One time I formatted it incorrectly, and it took the down the whole app in [major US city] for several hours. It was at least a $30,000 mistake.

— Ops Manager at major scooter company

 

Most of the “engineers” can barely code. You’d be shocked at how manual everything is. A lot of engineers take pride in the fact that they’re waking up at 8am to manually run jobs.

— Senior Data Scientist at Big 4 Consulting Firm

 

Our Retool apps aren’t built on the production database. They’re built on a copy of it. So there are certain things we just can’t do. At a minimum they take a week. Normally it's closer to a month.

— Retool guy at Seed-stage transportation company

 

I used to walk into the CTO’s office and he’d build me a tool with 10 lines of code. Now it takes a month to get anything because we’ve grown so much

— Operations Manager at Series C healthcare company

 

I’m responsible for onboarding our new hires. A lot of our workflows are complicated. They have to jump between different pages. It’s a bit of a workout to jump around. There’s a TON of room for human error.

— Director at Series C healthcare company

 

We would have to bribe engineers to get our tools built. Not literally bribe them with money. But I would send girls from our group to go flirt with them. It was crazy. We had a huge backlog. The attitude was that we’d hire more Ops people until we automated it all, but people were kicking and screaming and quitting.

— PM at Series E logistic company

 

Engineers generally didn't give the Ops team the SQL queries we needed. Some Ops people had the skills to write the queries themselves (they tended to share with everyone else), while some people got an assist from a friendly data scientist / data analyst. I think this is worth mentioning because it meant that such a critical business process as paying contractors wasn't standard across the company.

— Ops at Uber

 

I don't think the issue was so much that engineering was uncomfortable ceeding some control to Ops, as it was that the lack of...well a lot of things...meant that engineering had no choice but to cede too much control to ops, which led to the possibility and to some extent the inevitability that everything would break.

— Ops at Uber

 

We have one workflow where customers have to answer an arbitrary math problem to avoid Canadian gambling laws and get a redemption card. This works fine unless the customer makes an arithmetic mistake. In that case, the only person with the credentials to see the answer is the CEO, so he has to get roped in.

— Ops at marketplace company

 

We normally do about 4-5 projects per year [each project is a couple form inputs fields, basically]

— CTO at recently acquired ad-tech company

 
Share this