The first time we saw it, we were sitting in on a Tuesday standup at a fintech we were about to start with. Three people on the ops team had spent the morning copying values between a Google Sheet, an internal admin panel, and Stripe. They estimated it took "about an hour."
We watched it twice that week. It took two and a half.
That's not the cost we're talking about, though.
The line-item that isn't there
When the CFO looks at the budget, manual ops is mostly invisible. Headcount is there. Tools are there. The work is everywhere and nowhere — absorbed into the salaries of people who do other things too.
A reasonable team can carry a surprising amount of it. For a while.
The thing you can't see in the budget is who is doing it, and at what cost to the rest of their job. The senior ops manager who spent her morning reconciling spreadsheets did not also spend her morning thinking about how to design a better reconciliation process. That trade is real, and it's expensive in a way that doesn't show up until she leaves and the team realizes nobody else was doing the design work either.
The second-order cost
Manual ops doesn't just take time. It selects.
The kind of person who can do an hour of reconciliation cheerfully every morning is not the same person you hired to think about ops. Over time, the work shapes the team — either by burning out the thinkers or by quietly attracting people who don't mind a Tuesday morning of copy-paste.
Both ends are bad for you.
What the math actually looks like
Take a process that costs forty minutes a day across three people. Five days a week. Fifty weeks a year. That's five hundred hours. At a loaded cost of $100 an hour, that's $50K in time, before you count the mistakes.
Now count the mistakes. A 2% error rate, with each error costing an hour to track down and fix, adds another fifty hours. Most teams underestimate this number by half.
Now count the things that didn't happen because the team was doing this instead. That number doesn't fit on a spreadsheet. It shows up later, as a competitor that ships something you should have shipped.
What we look for first
When we come in, we don't look at the loudest pain. We look at the quietest:
- Where is someone doing the same thing every morning?
- What gets done at 7pm, after everyone else has gone home?
- What never gets done at all, because the person who used to do it burned out?
The work that's quietly degrading the team is almost always the work that's most worth automating. Not because it's the most expensive on paper. Because removing it changes the shape of the team.
You don't automate to save money
You automate to give the people back the kind of work they were hired for. The money savings are real, but they're the second-order effect. The first-order effect is that the senior person who was doing reconciliation gets her mornings back, and the next person you hire arrives into a team that thinks instead of types.
That's the math we keep coming back to.