The Problem You've Been Meaning to Fix Since March
This is Learn Leadership. Every issue takes one thing that quietly goes wrong between managers and their teams, shows it happening through a real story, and leaves you with something to try this week.
Today it's the known problem everyone works around and nobody gets to, because it's never this week's emergency. You'll leave with a twenty-minute backlog review that gives each one a fix date or a deliberate decision to live with it.
The reminders that never went out
On September 29, Rehan's cash forecast came up $41,300 short, three days before payroll. Rehan runs finance operations at a 40-person agency, and the gap traced back to one client. Rafferty & O'Neill had missed its monthly payment run because nobody had sent the reminder.
He opened the automated reminder log. The job had skipped Rafferty & O'Neill, along with two smaller accounts, and all three names shared one thing: an apostrophe.
The job had choked on apostrophes since March, and everyone on the team knew it. Until September, it had never cost them a cent, because Sofia had sent those three reminders by hand every month-end. This month she was on leave.
Six weeks earlier
Walk it backwards. At the end of August, Sofia ran her usual month-end routine: export the overdue list, find the three accounts the job had skipped, and send each reminder from her own inbox. It took about 40 minutes, and she'd done it every month since March.
The actual fix sat in the engineering backlog, labelled low priority. Dan, the developer who looks after billing, had estimated it at half a day. Newer and louder requests kept landing above it, so it slid a little further down the board every sprint.
Here's the arithmetic nobody ran. Six month-ends at 40 minutes each comes to four hours, which is the half-day fix. By August, the team had already paid for the repair once and still owned the bug.
Nobody ran it because the two numbers lived in different places. Dan's estimate was on the engineering board, and Sofia's 40 minutes were in her own calendar under "month-end." Dan never knew how often someone was patching the problem by hand, and Sofia never saw the estimate. So each of them could reasonably believe it was a small problem.
The signal everyone saw and ignored
The problem was written down in at least three places: the March retro notes, the backlog ticket, and Sofia's handover doc. Nobody hid it. It stayed alive because a capable person absorbed it every month, and a problem somebody absorbs never looks urgent to anyone else.
Anita Tucker and Amy Edmondson of Harvard Business School found the same pattern in their study of hospital process failures, published in 2003. Across 239 hours of watching 26 nurses in nine hospitals, they saw nurses handle 93 percent of problems by doing just enough to keep caring for the patient, and nothing to stop the problem coming back. One nurse ran out of clean linen overnight, so she walked to another unit and took theirs.
The workarounds felt good. Seven of the nine nurses they asked said they felt gratified when they got around an obstacle, and one told them, "Working around problems is just part of my job." The cost sat quietly in the hours: about 33 minutes lost in every eight-hour stint on the ward.
The researchers noticed something else. Nurses rarely told the department that caused a problem that it had happened, so the people who could fix it never learned it needed fixing. The patch worked well enough that nobody upstream ever saw the problem.
Sofia was the nurse with the linen.
The moment it was still fixable
There were two cheap moments. The first came at the end of April, the second time Sofia sent those reminders by hand. Any workaround somebody does twice has become a process, and processes belong on your list with an owner's name on them.
The second came on August 14, when Sofia booked her leave. Her handover doc said "Send apostrophe reminders manually," and nobody who read it asked why a person was doing a machine's job. That line was the last free warning. After it, the problem could only show up as a cash gap.
What you'd do differently
Dan fixed the job in three hours on October 1. Rehan then did something more useful. He went looking for every other problem on his team that somebody was quietly absorbing, and he found six.
Two were worth fixing that month: a bank export that someone reformatted every Friday, and a supplier portal that locked people out every 30 days. Three got fix dates later in the quarter. The sixth, a clunky expense form, cost about ten minutes a month, so Rehan wrote down that they'd live with it and put a review date in January.
You can run the same review on your own list this week in about twenty minutes:
- Pull every known issue older than 90 days from your backlog, retro notes, and handover docs.
- For each one, find out who's working around it and roughly how many minutes a month that costs them.
- Compare those minutes with the fix estimate. If the workaround has already cost more, move the fix to the top.
- Give every remaining item one of two outcomes: an owner with a fix date, or a written decision to live with it.
Managers skip the second outcome, and it matters as much as the first. Killing a problem deliberately means writing down what it costs, who absorbs it, and the date you'll look at it again. Sofia's month-end chore would have read very differently with her name and 40 minutes a month written next to it.
One more habit costs nothing. At every handover, ask, "What do you do by hand that the system should be doing?"
Your version of this
You've got an apostrophe bug somewhere. It might be a report someone reformats every Monday, a supplier who always needs chasing twice, or a login that only works in one browser.
A capable person on your team is absorbing it every week, which is exactly why it has never reached you as a problem. They probably don't even think of it as a problem anymore. By now it's just part of their Friday.
Find that person this week; before their next week off, find the problem for you.
