Finally, I have properly got my hands into the operations of an enterprise client. Operations is the part of this business I actually enjoy, building, fixing, making something run smoother than it did the week before. It's the domain I'm good in, and the work I like doing most.
When I start with a client, the first job is never the tool. It's finding out what's actually broken, and that only comes from asking the right questions out loud, then not stopping at the first answer.
The questions I always start with
What's frustrating you right now, this week, not in general. How much time does this actual task cost you, in hours, not a vague sense of "a lot." If a magic wand fixed this overnight, what would you be looking at instead of what you're looking at now.
Then, whatever they answer, I keep asking why. Toyota's production engineers called it the Five Whys: ask why five times and you usually land on the real problem, not the first one anyone mentions. On this one it went something like: the board report is late every quarter. Why. Because I open every card by hand. Why. Because the list a card sits in doesn't match the status label on it. Why. Because whoever moves the card to Done doesn't always flip the label. Why. Because nothing checks the two against each other. Four whys in, and the fix isn't a faster report. It's something that reconciles the two instead of trusting either one.
The weak point in Trello reporting
This time the frustration was Trello, or more precisely, everything built on top of Trello. The boards themselves are fine, the team keeps them reasonably current, so swapping the tool would cost months and buy a new set of problems for no reason. The problem is getting a clean report out the other end.
I looked at the usual reporting add-ons, Blue Cat Reports among them, and they're all doing the same thing under the hood: reading the list a card sits in and calling that the status. None of them catch a card sitting in the Completed list with a status label that still reads In Progress, because nothing in the add-on checks the two against each other. An add-on like that produces a number. It doesn't produce a number you'd stand behind in front of a board.
What the COO was actually building by hand
The client's COO had been building the real board report by hand every quarter: pulling cards off eight boards, cross-checking each one against a status field, deciding by eye which projects genuinely mattered to the founder and which one just had a loud label on it. Hours of it, every quarter.
That's not right, and it needed fixing fast.
My first version read every board straight through and handed back three hundred and nineteen lines flagged as needing attention. Nobody is opening a document that long before a board meeting, and I wouldn't have either. So the fix wasn't a cleverer list, it was scoping the read to the actual job: only the current quarter's completed items, checked list against label, and every disagreement between the two listed by name instead of silently picked one way or the other. Twenty-six lines. Every one defensible.
The prototype
What he actually wants isn't a dashboard. He doesn't want to log in and poke around. He wants to press a button: pick the quarter, pick which boards, get the PDF and the matching Excel workbook back, on his own, without pinging me first.
So that's the build. One page. A quarter dropdown. A checkbox per board. A run button. Two downloads. All the actual work, the reconciliation, the completion counting, the business-critical flagging, stays underneath in code that gets tested, not scattered across a UI. The front end just collects three choices and runs it, which is usually the part people over-build.
This is the same discovery process behind the AI financial controller work I've written about before, and the same instinct behind turning a forty-page board pack into one page: ask what the person is actually doing with their time, not what tool they're currently doing it in. The tool is rarely the problem. The reconciliation nobody built is.
Ninety-nine days into doing this out loud, the pattern holds. Every one of these engagements starts the same way: not a pitch, just questions, and then enough whys to find the thing actually costing someone their Tuesday.
Monthly Revenues $11,000 | Clients 2 | Prospects 1 warm lead | Team: Me with Jan on standby
Day 99 of 365.