The spreadsheet you keep rebuilding

You already built the sheet. That was never the problem.

Most of us already have the spreadsheet. We built it, and it works.

Then the department buys a new piece of tech, and you lose a week reworking the whole thing so the new export fits into what you already had.

Or you want to add one layer. A rolling average, a flagging system, a view your staff can read at a glance. Sounds like twenty minutes. It's three afternoons.

Or you never get to it, and you end up where a lot of us end up. You have four or five sheets, one per system, clicking between them to see one athlete's full week.

That's the tax nobody schedules. Not the collecting, not the reading. It's building and rebuilding the thing that holds your data, every time something changes.

It's a build job, not a data job

Every export looks a little different, so integrating a new one always feels like starting over.

It isn't. Your GPS platform isn't going to change its export format next week. Same headers, same layout, same quirks, every time. You're not solving a new problem. You're redoing the same one by hand.

That's a systems problem, same as the last two issues.

Build the bridge once.

You tell the AI what the export looks like and what your sheet needs to look like, and it hands you back a repeatable way to get from one to the other. Formulas, Power Query in Excel, a script in Sheets, whatever fits how you work.

You build it one time. After that, a new export just drops in and you hit refresh, and the week you used to lose is gone.

And you can do all of this without putting a single athlete's data into the AI. You're describing headers and structure, not handing over anybody's numbers. Send the shape, not the sheet. Your files never leave your drive.

Inside the Lab: how to describe the bridge

It only needs four things from you.

• What the export looks like. Headers in order, and what one row is. "One row per athlete per session, name as Last comma First, date as DD/MM/YYYY, then the metrics."
• What your sheet looks like. Give it the same detail for the finished sheet your staff reads: the columns you want, in the order you want them, plus any graphs or charts you want built off that data and what each one should show.
• What you fix by hand every time. This is the one that matters most. Names that need flipping, dates that need converting, summary rows you delete, the athlete who shows up three different ways because he got entered wrong back in September.
• What you're building it in. Excel or Sheets. Say it every time, and have it explain the steps back so you can fix the thing yourself in February.

Give it those four things and have it build the sheet. If describing your export and your old layout from scratch sounds like a pain, you don't have to. Pull a few rows into a test file, delete the player names, and upload that instead so it can see exactly what you're working with. Blank out the names and you can hand over the real structure without handing over a single athlete.

Once it's built, ask for ten rows of fake data in your export's format and test on that first. Break it somewhere safe before it gets near a real file.

What it looks like in the real world

New system lands. Its export doesn't match the sheet you've run for two years, and normally that's a week gone.

Instead you describe both and tell it:

"Excel. Give me a repeatable way to pull that export into this sheet. Flip the name format, convert the dates, drop the summary rows, and put any athlete it can't match on a separate list instead of guessing. Walk me through the steps so I can rebuild it if it breaks, and give me ten fake rows to test on first."

That last part, telling it to set aside any athlete it can't match instead of guessing, is the piece people leave out, and it's the one that causes real damage. If it matches on a close-enough name, it'll hang one athlete's numbers on another kid's row, and nothing about the sheet will look wrong. A few unmatched rows take two minutes to sort out. A wrong row you never catch can quietly steer a whole block of training.

Where the coach stays in charge

Check it by hand the first few times. Take two athletes and follow their numbers all the way through, raw export to finished sheet, and make sure they land where they should. Check it again after any software update, because when a vendor changes their export, your bridge tends to break quietly instead of throwing an error.

And the bridge doesn't tell you anything. It hands back the afternoons you were losing to copy and paste. Reading a clean sheet is still the job, and reading it was never the part slowing you down.

The Coach's Scouting Report

• Signal. You're not bad at spreadsheets. You're doing a build job by hand that only needed building once.
• Noise. A new sheet for every new system. Letting a tool guess on the athletes it can't match.
• Next Rep. Grab the export that never fits anywhere. Describe its columns, describe where you want it to land, list what you clean up by hand every time, and ask for a repeatable way to connect them. Test it on fake rows first.

See you next Sunday.