The Forty-Tab Workbook

The contract gets signed in January, and the kickoff call happens the week after. The vendor's implementation specialist is friendly and organized, and she sends over the thing her whole project plan hangs on: a spreadsheet workbook. Products on one tab. Customers on another. Rate tables, truck list, job types, fee schedules, user roles, approval workflows. Forty tabs, give or take, each one required before configuration can start. Go-live is penciled for April.

Somewhere in your county there's an owner who got that workbook, scrolled the tabs for ten minutes, and quietly decided not to buy the software. For twenty years the industry called that guy a late adopter. Look closer at what the workbook was actually asking for, and he starts to look like the one who read the fine print.

A complete description of a business that doesn't have one

The workbook asks your operation to enumerate itself. In advance. In full. In February.

Some tabs are easy. You know your trucks. You mostly know your products, though the same stone is 57s on the scale, 3/4 clean on the rate sheet, and "road rock" to half your customers, and the workbook wants one answer. Then come the tabs that want your pricing exceptions, your fee logic, your workflows, and here the trouble starts, because that knowledge doesn't live in a list anywhere. It lives in your billing clerk, who knows the orchard account pays last year's rate until the handshake renewal in March. It lives in dispatch, who knows which two customers get the fuel surcharge waived and roughly why.

The real spec of a quarry or a hauling outfit is its exceptions, and exceptions don't announce themselves in a conference room. They arrive on trucks, in season. The February version of you cannot list what the August version of you will run into: the township job that pays by the yard off a tonnage scale, or the load split across two jobs at the driveway. So the tabs get filled with best guesses, because the project plan needs them filled, and the guesses get poured into configuration like concrete.

What the guesses cost

April arrives and the configuration isn't done, because configuration is where guesses get discovered one at a time. Now the go-live has a problem the project plan never mentions: there is no good day to cut over a billing system between May and November. Nobody retrains a scale house in the busy season. So go-live slips to winter, which means the software you bought in January runs its first real ticket eleven months later, and the subscription ran all year.

When the real season finally hits the guessed configuration, the breaks come fast. The system has no way to hold the township's yard-priced tonnage job, so that account gets billed from a side spreadsheet. Each workaround moves a little more of the real operation back onto paper, and each fix is a support ticket or a professional services invoice. A year in, the system holds the clean accounts, the paper holds the strange ones, and the strange ones were the reason you wanted help.

Here's the claim your software-minded friends will fight: an implementation measured in months is not evidence that your operation is complicated. It's evidence the software can't read. Any system that needs a complete description of your business before it can process one ticket has admitted the paper means nothing to it, and the workbook is that admission, formatted nicely. The owner who refused to fill it out had simply priced the project right: the sticker said $500 a month, and the real cost was his office manager's spring, his clerk's guesses hardened into config, and a cutover window his calendar didn't have.

The spec already exists. It's the tickets.

The workbook was trying to extract, the hard way, something that already exists. Your operation describes itself every day, one ticket at a time, and has for decades. The rate the orchard account actually pays is on last March's tickets. The fee you actually charge for winter conditioning is on the tickets from January. The three names your best product goes by are all sitting in the same drawer, in ink. A year of ticket books is the complete, accurate, exception-rich description of your business that no February conference room could produce.

The old software couldn't use it. AI-era software reads it as written, and that flips the order of the whole project.

flowchart TD
  A["Day one"] --> B{"What does the software need first?"}
  B -- "a description of your operation" --> C["Workbook: products, rates, workflows"]
  C --> D["Guesses keyed in February"]
  D --> E["Season hits, exceptions break the guesses"]
  E -- "back to the tabs" --> C
  B -- "yesterday's tickets" --> F["Paper read as it was written"]
  F --> G["Rules you add when a real ticket asks"]

Day one stops being a workbook and becomes a stack of yesterday's tickets. That's how Termco starts: it digitizes the tickets your operation already produces, exactly as the scale house and the drivers already write them. The reference data and the billing rules you build from what that paper actually says, one real decision at a time. The aliases on the tickets get reconciled into your customer list once, by you, and a rate question gets settled when an actual ticket raises it, not guessed at in advance to hit a milestone. A ticket that breaks a rule you've set gets flagged that day, and invoice totals assemble from the tickets that cleared. Nobody at the scale changes anything. The drivers never see an app, and the carbon books stay where they are.

There's no cutover to schedule, because nothing about the operation cuts over. That difference, software conforming to the work instead of the work conforming to the software, is the whole argument of the last twenty years settled in your favor, if you held out.

Dig out the binder

If you bought one of these systems, the test is in a drawer: find the old implementation workbook or the configuration binder and read your own answers from that February. Count how many survived one full season. Every answer that didn't is a guess you paid to install and paid again to fix.

And if you held out, your test costs a stamp's worth of effort: pull 10-20 tickets straight out of the book, as they are, and send them to hi@termco.ai. Findings come back within 48 hours. No tabs to fill out first, which is the point.