Split by the Alphabet

The second clerk started in March, because last April nearly buried the first one. Eleven hundred tickets a month is past what one person can key, price, and invoice while also answering the phone, so you hired, and the split was obvious and fair: Donna keeps accounts A through K, the new hire takes L through Z.

Six months later you're running two billing policies. The dividing line is the alphabet.

The rules were never written

Nobody changed a rate. Your rate sheet says what it said in February. The problem is everything the rate sheet doesn't say.

It says "minimum haul charge: $85." It doesn't say whether the fuel surcharge calculates before the minimum or after it. It doesn't say whether the second leg of a split delivery is its own load for minimum purposes. It doesn't say what happens on the handful of old-line accounts that haven't been charged a minimum since 2014, and whether that's an exception or an oversight.

Donna knows. She learned it in 2011 from the woman who trained her, kept what made sense, and has applied her version so long it doesn't feel like a version. It feels like how billing works. The new hire learned by asking Donna, in the gaps between Donna's own deadlines, and what she didn't know to ask about she settled herself, reasonably, one ticket at a time.

Walk the ambiguities and watch them each get settled twice. The scale prints 9.38 tons: bill 9.38, or round to the half ton the way the old system used to? Winter rates start November 1, or when the plant actually switches mixes? Does "contractor rate" in the corner of a ticket mean the account's quoted rate, or the flat discount the counter gives anyone with a dump trailer? Each question has two live answers in your building right now, and which one a customer gets depends on the first letter of his company's name.

What the fork costs

Take one rule and run the arithmetic. Donna applies the $85 minimum to every load under 10 tons. The new hire waives it wherever the quote is silent, which is most places, because she was never told otherwise and waiving felt safer than surprising a customer. Say L through Z produces 45 small loads a month and the average gap between the minimum and the actual charge is $22. That's roughly $1,000 a month leaking out of one unwritten rule, on one half of the alphabet. Stack the rounding question and the winter-rate question on top. Whatever the total comes to, nobody decided any of it.

It stays invisible because each desk is internally consistent. Donna's accounts always get Donna's policy, so no customer of hers ever sees a change worth calling about. Audit either desk alone and it passes.

Then it surfaces, the way these things do, sideways. Picture the customer who runs two entities, Hargrove Construction and Lakeside Topsoil, same owner, same trucks, same kitchen table where both stacks of invoices land. H goes to Donna's desk, L goes to the new hire's. One night he puts two invoices side by side and finds an $85 minimum on an 8.2-ton load billed to Hargrove, and no minimum on the same 8.2 tons billed to Lakeside nine days earlier. He's not angry. He just wants to know which one is your policy, and you already know which answer he's rooting for, applied to both companies, going back a while.

Disputes fork the same way, because the phone doesn't respect the alphabet. When a customer calls about a minimum charge, whichever clerk picks up settles it, by her own rules. So an L-account dispute that lands on Donna's ear gets held firm, and an identical H-account dispute that reaches the new hire gets credited. The company's position on its own fee now depends on who was nearer the phone, and every one of those credits quietly teaches a customer which desk to ask for.

flowchart TD
  A["Rate sheet line: minimum haul charge $85"] --> B["What it doesn't say: applied when the quote is silent?"]
  B --> C{"Who keys the ticket?"}
  C -- "Donna, A-K" --> D["8.2 tons bills at $85"]
  C -- "new hire, L-Z" --> E["8.2 tons bills at $63"]
  D --> F["One rule on paper, two in practice"]
  E --> F

The binder won't save you

Here's the claim your office manager will fight: writing the procedures down fixes almost none of this. Most shops that spot the fork respond with a binder, or a Monday meeting, or both, and the fork survives.

A binder answers questions. The leak lives on tickets where the clerk doesn't know there's a question. Habit doesn't present as doubt; it presents as knowledge, fluently, at keying speed, a full tray before lunch. Nobody flips to tab four to confirm something she isn't aware she's assuming. And the binder goes stale the first time a rate sheet changes without it, after which consulting it is optional and everyone knows it.

"We talk constantly, we're consistent" is a testable belief, and the test is at the bottom of this page. Before you run it, be honest about the stakes: if two trained, well-meaning clerks can hold the same job in the same office and quietly diverge, then the thing you've been calling a billing policy is an oral tradition. It has as many versions as it has speakers.

Put the rule where the pricing happens

A rule only stays one rule when it lives in a structure, with its edge cases attached: minimum $85, computed after surcharge, waived on the six accounts flagged as legacy, effective these dates. Getting it there forces the argument you've been skipping, once, out loud, with whoever owns the price in the room, and the answer becomes an entry anyone can read. None of this changes who your clerks are. It changes what they key against.

That's the office Termco builds. Tickets get digitized as your drivers and operators already write them, carbon smudges and corner notes included. Billing rules and reference data live in one structured place, fine print and all. Once the minimum's edge cases are on the books, a ticket that skips one, or prices against a dead rate, gets flagged the day it's keyed. Invoice totals assemble from the tickets that cleared. Nothing reaches the trucks or the scale window. The books and pens stay, and the drivers install nothing. The alphabet goes back to being a workload split instead of a policy border.

Give both desks the same twenty tickets

The test costs an afternoon. Photocopy twenty real tickets: small loads, a split delivery, a late-November date, a couple of the old legacy accounts. Hand a set to each clerk and have them price every ticket independently, no conferring. Compare line by line.

Every disagreement is a rule your company doesn't have. Multiply each gap by a month's worth of similar tickets, and that's the cost of leaving it unwritten. If you'd like a third reader, send 10-20 sample tickets to hi@termco.ai and findings come back within 48 hours.

Donna isn't wrong, and the new hire isn't wrong either. Each of them is applying your billing policy exactly as she received it. The trouble is that the company never had one. It had Donna's, and now it has two.