Instinct Said Raise Everything: Two Years of Job Sheets Said Something Far More Specific

The business was busy, the crew was fully booked through the season, and the year-end figures showed a thinner margin than the previous year on higher revenue. The owner's instinct, which is almost everybody's instinct in that position, was that prices had fallen behind across the board and needed a general increase of some percentage nobody could quite justify. Before doing that, they spent an afternoon reading two years of completed work orders. What the paperwork said was far more specific than the instinct had been, and it produced an entirely different set of changes.

What the Job Sheet Was Already Recording

Every completed job carried a work order with the quoted price, the date booked, the date completed, hours logged by each person, materials used with supplier costs, and a notes field for anything unusual including return visits. None of that was designed as a management system and none of it had been set up with analysis in mind. It existed because somebody years earlier had decided the crew should write down what they did, and it had been filled in with reasonable diligence ever since.

Every one of those fields turned out to matter, and one of them had never once been looked at after the job closed. That field was hours on fixed-price work, and its emptiness was the reason the problem had stayed invisible through two years of monthly reviews that focused entirely on revenue.

Hours Went Unrecorded on Exactly the Jobs That Needed Them

Large jobs had hours logged carefully, because those jobs were billed against them and everybody understood the number would be checked. Small fixed-price jobs had hours logged loosely or not at all, because nobody was billing against the figure and there was no obvious consequence to leaving it blank.

Reconstructing those hours from the crew's daily schedules produced a very different picture from the one on the sheets. Recorded hours on small jobs were routinely lower than the elapsed time between the first and last entry for that job, and the gap widened considerably on anything requiring a return visit. That gap is where the margin had been going for two years, and it had been invisible for the simplest possible reason: the one field that would have shown it was the one field nobody filled in.

The Travel Was the Job

The work orders carried addresses and completion dates, which meant sorting jobs by day showed how many separate sites the crew visited in a working day without anybody having to remember. On days built around one large job, the answer was one site. On days made up of small jobs, it was three or four, and the driving between them appeared on no sheet anywhere.

The quoted price for a small job assumed the work itself. It did not assume a third of a crew-hour of driving, a setup and a breakdown at each stop, and it certainly did not assume two or three of those inside one day. Priced as work, the small jobs looked acceptable and always had. Priced as a portion of a day that the crew could otherwise have spent somewhere else, they did not, and the difference between those two framings was most of the missing margin.

One Category Came Back Twice

The notes field was informal and inconsistently filled, and it still recorded return visits often enough to be counted. One category of work showed return visits at a materially higher rate than anything else, and reading the notes on those specific jobs revealed a consistent pattern: the same component failing, sourced from the same supplier, under the same conditions. That was not a pricing problem in any sense. It was a materials problem that had been quietly funded out of margin for two years because nobody had ever aggregated the notes. A return visit costs far more than it appears to on a sheet, since it carries the travel again, the setup again, and a scheduling slot that could have held a paying job instead, which means a job that comes back once is frequently worse than no job at all.

Quotes That Never Followed the Supplier Invoices

Materials costs sat on the sheets alongside supplier prices, so comparing the first year against the second showed the direction of travel immediately, and comparing both against quoted prices showed that the quotes had simply not followed. The quoting was being done from a price list assembled at some earlier point and revised occasionally by feel, which meant every job in the second year silently carried the difference between an old assumption and a current invoice.

What Changed, What Did Not, and How It Was Explained

Not a general increase. Four specific changes came out of the afternoon. A minimum charge covering travel, setup and the first period on site, which addressed the small job problem directly and made the pricing honest about what a visit costs rather than pretending the work is the whole cost. Zone pricing by distance in three published bands, with the farthest band priced to reflect what a day actually looks like when a crew is out there. A different specification for the failing component, at a higher unit cost, which stopped the return visits and on its own recovered more than the general increase would have. And quotes tied to current supplier pricing, with the list reviewed quarterly and a written validity period so that an old quote stops binding indefinitely.

Large jobs, which the owner had assumed were the problem, turned out to be perfectly healthy and were left entirely alone. That is precisely the part a general increase would have got wrong, by raising prices on the work that was already competitive and profitable while leaving the actual leak untouched. The changes were announced in a short letter to the existing customer list six weeks before they took effect, carrying three points, no apology and no commentary about the economy: there is now a minimum charge for a visit and here is what it is, travel outside the immediate area is priced in three bands and here is the map, and existing quotes are honored to their stated expiry date.

Two customers asked about the minimum and both booked anyway, with one combining two small jobs into a single visit, which is exactly the behavior the minimum was designed to encourage. Nobody left over it. The reason it went smoothly is that every change corresponded to a cost the customer could see for themselves, since a minimum charge for a visit is intuitive to anybody who has ever waited half a morning for a van. A general increase invites the question of why, and the only honest answer available would have been that the shop did not know where its money was going.

What made the whole reading possible was unglamorous and mostly already in place. Four fields did all of the work: hours, materials with costs, address, and completion date. The single gap, hours on fixed-price jobs, was the one that mattered most, and closing it was a policy change rather than a software purchase, since hours now get logged on every job whether or not anybody is billing against them. The method generalizes easily enough. Pull two years of completed jobs into one sheet with a row for each, add columns for quoted price, recorded hours, materials cost and any return visit, then calculate margin per job and margin per crew hour, because the second of those reorders everything. Sort by margin per crew hour, read the bottom quartile, and the pattern that emerges is almost always a category rather than a customer. Read the notes on anything that came back and count the causes. Then change the specific thing the records name and leave everything else alone, which is the discipline that keeps the increases holding, because a customer will accept a fact about a business far more readily than a percentage applied to everything.