The overtime that billed itself at straight time for 14 months

Meridian Workforce is a composite of the light-industrial agencies we built Lumitide for: about 240 temps across eleven warehouse and food-production clients, weekly invoicing out of TempWorks, payroll through a bureau, and margin tracked in a spreadsheet that was accurate the day it was built.

Segment
Warehouse & food production
Workers on assignment
≈240 weekly
Annual billing
≈$11M
Back office
TempWorks, QuickBooks, Excel, 3 people

The setup

Meridian’s book looked healthy at the blended level: gross margin in the high teens, normal for high-volume light industrial with a couple of program accounts. The weekly close was a three-person relay — timesheets from client portals and paper into TempWorks, payroll out Thursday, invoices out Friday, and a margin spreadsheet updated “when there was time.”

Payroll was reliable, because workers call immediately when a check is short. Billing had no equivalent alarm. That asymmetry is where every leak in this story lived.

What the first reconciliation found

Running billed-versus-paid line by line for a single week surfaced three patterns, all invisible on the invoices themselves:

Overtime billed at straight time — two accounts

Both client contracts allowed OT to bill at 1.5×. When the accounts were onboarded, the OT bill-rate multiplier never made it into the billing setup — so pay ran at time-and-a-half while billing ran straight. Across roughly 45–50 OT hours a week between the two sites (peaking much higher in Q4), the unbilled premium averaged about $2,300 a month. It had been running for 14 months — call it $30,000 — and every invoice had looked routine, which is exactly why nobody caught it. Industry taxonomies of timesheet errors list incorrect OT calculation among the most common and most expensive failure modes, and this is the shape it usually takes.

A shift differential paid but never billed

A $0.75/hour third-shift differential negotiated with one client the previous summer was set up in payroll the same week — and never added to billing. Roughly 700 third-shift hours a month meant about $800 a month in differential-plus-markup that Meridian simply absorbed.

Two workers paid with no billing counterpart

A backfill added mid-week by a site supervisor, captured on a paper timesheet that reached payroll but not billing — twice, at two different sites, roughly $600 in wages plus burden with no invoice line anywhere. This is the classic “paid-but-never-billed” leak: payroll completes because people complain when they aren’t paid; nobody complains about a missing invoice line.

The fix, week by week

  • Week 1: full billed-vs-paid reconciliation of the current week; the three patterns above surface immediately.
  • Week 2: OT multipliers corrected in billing setup for both accounts; differential added; both verified against the next invoice run.
  • Weeks 3–4: lookback quantified. Meridian chose to invoice the most recent quarter of unbilled OT (about $6,900) and wrote off the rest to protect two anchor relationships — a judgment call, and probably the right one.
  • Week 5: reconciliation becomes part of the Friday close: every payroll line must match a billing line at the contracted rate before invoices go out, with unmatched lines held in a review queue instead of shipped and forgotten.

Where it landed

The recurring fix was worth about $3,700 a month — roughly 0.4 points of gross margin on an $11M book, or the margin contribution of adding a dozen placements, for zero sales effort. The one-time recovery covered the cost of the entire exercise several times over. And the Friday close now produces a short exception list instead of a shrug: SIA’s revenue-leakage estimate of 5%+ of billable revenue stops being an abstraction the first time you watch two accounts quietly under-bill for fourteen months.

$3,700
recurring monthly leak found
+0.4 pts
gross margin recovered
5 weeks
from first reconcile to fixed

Composite case study. Details are drawn from documented industry failure modes and pilot experience; the agency is fictional and the numbers are kept deliberately conservative.

Wonder what your own numbers look like?

Reconcile a week of your own — or join the list for the next case study.