How an HVAC Company Fixed a Broken Bonus System and Doubled What Techs Took Home

Every incentive plan runs on one assumption technicians rarely say out loud: that the number on the paycheck is actually correct. An HVAC and mechanical services company found out how fragile that assumption is when technicians started quietly comparing notes and noticing their commission did not always match what they thought they had earned. Nobody was being shorted on purpose. The plan had a bug. But a bug in a pay calculation does not read as a bug to the person cashing the check. It reads as being cheated.

The root cause turned out to be a data problem, not a math problem. Some jobs were showing up twice in the commission calculation, the same completed work counted once in the raw job export and again in a secondary sync, quietly inflating certain technicians' payouts while making the whole system look unreliable to anyone auditing it closely. At the same time, a separate mapping issue meant emergency and holiday jobs were not always being tagged correctly, which caused technicians to miss a $25 on-call award they had actually earned. A third issue was subtler still: credit card processing fees were being deducted from technician service commissions in some cases, quietly shrinking payouts by a few dollars here and there in a way almost impossible to spot without pulling the raw numbers apart.

Small Errors, Big Trust Problem

None of these three issues individually would have been a crisis. A duplicate line item here, a missed $25 award there, a few dollars of misapplied fees, none of it adds up to a huge dollar figure in isolation. But incentive pay does not run on dollar figures alone. It runs on trust, and trust does not degrade proportionally to the size of the error. A technician who catches one discrepancy starts checking every paycheck after that. A technician who hears a coworker caught a discrepancy starts wondering about their own numbers even if nothing was ever wrong with them.

That is the real danger of small, compounding payroll bugs in an HVAC incentive plan. The dollar amount at stake might be genuinely small. The trust cost is not. Once technicians stop believing the paycheck reflects the work, the entire incentive structure stops functioning the way it was designed to, regardless of how well the underlying formula was built. A perfectly designed commission plan running on unreliable data is functionally the same as a bad plan, from the technician's point of view.

Fixing it required going past the plan design itself and down into the data feeding it. That is a different kind of problem than most incentive redesigns solve. It is not about choosing the right percentage or the right threshold. It is about making sure the numbers going into the formula are the correct numbers in the first place.

Three Fixes, One Backfill

ShareWillow addressed each issue at its source rather than patching around it. The duplicate commission line items were resolved with a deduplication fix at the data layer, so a completed job could only be counted once in a payout calculation no matter how many times it appeared across different sync feeds. The missed on-call awards were solved with proper job-type mapping, so emergency and holiday work is now correctly tagged and automatically triggers the $25 award instead of depending on someone catching the miss manually. The fee-deduction issue was corrected at the calculation logic level, so credit card processing fees stopped being subtracted from technician commissions they had rightfully earned.

Just as important as fixing the logic going forward was backfilling the pay periods that had already been affected. A fix that only applies to future paychecks does nothing for the trust already lost on past ones. Technicians who had been shorted, even by a small amount, got made whole for those specific pay periods, not just told the problem was fixed going forward.

Payout Accuracy Within About $100

The clearest measure of whether the fix actually worked is not a percentage. It is how closely ShareWillow's calculated payout now matches the company's own internal numbers when they check technician by technician. Today, most technician differences land within roughly $100 of what the company's own calculations show, a tight enough margin that discrepancies stopped being a recurring conversation. That number matters more than it might sound like it should. A $100 variance on a real paycheck, across a real pay period, with a real explanation available if anyone asks, is the difference between a system people trust and one they quietly audit every single cycle.

That accuracy did not just stop complaints. It changed what happened to the incentive payouts themselves once technicians trusted the number behind them. Weekly technician incentive payouts more than doubled, moving from an average of roughly $2,090 a week across the plan's first five confirmed weeks to roughly $4,920 a week across the eleven confirmed weeks since. That is not a case of the company suddenly paying out more generously. It is what happens once technicians trust that engaging fully with the incentive plan, taking the extra call, chasing the on-call award, closing the upsell, actually shows up correctly in their paycheck.

Put another way: the accuracy fix and the payout growth are not two separate wins. They are the same win, seen from two angles. A technician who does not trust the math has no reason to change behavior to earn more under it. A technician who does trust the math has every reason to.

Why Backfilling Mattered as Much as the Fix Itself

It would have been faster, and cheaper, to fix the calculation logic going forward and leave the past pay periods alone. Nobody outside the company would have known the difference. But technicians would have known, because they were the ones who had noticed the discrepancies in the first place. Backfilling the affected periods was the part of the fix that actually addressed the trust problem, not just the technical one.

There is a lesson in that sequencing for any company running incentive pay through a system prone to sync issues or duplicate data, which is most systems stitched together from a field service platform, a payroll system, and a spreadsheet somewhere in between. The technical fix is necessary but not sufficient. Technicians do not experience a data pipeline. They experience a paycheck. Making the paycheck right, retroactively, is what actually rebuilds the trust a bug like this quietly erodes.

Backfilling also sent a specific signal to the team that mattered beyond the dollars involved: that the company would rather absorb the cost of correcting past mistakes than let a technician quietly eat the difference. That signal is worth more than the backfilled amount itself in almost every case, because it is the thing technicians actually remember the next time a number on their paycheck looks slightly off. Instead of assuming they got shorted again, they now have a reason to ask, and a track record showing the answer will be a real fix, not a shrug.

How to Catch This Before Technicians Do

The uncomfortable part of this story is that the company did not catch the duplicate commission issue through an internal audit. Technicians caught it first, by comparing notes with each other, which is close to the worst way for a business owner to find out their payroll math has a bug in it. By the time leadership hears about a discrepancy, it has usually already been discussed by several technicians privately, which means the trust damage happened before anyone in the office even knew there was a problem to fix.

For a business owner or operations lead running incentive pay through any combination of a field service platform, a payroll system, and manual reconciliation, there is a simple test worth running before technicians run it for you: pick five random technicians from a recent pay period and manually trace their commission from the completed job record all the way to the paycheck. If that trace is fast and the numbers match cleanly, the underlying system is probably sound. If it takes real digging, or if a manager has to explain an adjustment that is not documented anywhere technicians can see, that is the same kind of quiet risk that built up here over time.

Duplicate job entries are a particularly common version of this problem because they are almost invisible from the top. A revenue report or a job-completion count will not flag a job that got counted twice in a commission calculation but only once in the main dashboard leadership actually looks at every week. The discrepancy only becomes visible at the individual technician level, which is exactly where most owners are not looking closely on a routine basis.

What This Looks Like From the Technician's Side

A technician does not see a deduplication fix or a job-type mapping correction. They see a paycheck that finally matches what they expected, week after week, without a discrepancy to flag. That consistency is what let payouts grow the way they did. Once a technician trusts that taking an extra emergency call will show up correctly as the $25 award it is supposed to be, they take the call. Once they trust a completed job will not accidentally get shorted by a misapplied processing fee, they stop mentally discounting their own commission before it even hits their account.

Automated, data-driven incentive tracking is what makes catching and fixing issues like duplicate job entries or mistagged job types realistic at scale. A company doing this by hand, cross-referencing a field service export against a payroll run every pay period, is far more likely to miss the kind of small, compounding errors that quietly built up here in the first place.

What Other Companies Running Incentive Pay Should Check

If your technicians have ever mentioned, even in passing, that their commission looked a little off one week, it is worth treating that as a data problem to investigate rather than a one-off mistake to smooth over. Small discrepancies rarely stay isolated. They tend to be a symptom of a sync or mapping issue that is quietly affecting more paychecks than the one someone happened to notice.

The story of the multi-trade company that rebuilt eighteen separate pay plans into one connected system, covered here, faced a related but different challenge: not broken data, but disconnected systems that made it impossible to see incentive spend clearly in the first place. Both problems point at the same underlying fix. Pay technicians off of one accurate, connected source of truth, and both the trust and the results tend to follow.

Accuracy is not the exciting part of an incentive plan. Nobody gets a case study written about how correct their payroll math is. But it is very often the actual bottleneck standing between a well-designed plan and a plan that changes behavior. Here, the fix was not a new formula. It was making sure the existing formula was finally being fed the right numbers, and making the technicians whole for the periods where it had not been.