Below: the four tests with illustrative numbers, what to do when one fails, and how to run the check in about a week.
Why the baseline matters more than the split
Most owners spend their planning time on the split: 50/50, 60/40, a bigger slice for the crew lead. That's the easy part. (The split is covered in Should I pay my crew hourly or by the job?) Every bonus dollar is the gap between the budget and actual hours, so a budget that's off by 10% moves payouts far more than a split that's off by 10 points.
"Baseline" means two different things here, and you need both:
- The job budgets crews get measured against.
- The company starting point: labor cost as a share of revenue, and revenue per labor hour, for the period before launch. This is what tells you, a year later, whether the program worked.
The first decides whether crews believe the program. The second decides whether you do.
The four tests, with worked numbers
The numbers below are illustrative, not customer data.
Say your install and enhancement division finished 200 jobs in the last twelve months. The budgets on those jobs added up to 10,000 hours. Crews actually logged 10,600 hours against them. Loaded labor rate for the example: $24/hour. Division revenue for the year: $1.2M.
At the top line, 106% of budget doesn't look terrible. That's the trap.
Test 1: The win rate
Count how many jobs came in under budget. In the example, 46 of 200, or 23%.
That's your win rate before any incentive exists, and it's how the budgets feel from the truck. My rule of thumb, not a law: a fair budget is one a well-run crew beats on somewhere between a third and two-thirds of jobs before you pay anyone extra. Below that, crews decide early the program is rigged. Above it, bonus money goes to savings that were already happening.
At 23%, this division is in the danger zone. You don't know why yet.
Test 2: The spread by job type
Break the same twelve months out by type of work:
| Job type | Jobs | Budget hours | Actual hours | % of budget | Beat budget |
|---|---|---|---|---|---|
| Hardscape | 60 | 4,000 | 3,800 | 95% | 33 (55%) |
| Plantings & enhancements | 90 | 3,000 | 3,540 | 118% | 9 (10%) |
| Cleanups | 50 | 3,000 | 3,260 | 109% | 4 (8%) |
| Total | 200 | 10,000 | 10,600 | 106% | 46 (23%) |
Hardscape budgets look honest: crews beat them on more than half the jobs. Plantings and cleanups almost never win. Launch on these budgets and your hardscape crews love the program while everyone else stops believing by the second pay period. Do the same cut by branch and by estimator. One estimator who always bids tight shows up here and nowhere else.
Test 3: Clean time coding
Before you blame the budgets, check the hours.
In the example, you dig into the plantings jobs and find 340 hours of loading at the yard and nursery runs that crews clocked straight to job tickets. In cleanups, 160 hours of drive time between properties landed on the jobs too.
That's paid time, not job time. Move it to company time and rerun:
- Plantings: 3,540 − 340 = 3,200 actual against 3,000 budgeted, or 107%
- Cleanups: 3,260 − 160 = 3,100 actual against 3,000 budgeted, or 103%
- Whole division: 10,100 job hours against 10,000, or 101%
Cleanups were mostly a coding problem. Plantings still run 7% over, so that budget needs a real fix, and you found it before a crew paid for it. Whatever you decide counts as company time, decide it once and apply it to every crew. Changing the rule mid-season is what starts fights.
Test 4: Lock the company baseline
Now write down the number the whole program will be judged against. In the example:
- Total paid labor: 10,600 hours × $24 = $254,400
- Labor cost as a share of revenue: $254,400 ÷ $1.2M = 21.2%
- Revenue per labor hour: $1.2M ÷ 10,600 = $113.21
Moving 500 hours to company time didn't change this at all. You still paid for them. Clean coding makes job budgets fair; it doesn't make the company look better on paper.
Date these numbers and agree on them before launch. If you're seasonal, keep them by month so you compare this June to last June. Without a locked baseline, every review becomes an argument about what "before" looked like.
The other direction: the padded budget test
Tight budgets kill morale fast. (That spiral is in Why do most crew bonus programs fail after 90 days?) Padded budgets are quieter, and they cost real money.
Suppose your hardscape crews had instead logged 3,200 hours against the 4,000-hour budget, beating it on 57 of 60 jobs: 80% of budget, a 95% win rate.
Launch a 50/50 split on that budget and, before anyone changes a habit, it pays:
- 800 "saved" hours × $24 = $19,200
- Crew share at 50%: $9,600 a year for savings you already had
That's not a self-funding bonus. It's a raise with extra paperwork. The fix is to reset from your own history, but not all the way to 3,200. That takes away slack crews earned with good habits, and they'll notice. Aim for a budget a good crew can still beat with effort, which brings you back to Test 1.
When the honest budget doesn't match the price
Sometimes the honest budget is higher than the labor you priced. In the example, plantings run 7% over even after the time is cleaned up. Keep the estimate's number and you're asking crews to fix a pricing problem with their effort.
Don't hide a pricing problem inside the bonus. Set the bonus budget at what a good crew can actually do, and fix the estimating template going forward. Your margin on those jobs was already lower than you thought. The check just made it visible.
Failure modes: how baselines go wrong
Checking the total, not the spread. 101% across a division can hide one job type at 120% and another at 85%.
Trusting the bid over history. The bid is what you hoped. The history is what happens. Test one against the other before anyone gets paid.
Recurring contracts that drift. Customers add beds and drop services, and the route budget still reflects last year's visits. Recheck when contracts change. (Route budgets are covered in Should I bonus maintenance routes stop-by-stop or by the week?)
Changing budgets quietly after launch. Adjust on a schedule, with a reason, and keep a record. A budget that moves mid-job after the crew started winning destroys trust faster than a tight one.
Skipping the baseline because the numbers are messy. Messy data is the reason to do this. If you can't trust the hours, you can't trust the bonus. (Clean gates on quality matter just as much; see Won't performance pay make my crews rush the work?)
How to run the baseline check in about a week
- Pull twelve months of finished jobs. Budget, actual hours, job type, branch, and estimator. Skip open jobs.
- Calculate the win rate and the spread. Flag anything that almost never or almost always wins.
- Audit the time on flagged jobs. Find drive, loading, and shop time on job tickets. Write the company-time rule once and recode history the same way.
- Fix budgets, not crews. Reset flagged job types from history. Where the honest budget is above the price, fix the estimate too.
- Rerun the would-have-paid math. What would each crew have earned last year? If it's "nothing" for half and "a lot for doing nothing" for the rest, you're not done.
- Lock the company baseline. Labor as a share of revenue and revenue per labor hour, by month if seasonal.
- Plan the first review. Ninety days in, compare against the baseline and tune outlier budgets.
What it looks like when it works
Promier Landscapes (Kingsport, TN; 35 field staff; numbers verified by an outside landscaping CFO at B Advisory) ran a year on savings-share performance pay with the same crew. Labor cost as a share of revenue fell from 21.8% to 19.9% while wages went up, and revenue grew 17.5% ($7.5M to $8.8M). That came to about $165,000 in field labor savings, and 4,877 hours moved back to billable work. Those numbers mean something because there was a measured starting point.
Mulhall's Landscaping (Omaha; 30+ years in business). Mick Mulhall: "Protiv has been the most impactful thing we've ever done to drive operational performance. We saw the impact in one month."
Those are landscaping results. Other trades, like HVAC, plumbing, and electrical, can set up the same baseline check on their own job history, but these case studies are proof points, not a promise.
Where Protiv fits
You can do this check in a spreadsheet. The hard part is keeping it honest after launch.
Protiv starts with a pre-launch analysis of your last twelve months. It pulls in jobs, budgets, and hours from the systems you already use and flags budgets crews would never win and recurring contract hours that drifted. Drive time, training, and shop work can be logged as company time, which doesn't count toward a job budget or a bonus. After launch, managers adjust budgets from the budget vs. actual view, with a budget history of every change, and a quarterly business review checks results against the pre-launch baseline. The framework is yours. Protiv keeps the numbers honest every week.
FAQ
How many months of history do I need for an honest baseline?
Twelve months is the standard, because it covers a full season and evens out the weird weeks. If you're seasonal, treat it as the minimum, since a spring-only sample makes every summer budget look wrong. With less history, use what you have, expect to tune budgets more often in the first ninety days, and tell crews up front that the baseline is a starting point, not a final answer.
Should my bonus budget come from the estimate or from past actual hours?
Start from the estimate, because that's what you priced. Then test it against past actual hours for the same type of work. Where they match, use the estimate. Where they don't, figure out which one is wrong: sometimes the estimate is light and needs fixing for future bids, and sometimes the history is inflated by bad time coding. The bonus budget should be what a good crew can beat with real effort.
What if my time data is too messy to trust?
Fix the time data first, at least for the job types you plan to launch with. Sample recent jobs, ask crew leads where the hours went, and look for drive time on tickets, yard loading, and shop days coded to whatever job was open. Write one company-time rule and apply it going forward. You don't need perfect history to launch, but you need clean current data, because every bonus pays on it.
Won't crews think I'm rigging the budgets against them?
Some will at first, which is why you share the method, not just the numbers. Tell them budgets come from the company's own history, show which job types got looser after the check, and explain how budgets get reviewed. The first time a crew lead flags a bad budget and you fix it on schedule, you've done more for trust than any kickoff speech.
How often should I reset budgets after launch?
Review on a fixed schedule, usually quarterly, and in between only when something real changes, like a contract change or a new estimating template. Job types that never win and ones that always win by a mile both have budget problems. Don't tighten a budget just because a crew got good at it. That's the improvement you were paying for, and taking it away teaches crews to slow down before reviews.
Soft next step
Before you announce anything to the field, pull last year's finished jobs and run the four tests. It's a week of work, and it's the difference between a program crews believe and one they write off after the first paycheck.
When you want to run that check on your own data, talk to the Protiv team.