Skip to main content
Back to Blog
Comparisons

Data: How Piece Rate Crews Outperform Hourly Crews

What the productivity research actually shows about piece rate vs hourly pay, what it does not show, and how to read the numbers without fooling yourself.

Tyson Faulkner·May 5, 2026·12 min read

Output-Based Pay Has Real Numbers Behind It

Most arguments about piece rate vs hourly pay get made with anecdotes. "My buddy switched and his crew is killing it." "I tried it once and it was a disaster." Both stories are real, and neither one settles the question.

There is actually solid research on this topic going back decades. The numbers are striking on the right kind of work, and they are weaker on the wrong kind of work. If you are thinking about moving your crew off hourly, it helps to know what the data actually says before you read it as a guarantee.

This article walks through the research, the caveats, and a worked example for a 3-person roofing crew. The numbers in the example are illustrative, not pulled from any specific customer. They are meant to show how the math behaves, not to predict your exact results.

A Quick Note on Where I Am Coming From

My background is in roofing — tear-offs, gutters, soffit and fascia, occasional siding. That is the work I ran piece rate on for years before building Piece Work Pro. The research I am summarizing here covers other industries too, including auto glass and manufacturing, so I am leaning on academic sources for the broader claims and on roofing context for the worked example.

I am not an economist. I am a contractor who reads this stuff because I want to know if my own intuition about piece rate is supported by people who do this for a living. So far, mostly yes — with caveats.

The Lazear Safelite Study: The Canonical Number

If you have ever read anything serious about piece rate, you have seen this study cited. Edward Lazear, an economist at Stanford, studied Safelite Auto Glass when the company moved its windshield installers from hourly pay to piece rate in the mid-1990s.

The headline number: productivity per worker rose roughly 44 percent after the switch.

That is a huge gain. But the more interesting finding is in the breakdown. Lazear estimated that:

  • About half the gain came from incentive effects — existing workers producing more under the new pay structure.
  • About half came from sorting effects — slower workers leaving the company and faster workers being attracted or staying.

Both effects matter. Even if your existing crew only picks up 20 percent more output, the long-run gain compounds because your hiring funnel starts pulling in people who actually want to be paid for what they produce.

Safelite also kept tight quality controls. Every install was inspected, and any callback was charged back to the original installer. That part gets ignored in most retellings, but it is the reason the productivity gain did not come at the cost of broken windshields.

What the Data Says Beyond Safelite

Lazear is the most cited paper, but it is not the only one. Studies on tree planting, fruit picking, manufacturing piece work, and call center performance bonuses all show output-based pay producing measurable gains over flat hourly. The size of the gain varies a lot:

  • Simple repeatable production tasks (windshields, fruit, simple manufacturing): often 20 to 50 percent gains.
  • More complex production work (construction trades, varied installs): typically smaller and noisier, often 10 to 25 percent.
  • Knowledge work, teamwork-heavy roles, customer service: piece rate often flat or negative because it distorts the wrong incentives.

The pattern is consistent. The cleaner the unit of output, the bigger the effect. The fuzzier the unit of output, the smaller and harder to measure.

Roofing falls somewhere in the middle. A square is a clean unit. But job complexity varies, weather varies, tear-offs versus new construction varies, and a single bad slope can wreck a "good" production day. So I would expect roofing piece rate gains to look more like the construction range — meaningful but not the windshield number.

What the Data Does NOT Say

Before anyone screenshots the 44 percent and puts it on a sales page, here is the boring fine print.

Piece Rate Is Not Always Better

The research shows piece rate beats hourly on the right kind of work. The studies that found big gains were on tasks that were:

  • Easy to measure objectively
  • Mostly individual (not heavily team-dependent)
  • Quality-checkable independently of speed
  • Pace-controlled by the worker

If your work fails one or more of those tests, the academic gains shrink fast. A piece rate program on a job where two crews depend on each other and one bottlenecks the other will not produce 44 percent gains. It might produce zero. It might produce a fight.

Quality Does Not Improve Automatically

This is the one I want everyone to read twice. Piece rate makes people faster. It does not, by itself, make people more careful. The Safelite gain held up because the company already had strong quality checks and chargebacks for callbacks.

If you switch to piece rate without a quality control system, the speed will go up and the rework will go up at the same time. You may end up no further ahead, or behind. This is not a hypothetical risk — it is the most common failure mode.

For more on this, see the psychology of piece work and the linked piece on piece rate vs hourly in construction.

Hawthorne and Observation Bias

Anytime you change something on a crew, performance moves a little just because people know they are being watched. A new pay program comes with new attention from the boss, new conversations about output, new tracking. Some of the early gain is just that.

The Safelite study controlled for this by looking at performance over months, not weeks. If you switch to piece rate and the first two weeks are 30 percent higher than usual, take that with a grain of salt. Wait two or three months and look again before you decide what the real gain is.

Sorting Effects Take Time

The other half of the Lazear finding — the sorting effect — does not happen in week one. It plays out over months and years as your slower workers leave on their own and your hiring funnel attracts faster workers. If you measure productivity 30 days after switching and only see incentive effects, you have not seen the full picture yet. The compounding piece comes later.

What Actually Drives Real-World Gains

The research, plus what I learned running roofing crews, points to four things that determine whether a piece rate program produces meaningful gains or fizzles out.

1. Clean Rate Cards

If your rates are negotiated on the spot for every job, your crew never knows what they are working for, and you cannot benchmark anything. A clean rate card — published rates per unit, with documented adjustments for pitch, height, complexity, tear-off layers — is what makes the system trustworthy on both sides. See the article on building a piece rate pay scale for how to structure one.

2. Quality Controls That Bite

A walk-through after the job. A callback policy that holds the crew that did the work accountable. Photos before and after. The crews that produce the academic-style gains are the ones where being fast and being sloppy do not pay. Being fast and being clean pays.

3. Overtime Compliance

Federal law requires an overtime premium on top of piece rate earnings for hours over 40. The math is the half-time premium method, not straight 1.5 of an hourly rate. If your program drifts into wage violations, the back pay and liquidated damages will eat any productivity gain, fast. The FLSA piece rate compliance article walks through the math.

4. Honest Measurement

Track output per labor hour, not just total output. Track callbacks and rework as a separate line. Track gross profit per job, not just revenue. The crews and contractors who learn the most from piece rate are the ones who measure it honestly. The article on crew performance monitoring covers what metrics are worth tracking.

A Worked Example: 3-Person Roofing Crew Over 30 Days

Here is an illustrative comparison. Same crew, same job mix, two pay structures. Numbers are made up to show the math, not pulled from any actual customer.

The Setup

  • 3-person roofing crew
  • 22 working days in a 30-day window (weather and weekends)
  • Hourly baseline: $30/hour blended, 9-hour days
  • Piece rate: $80 per square (residential tear-off and re-roof, blended rate)

Hourly Baseline

ItemValue
Hours per day per worker9
Total crew-hours over 22 days594
Hourly rate (blended)$30
Total wages$17,820
Squares completed110
Labor cost per square$162

The crew produces 5 squares per day on average. Some days are higher (8 squares on a clean tear-off), some are lower (2 squares on a steep complex job). Output is steady but not pushed.

Piece Rate

Now imagine the same crew, same jobs, switched to piece rate at $80 per square. Assume a realistic 20 percent production gain — well below the Safelite number, well within the construction-research range — driven by less standing around, faster cleanup, faster lunch breaks, and the crew calling for the next job sooner.

ItemValue
Squares per day (20% gain)6
Squares over 22 days132
Piece rate per square$80
Total piece rate wages$10,560
Hours worked (estimated)540
Implied regular rate$19.56
Plus OT half-time premium (est.)$530
Total labor cost$11,090
Labor cost per square$84

The headline numbers move a lot:

  • Output up 20 percent (110 to 132 squares)
  • Labor cost per square down 48 percent ($162 to $84)
  • Total wage spend down, even though more squares were produced

That second number — labor cost per square — is the one that matters for your job costing. You can estimate your fully burdened labor rate and plug it into bids, but on production work the per-unit cost is what controls margin.

What This Example Does NOT Prove

A few caveats before anyone takes these numbers and runs.

  • The 20 percent gain is an assumption, not a measurement. Your real number could be 10 percent or 30 percent. Some crews see nothing in month one and meaningful gains in month three.
  • Hours worked under piece rate often go down, not up, because the crew finishes sooner. That is part of why the per-square cost drops. But if hours go down too far, take-home pay drops, which can drive turnover. Watch this metric.
  • Quality is not in the table. If callbacks double, the savings disappear. The cost of a single redo on a residential roof can wipe out a month of gains.
  • Slow weeks are not in the table. If a week of rain knocks the crew down to 12 hours of paid work, you still owe minimum wage make-up plus any state-specific rest and recovery pay (the California piece rate calculator covers that state's rules). Hourly absorbs that risk; piece rate does not.

A Note on Methodology

The research I am citing — Lazear 1998, the broader productivity literature — is real and reproducible. The construction-specific gains are noisier, partly because there are fewer clean studies and partly because the work itself varies more.

If you want to read the source material, Lazear's paper is "Performance Pay and Productivity," published in the American Economic Review. It is dry, it is honest, and it is one of the most-cited applied micro papers in labor economics for a reason. BLS productivity data and BEA construction output series can also give you sector-level context, though neither cuts the data by pay structure.

What I want you to walk away with is not "piece rate gives you 44 percent." It is "piece rate has a real effect on production work, the size depends on your setup, and the gain only sticks if you measure it honestly and protect quality."

Closing

Piece rate works because it ties pay to the thing the business actually sells: completed work. The research backs that up across industries, going back decades. Roofing is not auto glass, and your gain will not be 44 percent. But a real, durable improvement in output per labor hour — paired with happier top producers and natural attrition of slower workers — is well within reach if you set the program up right.

If you want to run the numbers on your own crew, the piece rate calculator and labor burden calculator can help you compare scenarios with your own rates. To put a piece rate program in place and actually track the output, callbacks, and per-square cost over time, start a Piece Work Pro account and run a couple of crews on it for a quarter.

For more on the structural side of switching, the article on whether piece work is right for your business and the piece rate vs hourly motivational comparison both go deeper into where piece rate fits and where it does not.

Frequently Asked Questions

Is there real research showing piece rate beats hourly pay?

Yes. The most cited study is Edward Lazear's 1998 analysis of Safelite Auto Glass, which moved technicians from hourly to piece rate and saw productivity rise roughly 44 percent. About half of that came from existing workers producing more, and about half from sorting — slower workers leaving and faster workers being hired or staying. The result has been replicated in other settings, but the size of the gain depends heavily on the type of work.

Does piece rate always beat hourly pay?

No. Piece rate works best on tasks where output is easy to measure, quality can be checked independently, and the worker controls the pace. On tasks that require slow, careful work, heavy teamwork, or constant supervision, piece rate can backfire. The research shows large gains on simple repeatable production tasks, smaller and noisier gains everywhere else.

Why might my crew not see a 44 percent gain when I switch to piece rate?

The Safelite number came from a clean, controlled environment with strong quality checks and consistent volume. A roofing crew faces weather, varied job complexity, drive time, and supply hiccups that all cap how much output-based pay can move the needle. A realistic improvement on production tasks is often in the 10 to 30 percent range, with the rest coming from better hiring and retention over time.

How do I know if the gains I see are real and not just observation bias?

Compare the same crew or similar crews over a long enough window — three to six months at minimum — and track output per labor hour, not just total output. Watch callbacks and rework separately. If output is up but rework is also up, you may not actually be ahead. Tracking the same metrics before and after the switch is the only way to be honest with yourself about the real gain.

Free Guide

How to Pay Your Crew 20% More and Double Your Profit

The math most contractors never run — and the mistakes that cost them $93K+ a year. This free PDF breaks down the math in ten minutes. Plus, you'll understand the payroll traps that can wipe you out.