management

Comp Plans That Survive a Slow Week

Why pure commission breaks canvassing teams, what the FLSA requires, and how to pay for the work instead of only the outcome.

The comp plan is the only management tool that works when nobody is watching. A canvasser four miles into a territory at 6:40pm on a Thursday is going to knock the next door or not knock it based on what that door is worth to them, not based on a sales meeting.

Most canvassing comp plans are designed for the good week. The problem is the slow week, because the slow week is when people quit.

Start with the legal question, not the motivational one

Before designing anything, settle two things. Getting them wrong makes the rest academic.

Employee or contractor. The Department of Labor's Fact Sheet #13 is blunt about what does not settle this: "What the worker is called is not relevant," and an independent contractor agreement "does not make a worker an independent contractor under the FLSA." A canvasser working your territory on your schedule with your script and no capital at risk is weak on exactly the factors that carry the most weight.

Exempt or not. The outside sales exemption, described in DOL's Fact Sheet #17F, requires that the primary duty be making sales and that the employee be "customarily and regularly engaged away from the employer's place or places of business." It carries no salary floor. But 29 CFR 541.503 says "promotional work that is incidental to sales made, or to be made, by someone else is not exempt outside sales work."

A setter who hands off to a closer is doing promotional work for someone else's sale. That is the structure most roofing and solar teams use, and on the face of the regulation it is the structure least likely to qualify. If the exemption does not apply, minimum wage and overtime do, and no commission-only agreement changes that. Have an employment lawyer review your actual structure.

Everything below assumes you answered those two questions. If not, the plan may be illegal rather than merely ineffective.

Pure commission fails for a structural reason

The argument for commission-only is that it aligns incentives and costs nothing when nobody sells. Both halves are wrong.

On alignment. Commission aligns the rep to the outcome, but the outcome in canvassing is heavily determined by things the rep does not control. A rep working a daytime pass in a commuter subdivision in February will contact fewer households than a rep working Saturday mornings in a retiree neighborhood, and the pay gap between them is mostly a map and a calendar, not effort. Reps know this. Comp that punishes them for the territory they were assigned does not read as alignment, it reads as a rigged game.

On cost. Commission-only costs nothing when nobody sells, and it also produces nobody. A rep in week two with no sales and no income does not stay long enough to become a rep in week eight. You are not saving money, you are converting payroll cost into recruiting cost and losing the ramp investment every time.

BLS lists a May 2025 median annual wage of $41,380 for door-to-door sales workers on its occupations not covered in detail page. Whatever your plan looks like on paper, a new rep is comparing their actual first-month take against alternatives in that range. If a bad territory and a rainy week put them well under it, the plan has a retention problem regardless of its upside.

Pay the input, bonus the outcome

The structure that survives a slow week pays for two things.

A floor tied to the work, not to the result. An hourly wage or a per-door rate for logged, verified knocks. This is the part that gets a rep through week two and through a rained-out week, and it is what makes a comp plan something other than a lottery ticket. If the rep is non-exempt, you need a wage floor anyway.

Per-door pay has one hazard worth naming: it pays for logging, which creates an incentive to log doors that were not knocked. That is manageable, but only if the log has a timestamp and a GPS location, so that forty doors in nine minutes from a parked truck is visible. Without verification, per-door pay buys you fiction.

Upside tied to the outcome. Per-set and per-sale bonuses, with the per-sale piece larger. This is where the real money is and should be.

The ratio between floor and upside is the real decision. Weight the floor too heavily and you get attendance without intensity. Weight the upside too heavily and you get the commission-only failure above. The honest answer is that it depends on your close rate, your average job value, and how long your ramp is, which is why the next section matters more than any recommended split.

Do the unit economics before you set a rate

You cannot price a knock without knowing what a knock is worth, and most teams set commission rates by copying a competitor.

Work it backward:

1. Doors per set, measured from your own data over several hundred doors.
2. Set-to-close rate, from your own closers.
3. Average gross profit per job, not revenue.

Multiply through and you get gross profit per door. That is the ceiling on what a door can cost you across all compensation, recruiting, and management. Everything you pay out, floor plus bonus plus the manager's time, has to fit under it with margin left.

Teams that skip this arrive at a plan that feels generous and is unprofitable, then cut it mid-season, which does more damage to trust than a tighter plan would have done from the start.

Think hard before adding a quota

Quotas have a specific, documented failure mode. Harikesh Nair and Sanjog Misra studied a large sales force that moved to a plan without a quota requirement; Stanford GSB's summary, Eliminating Sales Quotas May Stimulate Profits, reports revenues rose 9 percent, around a million dollars a month.

The mechanism is the useful part. Quotas create a "perverse incentive to postpone their effort to the next cycle." A rep who has hit quota stalls the next deal into next period. A rep who cannot reach it banks the effort for a fresh start. Both are rational.

Two caveats: the 9 percent came from a whole plan redesign rather than quota removal alone, and that page is the school's summary, not the paper. But the pattern is checkable in your own numbers. If sets spike at period end and go quiet right after, you are paying for gaming.

If you keep a quota, keep it on an input the rep controls, like logged doors, where there is no reason to stall.

Accelerators, draws, and the two traps

Accelerators work. A higher rate past a threshold gives a rep a reason to knock the last hour of the week. Keep them simple enough to compute on a porch.

Draws are how most teams bridge the ramp: an advance against future commission. The trap is the recoverable draw that accumulates into a debt the rep can never clear. A new rep who learns in week six that they owe the company money quits that day and tells everyone. If you use a draw during ramp, make it non-recoverable for the ramp period. It is a training cost. Call it one.

Clawbacks on cancellation are legitimate, because door-to-door sales carry statutory cancellation rights and a sale inside the window is not a sale. But a clawback that reaches back months turns a rep's paycheck into something they cannot predict, and unpredictable pay is the thing people leave over. Tie the clawback to the cancellation window and stop there.

Make the number visible daily, with the gap attached

Comp plans fail at communication more often than at design. A rep who cannot compute today's earnings will not trust the plan, and a plan nobody trusts does not motivate.

Two requirements. The rep should be able to see what they have earned today, not at the end of the pay period. And they should be able to see what the next threshold is worth and how far away it is.

That second part has evidence behind it. Iwan Barankay's randomized experiment with 1,754 salespeople, Rank Incentives, found that removing bare rank feedback raised sales performance by 11 percent, while adding benchmarks that showed what was needed to climb "significantly raises performance." A number with no visible path attached discourages. The same number with the gap attached motivates.

The test

A comp plan is working when a rep who has a genuinely bad week, because of weather or a bad territory draw, still earns enough to show up Monday, and a rep who has a great week earns enough that they tell a friend to apply.

If only the second is true, you are running a recruiting treadmill and paying for it twice.

Canvass Pro timestamps and geolocates every logged knock, which is what makes a per-door or per-input component auditable rather than self-reported.