Compensating Sales on Gross Profit, Not Revenue: Discount Control That Survives Multi-Product Deals

TL;DR: Revenue-based sales compensation pays the same commission on a discounted dollar as on one closed at scheduled net, so reps trade margin for velocity and the plan cannot see it. Moving comp to gross profit prices that trade, but only once a scheduled net price exists to measure the landing against.

A sales rep paid on revenue earns the same commission on a dollar of discounted revenue as on a dollar closed at the scheduled net price. A concession costs them in proportion to the revenue surrendered and nothing in proportion to the margin destroyed, which on a thin line is very nearly all of it. On a multi-product enterprise deal, where the buyer probes every line, the rep’s rational move is to buy the close with margin the plan never priced. The sales team is doing what the plan pays them to do.

Gross-profit compensation re-prices the rep’s own decision: a point of unnecessary discount comes out of their number, so the rep lands the deal at the commitment level that maximizes contribution, not contract size. But comp is the behavioral lever, not the fix. Move compensation to gross profit while the pricing architecture underneath is broken and you relocate the argument from the discount approval queue to the comp-plan exception queue.

Why Revenue-Based Sales Compensation Pays Reps to Discount

Reps are paid on what lands, not on what was listed. So a fifth given away to get a deal signed costs the rep a fifth of that deal’s commission, and nothing beyond it. Set against a faster close, a better win rate, and a quarter that makes its number, that is a cheap trade, and reps take it correctly. The same concession under a gross-profit plan comes out of contribution rather than revenue, where a fifth off the landed price can be most of the margin on the line. The trade stops being cheap.

Mix is where a revenue plan goes fully blind. Twenty points off a product carrying most of its price as margin and twenty points off one carrying very little are the same event to the plan and completely different events to the business.

Most sellers have met this dynamic from the other side of a house sale. An agent working on commission has a strong interest in a close and a thin interest in your last few points, which is why the first offer so often arrives described as the best one you will see. Agents selling their own homes hold out longer and do better on average. Reps sit in the same position, and they are not being cynical about it. Making the number across many deals is more reliable than winning a few points of margin on one and stalling everything queued behind it.

Peer-reviewed work on incentive contracting in enterprise software finds deals bunching at the end of a period and closing at deeper discounts. The mechanism is the plan’s own geometry: attainment resets on a fixed clock, and the payout rate jumps at quota thresholds rather than staying flat per dollar. A rep just under a line with days left will trade real margin for the close, because crossing it pays more than the concession costs.

The revenue plan also fails silently at the layer where the damage happens. The number that matters on a closed deal is the realized price: what the deal actually yields against the scheduled net price, once every concession is counted. Revenue comp never looks at that gap. A rep whose book lands 15 points below schedule on every deal posts the same attainment as a rep landing on schedule. Finance sees the difference quarters later, when nobody can trace it to any single decision.

Fix the Pricing Architecture Before the Comp Plan

Here is the part comp redesigns routinely skip: most discounting that looks like rep behavior is a rep compensating for a gap in the pricing architecture. In our diagnostic work on pricebook deviation, the deviation pattern usually traces upstream of the sales team entirely.

A value metric that misfits usage forces the discount. When the count runs high relative to how the customer actually extracts value, the rep discounts to correct the meter’s overcount, and the deal desk approves it because the partial-use story sounds reasonable every time.

Packaging that embeds volume thresholds in edition boundaries forces the discount. When a customer has to jump editions just to add units, the rep discounts the jump, because the buyer is being asked to pay an upsell price for an expansion motion.

A discount schedule that drifted into open negotiation forces the rest. Once the published schedule stops governing, every deal starts from the last exception, and the rep has no defensible position to hold.

None of those gaps are fixable at the comp plan. Pay that same rep on gross profit and they still face a misfit metric, a forced edition jump, and a schedule nobody believes. The difference is that every architecturally forced discount now becomes a comp dispute. With no uniform scheduled net price to measure against, there is no way to separate the concession the architecture forced from the one the rep chose, so all of it reads as discretionary and the rep pays for it. The exception queue fills up, and within two quarters the carve-outs have swallowed the plan.

The fix order runs in one direction. Architecture first: the licensing model, then packaging, then the pricing model, the sequence laid out in our enterprise SaaS pricing guide. Then a margin-calibrated pricing surface that schedules a net price at every commitment level a customer might make. Then compensation tied to where the rep lands on that surface. Comp is the last decision because it only has something rational to measure once the first two exist.

What Gross-Profit Compensation Changes on Multi-Product Deals

Single-product deals hide the worst of revenue comp’s blindness. Multi-product deals expose it, because margin differs by product and revenue comp cannot always see mix.

Consider a portfolio where one product is licensed IP with gross margin in the nineties, another carries hosted infrastructure cost, and a third is an AI-bearing product whose COGS scales with LLM inference. A $1M deal weighted toward the first and a $1M deal weighted toward the third are very different outcomes for the company. Under revenue comp they can often be the same deal, so the rep assembling the package has no reason to care which lines carry the volume.

The allocation choice matters even more than the mix. Every multi-product negotiation ends with the buyer wanting a number off the total and someone deciding which lines give it up. A discount absorbed by the high-margin line costs a fraction of the same discount absorbed by the inference-heavy line. Revenue comp prices those two choices identically, so the allocation gets made by whatever closes fastest. Gross-profit comp makes the allocation visible to the one person actually assembling the deal, at the moment they assemble it.

Multi-product deals are also where legacy discount structures interact worst. When each product carries its own climbing volume schedule, the composite math produces blended outcomes no one designed, the pattern margin-calibrated discounting exists to fix with a single composed surface. Across decades of patterns in our corpus, the negotiations that tier-step schedules invite vaporize 9 to 15 percent of total revenue across the customer base. A comp plan cannot recover that. A surface can, and a comp plan can hold it.

Revenue Compensation vs Gross-Profit Compensation

The last row is the one that decides whether the change survives. A revenue plan runs on a number every company already has. A gross-profit plan runs on a scheduled net price at every commitment level and a cost basis a rep can recompute, and neither exists by accident.

Tying Sales Compensation to the Pricing Surface

Once the surface exists, the comp design gets concrete. The surface produces a scheduled net price for any commitment a customer might make, across volume, product mix, Customer Group, and channel. The plan pays on gross profit and measures each deal by where it landed relative to schedule.

Under a discount-matrix regime, comp and pricing governance are separate systems that meet in an approval workflow. Under a surface regime they are the same system: the surface says what the deal should yield and the realized price says what it did; the rep’s number moves with the gap. A rep who lands at or above schedule keeps full economics; one who lands below funded the concession partly out of their own number. The plan now prices the decision at the desk where it gets made.

Two design details keep this workable. The acceptable variance around scheduled net price starts wide and tightens as the organization matures; a team emerging from years of chaotic discounting cannot jump to a tight band without breaking its own pipeline. And the surface, not the plan document, carries the pricing logic: when margin pressure shows up at a commitment level, you retune the surface once and every rep’s scheduled net prices update with it.

For a fast read on whether your current architecture could support this today, the Pricing Architecture Assessment scores the licensing, packaging, and pricing decisions that have to hold before a gross-profit plan can.

What Finance Will Ask of a Gross-Profit Plan

Gross-profit plans die in finance review for predictable reasons. Each one is a design decision with a real tension in it, and a plan that has not resolved them arrives at the review with the argument still open.

Which cost basis the rep is paid against. The moment allocated overhead enters the rep’s number, the plan stops being an instrument for deal behavior and becomes an argument about allocation methodology, which finance will win and the sales team will never accept. The tension is between a basis close enough to true contribution to point the rep at the right deal and one stable enough to survive being audited. Where that lands depends on how your COGS is actually structured.

No black-box math. A rep must be able to recompute their number from the lines on their own deal: quantity, scheduled net price, realized price, standard cost per line. If the calculation needs a finance analyst to explain, reps will not trust it, and a plan reps do not trust changes nothing except attrition.

Whether the plan’s cost table moves when costs move. Product-level costs shift, especially on AI-bearing products where inference economics change underneath you. If the plan’s costs track them, a rep’s attainment can move retroactively because an input did, which is the fastest way to lose the plan’s credibility. If they do not, the plan drifts from the economics it exists to protect. The surface and the plan do not have to reprice on the same clock, and deciding they do not is the point.

How much changes at once. Changing what the plan pays on and changing the booking-credit and clawback mechanics in the same cycle doubles the surface area of dispute, and the disputes arrive together with no way to tell which change caused which argument.

Do not stack a model change on a measure change. If the company is also moving from perpetual to subscription, or from subscription to consumption, then when revenue arrives is changing at the same moment as what the plan pays on. Those are two different problems, and running them together is how a transition produces the win-lose deals that follow a new pricing model: the rep is paid at signature while the company’s revenue now depends on renewal and on usage that may arrive later, or decline. If both are in flight, ask first which one your reps are currently being paid against.

Sequencing the Transition for an Installed Sales Team

An installed sales team has priced its own expectations against the current plan, so the transition is staged, not flipped.

Run a shadow period first, where gross-profit attainment and landing position against schedule are reported next to revenue attainment with no pay consequence. How long it runs is a judgment call about your own sales cycle, and running it shorter than one is the common mistake. The shadow data shows reps their own pattern before the pattern has a price, and it surfaces the architecturally forced discounts that must be fixed on the surface before the plan goes live.

Then blend, shifting weight from revenue to gross profit on a published schedule rather than in one cutover. In-flight pipeline closes under the plan it was sold under; a deal negotiated for nine months against one comp structure should not reprice the rep’s economics at signature.

Re-baseline quotas so on-target earnings hold at plan level. The transition re-prices behavior, not the sales team’s income; a plan read as a disguised pay cut will be worked around rather than worked with. The reps already landing near schedule come out ahead.

If deals in your book keep landing well below schedule and the pattern survives every comp tweak you have tried, the constraint is probably upstream of the plan. Talk to a pricing expert: describe what you are seeing in your own discount and margin data, and a pricing expert will reply with a read on whether the fix sits in the architecture, the surface, or the plan.

FAQs