On this page
Four labels can appear beside the same sales result: net sales, gross margin, contribution margin, and profit. They are not four names for one number. They are different economic objects, and each becomes meaningful only after its scope and calculation rule are written down.
The direct answer is therefore a definition rule: gross margin is not contribution margin, and neither is automatically net sales or profit. A compensation base can signal which result an organization can measure and wants to influence. It cannot, by its label alone, prove the profit created by a seller or the value of work that the number does not observe.
That distinction matters before anyone debates a rate. A plan tied to net sales may observe top-line commercial activity while ignoring product economics. A plan tied to a defined gross margin may reflect a product-cost boundary while omitting service or acquisition costs. A contribution-margin base may include some variable delivery or service costs, but only if the organization has defined and assigned them. A profit measure may contain overhead and allocation choices outside a seller’s control.
The coverage problem in compensation asks which commercial work a plan leaves visible or invisible. The cost boundary in a customer P&L asks which costs belong in the local result. REV-04 asks a narrower question first: which metric object is the plan actually using?
Why is gross margin structurally different from contribution margin?
Margin is a word, not a universal calculation. A management record should define its fields before it uses the label.
Net sales can be treated as the sales amount after the reductions the organization has explicitly classified as returns, rebates, discounts, credits, or other deductions. The exact definition depends on the accounting policy and the data cut. The important point is that sales without the reduction rule is not a stable object.
Gross margin can be treated as net sales less the defined cost of goods sold or cost of sales for the scope being measured. Some organizations use a product-cost boundary. Others use a generated gross-margin measure in a commercial system. The label does not settle whether freight, support, implementation, returns, payment costs, or acquisition costs are included.
Contribution margin usually asks what remains after the variable costs assigned to a decision or incremental sale. The assigned costs may include delivery, service, transaction, or acquisition items that sit outside a product-cost definition. That makes contribution margin potentially useful for a decision, but also makes its boundary especially important.
Profit is a result for a declared scope after the costs and other items assigned to that scope. It may contain fixed costs, overhead, financing, taxes, shared functions, or allocation rules. A seller may influence some of those items and have no control over others. Profit therefore needs a responsibility map beside its arithmetic.
These working definitions are not a license to impose one accounting convention. They are a reason to write the local convention in the plan, data dictionary, report, and dispute process. If two teams use gross margin for different cost sets, they do not have a metric disagreement only. They have two different objects carrying one label.
A compensation base is a signal about observability
Every compensation base answers a practical question: what result can be calculated at the required frequency, attributed to a person or team, and explained when a payment is challenged? That question is different from what is the complete economic value of the sale?
Net sales is often easier to observe than a fully assigned profit. It can nevertheless encourage attention to volume, timing, or transactions that do not create attractive economics. Gross margin can introduce a stronger product-economics signal, but it can also be harder to calculate consistently and may omit work that happens after the sale. Contribution margin can bring more variable economics into view, but it requires credible cost assignment and a governance rule for items that move between fixed and variable categories. Profit can be closest to an overall result while being furthest from a seller’s direct control.
The right interpretation is not that one base is inherently superior. The base reveals a tradeoff among observability, controllability, strategic coverage, calculation burden, timing, and behavioural risk. A strong plan makes that tradeoff explicit and tests the outcome rather than hiding it behind a familiar label.
What the 1959 report actually shows
Cassidy and Wischkaemper’s USDA report is useful precisely because its scope is limited. It presents case studies of nine selected institutional wholesalers (Cassidy & Wischkaemper, 1959). The report is dated 1959, and the firms were selected purposively rather than as a representative sample. Its practices should therefore be read as a historical description, not as current evidence or a benchmark.
The report says, “Three basic sales compensation systems were used by the wholesalers” (Cassidy & Wischkaemper, 1959). It describes commission on generated gross margin, commission on net sales, and a combination of salary, commission, and bonus. Four of the selected wholesalers used commission on each salesperson’s generated gross margin. Four used net-sales commission with variations in rates or categories.
One described system combined salary, net-sales commission, and a bonus tied to average daily generated gross margin.
Those counts are source-reported descriptions of the selected firms. They do not tell us which system performed better. They do not define modern gross margin. They do not prove a causal profitability effect. They show that compensation design has long contained different measurement boundaries. The report also records the phrase “it would not fit”.
The report makes another boundary especially relevant. It distinguishes work measurable through net sales or generated gross margin from service and account-development work that is not directly related to those measures. A number can be accurate and still be incomplete as a signal of work.
| Metric object | What it measures in this framework | Work it may signal | What it may leave out | Outcome question still open |
|---|---|---|---|---|
| Net sales | Sales after explicitly defined reductions. | Transaction creation, volume, and commercial timing. | Product economics, service burden, acquisition cost, and shared cost. | Did the activity create acceptable economics and durable customer value? |
| Gross margin | Net sales less the defined product or sales cost boundary. | Product mix, price discipline, and generated gross-margin result. | Costs outside the selected product-cost boundary and less measurable account work. | Is the margin definition stable, attributable, and sufficient for the decision? |
| Contribution margin | Net sales less variable costs assigned to the decision or sale. | Incremental economics when cost assignment is credible. | Fixed cost, shared overhead, unassigned work, and policy-dependent items. | Are assigned costs truly incremental and within the decision’s influence? |
| Profit | Result after the declared scope of costs and other items. | Overall economic result for a business scope. | Causal attribution, uncontrollable allocation, and non-financial work. | Which part of the result can the compensated role control? |
Figure 1Four metric objects, four unanswered questions
The metric label is only the beginning. Each base needs a scope, responsibility map, behaviour check, and outcome test.
Source: Author's framework grounded in Cassidy and Wischkaemper (1959). Rows and definitions are synthetic, scope-dependent examples.
The work signal is not the same as the economic result
The historical distinction between sales-measurable work and service or account-development work helps explain why metric debates become behavioural debates. A seller can win a transaction that produces net sales while another team performs implementation, support, renewal preparation, or relationship repair. The transaction number may be correct. The work system is larger than the number.
Sabnis and colleagues studied lead follow-up rather than margin definitions. Their usable analysis came from 461 sales representatives across four B2B firms, and the compensation systems were primarily based on quota achievement, commissions, and top-line bonuses. Sabnis et al. (2013) model time allocation across marketing leads, self-generated leads, and non-acquisition activities.
That evidence belongs in its own lane. It does not prove that a gross-margin base improves lead follow-up, nor does it establish an optimal compensation plan. It does show why a base can shape attention through what it makes visible and payable. If lead follow-up, account development, or service recovery is outside the base and outside the scorecard, the plan can leave that work weakly represented even when leaders say it matters.
The correct response is not to add every activity to one metric. It is to decide which work belongs in the role, which work needs a separate measure, which work is a shared responsibility, and which outcome will test whether the combination works.
Gross margin and contribution margin can diverge on one deal
Consider a synthetic deal with a high invoice value, a moderate product margin, and a substantial implementation burden. The gross-margin field may subtract the product cost and show a positive result. The contribution-margin field may also subtract variable implementation labour, payment costs, delivery, or a support commitment if the organization has defined them as incremental and assigned them to the deal. Profit may additionally absorb fixed team cost or an allocation of shared overhead.
No label tells us which result is correct. Each can be correct for its declared question. The problem starts when a plan calls the first number margin, pays on it, and later interprets the payment as if it represented the third number.
The diagnostic sequence is simple:
- Write the numerator and every deduction in plain language.
- Mark each item as transaction-specific, variable, fixed, shared, or not assigned.
- Map each item to the person or team that can influence it and the date when it becomes observable.
- Record exclusions such as returns, credits, service, acquisition, financing, tax, or overhead.
- Test whether the resulting measure predicts the outcome the organization actually cares about.
This sequence does not choose a base. It prevents a label from choosing one silently.
Figure 2The metric-definition worksheet
A compensation base is not ready for interpretation until its formula, responsibility boundary, timing, exclusions, and outcome check are visible.
Source: Author's worksheet. Blank fields are reader inputs; no company or compensation data is supplied.
Timing can turn a metric into a behavioural lever
The timing of a metric can matter as much as its definition. If payment depends on crossing a threshold before a period closes, a seller may have reason to move a transaction, discount a price, delay a booking, or concentrate effort at the boundary. The resulting number may be real while the timing pattern makes the underlying economics harder to read.
Oyer’s 1998 study gives this mechanism a specific empirical setting. Its analysis uses 31,936 quarterly observations covering 981 manufacturing firms and fiscal years from 1985 through 1993 (Oyer, 1998). Oyer describes nonlinear pay as creating an “incentive to manipulate prices, influence the timing of customer purchases, and vary effort over their firms’ fiscal years”.
The finding is not a modern sales-plan benchmark. It concerns manufacturing-firm seasonality and a particular incentive mechanism. Oyer’s evidence is consistent with incentive responses, but it does not establish an optimal compensation plan or tell us how much any metric costs a specific organization.
The lesson for metric design is narrower: define the observation window and the reversibility of the underlying transaction. A monthly gross-margin figure, a quarterly contribution figure, and a recognized profit figure may describe different moments. If a seller can influence timing but not the final cost settlement, the metric may pay before the economic result is stable.
Build a diagnostic, not a verdict
The following matrix is an author framework. It is a way to challenge a metric before it becomes a plan. It does not say gross margin is better than contribution margin, net sales, or profit.
Figure 3From measured object to outcome check
A metric becomes a compensation input only after work, timing, possible behaviour, and outcome have been tested separately.
Source: Author's schematic framework grounded in Cassidy and Wischkaemper (1959), Sabnis et al. (2013), and Oyer (1998). Presence coding is not empirical data.
Use the matrix to ask five separate questions. What is measured? What work is rewarded? Which costs are included? What behaviour could the timing invite? Which outcome determines whether the measure is useful? If one answer is missing, the plan may still be a draft, but the metric should not be described as a complete proxy for profit.
Implementation questions without choosing a rate
An implementation conversation can remain rigorous without recommending a particular commission rate. Start with the definition ledger. Give every metric a formula, owner, source system, period, adjustment rule, exception policy, and version. Do not allow a label to change meaning between compensation documents, finance reports, and commercial dashboards.
Then build the responsibility map. A seller may influence price and customer selection but not supply cost, returns, implementation staffing, or shared overhead. A manager may influence territory and coaching but not the accounting close. A metric can still include an item outside direct control, but the plan should say why and how the resulting exposure is governed.
Next define the timing rule. State when a transaction becomes eligible, when deductions can reverse it, and how disputes are handled. Compare the payment date with the date when the economic outcome becomes reliable. This is where the Oyer mechanism becomes an operational question rather than a historical analogy.
Finally separate diagnosis from redesign. A pilot or retrospective can ask whether the metric predicts the declared outcome, whether the work signal is credible, whether exceptions cluster, and whether timing behaviour changes. It should not present the observation as proof that one base is universally optimal.
What this source set does not establish
The Cassidy and Wischkaemper report documents compensation variants among nine selected wholesalers in 1959. Sabnis and colleagues study sales-representative time allocation in four B2B firms. Oyer studies a fiscal-timing mechanism in a large manufacturing sample. These sources illuminate measurement and behavioural boundaries. They do not establish a current benchmark, an optimal gross-margin definition, an optimal commission rate, or a causal profitability effect for a named organization.
The reliable conclusion is modest: keep gross margin, contribution margin, net sales, and profit as separate fields until an explicit definition proves that two labels describe the same object. Then map the chosen object to work, control, timing, behaviour, and outcome. If the metric cannot answer those questions, it may still be a useful report. It is not yet a complete compensation argument.
References
- Cassidy, P. J., & Wischkaemper, P. (1959). Policies and practices of some leading institutional wholesale grocers (Marketing Research Report No. 335). Agricultural Marketing Service, U.S. Department of Agriculture. https://doi.org/10.22004/ag.econ.311161
- Oyer, P. (1998). Fiscal year ends and nonlinear incentive contracts: The effect on business seasonality. The Quarterly Journal of Economics, 113(1), 149-185. https://doi.org/10.1162/003355398555559
- Sabnis, G., Chatterjee, S. C., Grewal, R., & Lilien, G. L. (2013). The sales lead black hole: On sales reps' follow-up of marketing leads. Journal of Marketing, 77(1), 52-67. https://doi.org/10.1509/jm.10.0047
