The tolerated pain
Many businesses can report sales quickly but cannot allocate service effort, fulfilment cost, support load, discounts, returns or infrastructure consumption to the revenue that caused them. The monthly management pack shows revenue by customer, product or service line in detail. Below the gross margin line, almost everything becomes a single pool of overhead.
That is tolerable for a while because revenue is easy to measure and cost is not. Sales figures come straight from invoices. The cost of serving a customer is scattered across timesheets, support tickets, delivery records, credit notes, payment terms and the goodwill of people who quietly absorb extra work. Nobody has deliberately hidden it. The systems were simply never designed to connect the effort to the account.
The problem is old and well understood. Robin Cooper and Robert Kaplan argued in the Harvard Business Review in 1988 that managers in multi-product companies were making pricing, product-mix and process decisions on distorted cost information, and that the distortion often stays invisible until competitive and financial performance has already suffered. The tools have changed since then. The pattern has not.
Why businesses tolerate it
- Revenue is rewarded. Sales targets, commission plans and board reporting usually measure what was sold, not what it cost to deliver.
- Allocation feels contentious. Any attempt to assign shared cost invites argument, so the business avoids the exercise rather than agreeing a reasonable method.
- The data lives in different systems. Finance holds the ledger, operations holds the effort and support holds the tickets. Joining them is a manual investigation.
- Large customers feel safe. A big account looks valuable by default, so its cost to serve is rarely questioned.
Operational consequence
Teams optimise volume because volume is visible. Pricing, account management and product decisions are then made without a reliable view of contribution.
The symptoms are familiar. Discounts are approved to win or keep large accounts without anyone testing what the account already costs to serve. Bespoke requests, rush orders and extended payment terms are granted one at a time and never added up. Support and delivery teams know which customers consume disproportionate effort, but that knowledge stays anecdotal because it never reaches a report. Product lines that look healthy on revenue may depend on heavy manual handling that nobody has priced.
When margin falls, the business cannot say why. Was it pricing, mix, service creep, payment terms or delivery inefficiency? Without attributable cost, every explanation is an opinion, and the resulting decisions tend to be broad cost cuts rather than targeted corrections.
How it becomes money
contribution = attributable revenue − attributable variable and directly traceable operating cost
The exact model should match the business. The danger is not imperfect allocation; it is making material commercial decisions while known cost drivers are invisible.
An illustrative cost model
The following example is entirely illustrative. Every figure is an assumption chosen to show the arithmetic, not a benchmark or an observed result.
Assume a services business reports a blended gross margin of 32% on delivery cost. It compares two customers:
| Line (illustrative) | Customer A | Customer B |
|---|---|---|
| Annual revenue | R4,800,000 | R1,200,000 |
| Gross margin at 32% | R1,536,000 | R384,000 |
| Account management (assumed share of a R650,000 loaded role) | 50% = R325,000 | 5% = R32,500 |
| Support load (assumed share of a R1,800,000 support team) | 30% = R540,000 | 3% = R54,000 |
| Bespoke reporting (assumed 8 hours a month at R450/hour) | R43,200 | R0 |
| Extra 60 days' payment terms financed at an assumed 11% | R86,795 | R0 |
| Contribution after cost to serve | R541,005 (11.3%) | R297,500 (24.8%) |
The financing line is the extra receivable balance (R4,800,000 × 60 ÷ 365 = R789,041) multiplied by the assumed 11% cost of funds. On the reported gross margin, both customers look identical. Once directly traceable cost to serve is included, Customer A earns less than half the margin percentage of Customer B, despite producing four times the revenue.
Now suppose Customer A asks for a further 5% price reduction at renewal. On a revenue view, R240,000 looks like an acceptable concession on a R4.8 million account. On a contribution view, it removes 44% of the account's remaining contribution and takes it to roughly 6.6% of revenue. Those are very different decisions, and only one of them is visible in a revenue report.
Note what the model deliberately excludes. Shared overhead such as rent, finance and leadership is not allocated, because forcing an allocation would add argument without changing the decision. The model uses only cost that can be traced to the customer with reasonable confidence, and it labels each assumption so that it can be challenged.
Secondary consequences
- Mispriced growth. If the least profitable customer profile is also the fastest-growing, revenue growth can reduce total profit.
- Misdirected sales effort. Commission and attention flow to the accounts that look largest rather than those that create the most value.
- Slow, disputed analysis. Every profitability question becomes a one-off spreadsheet exercise, with numbers that are hard to trace back to source. The same weakness is explored in the audit cost of untraceable reporting numbers.
- Reconciliation labour. Joining finance and operational data by hand each month consumes skilled time, as described in the labour cost of manual reconciliation between systems.
Commercial trigger
If a high-revenue customer or service cannot be tested for contribution without a manual investigation, margin visibility is too weak for confident growth decisions.
The trigger becomes urgent at specific moments: a major renewal or tender, a price increase, a request for a large discount, a decision to launch or retire a product, falling margin despite rising revenue, or a plan to scale a service line. Each of these commits the business to a cost structure for months or years. Making the commitment without contribution data is a decision taken on partial evidence.
What good looks like
Good profitability visibility does not require a perfect costing system. Kaplan and Steven Anderson's 2004 Harvard Business Review work on time-driven activity-based costing argued that informed managerial estimates of time and capacity can produce a more flexible and practical cost model than large, survey-based exercises. The principle is useful for any business: estimate the main cost drivers sensibly, connect them to transactions and refine them as evidence improves.
In practice, the business can produce a contribution view by customer, product or service line on a regular cycle, from connected data rather than a fresh investigation. Assumptions are visible in the report. Commercial decisions such as discounts, renewals and service changes are reviewed against contribution before they are approved.
Resolution
- Define the unit whose profitability matters.
- Agree revenue and cost definitions.
- Connect operational drivers to financial records.
- Separate directly traceable cost from allocation assumptions.
- Expose assumptions in reporting rather than hiding them.
- Review pricing and service design using contribution, not revenue alone.
A sensible first step is to run the model above on your ten largest customers using your own figures. If the result surprises anyone, the gap in visibility is already costing money. Where the data needed sits across several disconnected systems, a digital systems architecture roadmap can set out how finance and operational records should connect.
Bottom line
Revenue tells the business what it sold. Profitability visibility helps it understand whether the way it sold and delivered that work is economically sustainable. Without it, the largest accounts can quietly become the most expensive ones, and each discount, concession and extended term widens the gap without anyone seeing it.
Sources and further reading
- Measure Costs Right: Make the Right Decisions (last verified 10 October 2026)
- Time-Driven Activity-Based Costing (last verified 10 October 2026)