A software-enabled marketplace can report a good quarter and still create a difficult management meeting. Volume has risen. Revenue has risen. Profit remains positive. Gross margin is lower, cash is thinner, and the latest forecast is weaker than the one used a few months ago.
The first explanation usually arrives quickly. Perhaps the customer mix has moved towards lower-margin accounts. Perhaps collections have slowed. Perhaps the forecast was optimistic. Each is plausible. None deserves to be treated as the answer before the numbers have been put back into their own time bases and definitions.
Use the marketplace as the thread for this article. We will trace the reported result through the statements, reconstruct the way the business earns, choose the comparison that actually matters, check the cash boundary, and decide what evidence would justify a commercial change.
Give each question to the right owner
The dashboard is mixing several jobs together.
| Question | Where the work belongs |
|---|---|
| What has been recorded? | Financial statements |
| What activity produced the result? | Business model and revenue architecture |
| What should actual performance be compared with? | Plan, budget, forecast and target |
| How long can the business fund the response? | Liquidity analysis |
| What should the team watch while it responds? | KPI system and metric contracts |
| Which explanations survive scrutiny? | Variance diagnosis |
| Which measures already have a definition and an owner? | Finance and FP&A metric library |
Higher revenue belongs in the first row. Transaction volume, subscription fees, commission rates, incentives and payment terms belong in the second. “Below expectations” is incomplete until a forecast or budget version is named. Falling cash needs a usable-funds and timing view. Customer mix can sit in the driver tree, but only as something to test.
That separation gives the meeting a sequence. It also keeps a popular KPI from being asked to answer every question at once.
Follow one transaction through the statements and into the model
Financial statements report different parts of the same business activity. The statement of financial position shows resources and obligations at a reporting date. The income statement records revenue, cost and profit over a period. The cash-flow statement separates movements in cash and cash equivalents into operating, investing and financing activities. They connect, yet each preserves a different view. The IFRS Conceptual Framework and IAS 7 set out that distinction.
Start with revenue in the example and then inspect receivables. The platform may recognise more revenue this quarter while collections lag behind it. That is enough to produce an improving profit figure alongside weaker cash. Capital expenditure adds a separate timing path: a payment may create an asset first and affect profit later through depreciation or impairment. IAS 16 is the standards context for that treatment. The exercise is a cross-check, not an attempt to force every line item into a one-to-one mapping.
The statements still leave one major question open: how did the business generate that revenue? Our marketplace has two earning mechanisms. Subscribers pay for continued access. Transaction commissions arise from qualifying activity on the platform. Their recognition triggers, timing and cost patterns may differ.
Keep the operating measures separate as well. GMV or GBV describes platform activity. Bookings often describe activity that has been ordered or contracted. MRR is a point-in-time recurring-revenue run rate. None should be substituted for recognised revenue. A take-rate calculation also needs a consistent numerator and denominator: period, currency, cancellation and refund treatment, transaction population, and gross or net basis.
At this point the useful work is specific. Break revenue into subscription and transaction components. Check what moved in volume, order value, commission rate, incentives and refunds. Check whether the new customers have different payment terms. The margin question will become easier to investigate once those mechanics are visible.
Keep the budget, forecast and cash view on the table
An actual result may be above budget and below the latest forecast at the same time. One comparison does not cancel the other.
A plan describes a direction and operating path. A budget carries resource allocation and accountability. A forecast estimates the future with the latest available information. A target states an intended outcome. Forecasts need a stated horizon, assumptions, method and update point so that later variance can be read in context. Revising a forecast does not erase the approved budget. The GFOA planning and budgeting framework treats planning, allocation and monitoring as connected but distinct activities.
Suppose the marketplace beats its original revenue budget but misses the forecast produced three months later. The budget comparison asks whether the original resources and commitments were appropriate. The forecast comparison directs attention towards recent demand, pricing, mix or execution assumptions. If both are labelled as one “miss”, the later bridge has no stable comparator.
Cash needs the same discipline. A reported cash balance does not show which funds are usable, in which legal entity and currency, or when they can be deployed. Put opening usable balance, expected receipts, expected payments, the low point and financing triggers on a near-term schedule. Slow collection, early incentive payments or concentrated capital expenditure can create a low point in usable cash during a particular period even where cash remains on the balance sheet. Burn, free cash flow, runway and minimum cash must keep their own definitions; a universal threshold would hide the assumptions that matter.
Build the dashboard around the decision now in front of the team
The marketplace does not need every available metric. It needs a few measures that describe the desired outcome, the boundaries around it and the candidate drivers worth investigating.
Revenue quality may be the outcome. Gross margin, minimum usable cash, refunds and risk exposure may act as controls. Subscription renewal, customer mix, commission rate, payment days, incentives and acquisition quality are candidate drivers. The distinction matters during the meeting: a driver can explain a movement only after evidence supports the mechanism, timing and affected population.
Give each material measure a metric contract. Record the calculation, population or legal entity, period, currency, source, update cadence, owner and action linkage. Public-company discussion of non-financial KPIs places similar emphasis on management use, calculation and changes in definition. It is a useful discipline for internal systems, not a universal operating rule. The SEC’s MD&A and KPI guidance is a relevant reference.
Once the team has agreed those contracts, “the mix got worse” can be turned into a checkable statement. Which customers are in the population? Did their margin and payment behaviour change in the same period? Did incentives move at the same time? The dashboard should make those questions possible; it should not answer them by implication.
Make the variance add up before assigning a cause
Choose the comparator first: actual against budget, or actual against the latest forecast. Lock the metric definition, scope, period, currency and data version. A bridge built on mixed bases can be numerically tidy and still answer the wrong question.
The marketplace bridge might separate volume, rate, mix and timing. Higher volume contributes positively. Some commission rates fall. The share of lower-margin, slower-paying customers rises. A portion of receipts lands after the period. Other businesses may need efficiency, foreign exchange, one-off, classification or scope components instead. The component list follows the business model; the total must reconcile to the original difference.
Any remainder stays visible as a residual. It represents something the current decomposition has not explained.
Only after the arithmetic closes should the team compare explanations. A genuine shift towards lower-margin, slower-paying customers would predict concentrated changes in margin, payment terms and receivables for that population. A promotion, revenue-recognition or data-classification timing explanation would leave different evidence: promotional batches, adjustments or corrections should account for most of the movement. A correlation can suggest where to look. It cannot decide between those explanations by itself.
If the evidence crosses the threshold set for the decision, the commercial owner can adjust incentives or payment terms for the affected group while finance checks the usable-cash boundary. Customer-level margin, collection days and renewal quality then become the monitoring set. Where the evidence has not crossed that threshold, the work remains a test rather than a recommendation.
Treat the metric library as an index
Finance and FP&A should make existing metric families easy to find. For the marketplace decision, that means planning and forecast measures, actual-versus-comparator and variance measures, cash and liquidity measures, profitability and margin measures, and KPI-governance measures.
Capital allocation and investment-return measures may matter elsewhere. Adding them here would make the article look more complete without improving this decision. Each formula and authoritative definition should remain with the domain that owns it. The library helps the analyst reach the right definition and owner.
Write the decision packet before changing the offer
For the marketplace, the next hand-off can be written directly from the work above.
| Field | Current case |
|---|---|
| Decision | Should a defined customer group receive different incentives or payment terms? |
| Fact base and owner | Finance traces revenue, receivables and usable cash; the commercial team tests the customer and promotion mechanisms. |
| Evidence contract | Customer population, commission rate, margin, payment days, receivables, period, currency and data version. |
| Comparator and boundary | Actual against the latest forecast; no change that breaches management’s usable-cash boundary. |
| Work still required | Complete the volume–rate–mix–timing bridge and examine concentration, promotion batches and recognition or classification adjustments. |
| Trigger | The commercial owner decides once the agreed evidence threshold is met; finance confirms the cash effect. |
| Guardrails and open items | Monitor customer-level margin, collection days and renewal quality. Keep any residual on the follow-up list. |
The next conflicting dashboard can start from the same packet. It tells the team which figures need tracing, which claim still needs a test and who is allowed to change the commercial terms.