Commercial diligence: testing revenue quality IDER in-house training materials Hypothetical cases and figures for teaching the methods. Learning objectives Reconcile reported revenue, corrected recognition, recurring activity and cash collections. Calculate cohort retention without contaminating the denominator with new customers. Design evidence tests that distinguish commercial risk from proven accounting error. Basic revenue, receivables and percentage calculations. Is reported growth commercially durable? Hypothetical case: GHS 12 million reported revenue Recognition, recurrence and cash are different tests Deliverable: evidence-linked revenue bridge A business reports twelve million Ghana cedis of revenue and strong growth. Your assignment is to determine what that figure says about durable commercial performance. We will not begin by applying a convenient valuation multiple. We will reconstruct what was earned, identify which activity is plausibly repeatable, examine customer retention and trace the route to cash. Every figure in this IDER in-house clinic is hypothetical. Keep separate columns for accounting corrections and commercial adjustments. That separation prevents a legitimate one-off sale from being described as an accounting error, and prevents a recurring invoice from being accepted without evidence of delivery. By the end, you will have a practical investigation sequence and a defensible explanation of unresolved risks. Build four distinct revenue views Reported → correctly recognised Correctly recognised → recurring analytical base Earned revenue → customer cash collections Four numbers can all be different without any contradiction. Reported revenue is the starting claim. Correctly recognised revenue reflects the appropriate treatment of actual contractual delivery. A recurring analytical base removes activities you do not expect to repeat under an explicit definition. Customer collections measure cash received, including amounts earned in earlier periods. Ask which question each number answers before comparing them. A signed order may support future demand but does not automatically establish current earned revenue. Equally, receiving an advance does not prove that a service has already been delivered. Build a dictionary that specifies period, currency, tax treatment and inclusion rules for each view. Most avoidable disputes begin with those definitions being left implicit. Create a transaction spine before testing totals Stable customer, contract and invoice identifiers Service dates, credits and collection references Reconcile extraction totals to the ledger Request a transaction-level extract that can connect the commercial and financial systems. You need stable customer identifiers, contract references, invoice and credit-note identifiers, service periods and payment links. Keep the original extract immutable and perform transformations in a reproducible working layer. Before analysing cohorts, reconcile row counts and totals to the ledger for the same period and scope. Investigate missing identifiers and duplicate invoice lines. A many-to-many join between invoices and payments can multiply revenue while appearing technically successful. Test the expected relationship before joining and retain unmatched records for review. This transaction spine is your audit trail: every attractive chart should be traceable back to an identifiable set of records and documented transformations. Follow the contract through actual delivery Contract terms and performance evidence Invoice timing and period-end credits Document exceptions; do not infer from labels Choose transactions around the reporting boundary and follow the evidence chain. Read the relevant contract terms, identify what was promised and inspect evidence that the promised delivery occurred in the reported period. Invoice date alone is insufficient. A subsequent credit note may reveal a cancellation or correction, but it needs interpretation rather than automatic subtraction. For bundled services, establish whether the recorded allocation and timing reflect the actual obligations; escalate technical accounting judgments to the responsible professional. Your research role is to surface and organise the evidence, quantify clearly supported exceptions and distinguish them from open questions. Preserve the original amount, proposed adjustment, reason, supporting documents and reviewer status for every exception you report. Correct recognition before judging recurrence Reported: GHS 12.0m Less wrong-period 0.6m and reversal 0.4m → 11.0m Less valid one-off 1.0m → recurring base 10.0m 12.0 − 0.6 − 0.4 = 11.0; 11.0 − 1.0 = 10.0 (GHS millions) Our hypothetical review identifies six hundred thousand cedis recorded in the wrong period and a separate four hundred thousand that must be reversed. Corrected recognised revenue is therefore eleven million. Within that corrected amount, one million comes from a valid one-off project. Excluding it produces a ten-million recurring analytical base. Notice the two bridges. The first corrects recognition; the second changes the lens for forecasting. The one-off project remains valid recognised revenue, and removing it does not create an accounting correction. State the evidence supporting its non-recurrence and check whether similar projects regularly replace one another. A business built on successive projects may have repeatable commercial activity even when individual contracts are not recurring subscriptions. Freeze the opening population before measuring retention Opening monthly recurring revenue: GHS 100,000 Churn 10,000; contraction 5,000; expansion 20,000 New-customer revenue 30,000 is separate Retention asks what happened to the customers you already had. Freeze the opening cohort and its revenue, then classify each customer's change using consistent rules. Our monthly opening revenue is one hundred thousand cedis. Existing customers lose ten thousand through complete churn and five thousand through reduced spending, while expansion adds twenty thousand. New customers contribute a separate thirty thousand. Keep new customers outside the retention numerator because they were absent from the opening denominator. Track reactivations and mergers under a documented policy so that identifiers cannot manufacture retention. Also align currency and billing periods. A price increase or exchange-rate movement can lift revenue retention even when usage or customer counts weaken, so investigate the mechanism behind the movement. Calculate gross and net revenue retention Opening 100,000; churn 10,000; contraction 5,000 Expansion 20,000; new customers 30,000 Find gross retention, net retention and closing run-rate Pause and calculate three measures from the monthly cohort. First find gross revenue retention, which deducts churn and contraction from the opening base while excluding expansion. Then find net revenue retention, which includes expansion by those same opening customers. Finally calculate the total closing monthly revenue run-rate, now including the new customers. Write the population and period beside each answer. Explain why a closing run-rate should not be described as revenue earned throughout the entire year. A useful way to check your calculation is to build a bridge from the opening amount to the closing amount before dividing anything. That bridge exposes missing categories and helps distinguish a retention improvement from growth purchased through new-customer acquisition. Expansion can conceal meaningful attrition Gross retention = 85% Net retention = 105% Closing monthly run-rate = GHS 135,000 GRR = (100−10−5)/100 = 85%; NRR = (100−10−5+20)/100 = 105% Gross retention is eighty-five thousand divided by one hundred thousand, or eighty-five percent. Add twenty thousand expansion from existing customers and net retention becomes one hundred and five percent. Including thirty thousand from new customers gives a closing monthly run-rate of one hundred and thirty-five thousand. Net retention above one hundred percent is compatible with losing customers and revenue elsewhere in the opening cohort. That is why you should inspect gross retention and the distribution of expansion, not celebrate the net figure alone. If one large account supplies nearly all expansion, the apparent strength may be concentrated. And multiplying the closing run-rate by twelve creates an annualised snapshot, not a reconstruction of what the business actually earned during the year. Separate price, usage and customer mix Track contract price and delivered quantities Compare cohorts at equal maturity Investigate concentration at group level To interpret growth, decompose it into customer counts, usage and realised price wherever the business model permits. A price increase can temporarily improve revenue while future renewal risk rises. Compare acquisition cohorts at the same age, because a mature cohort and a newly acquired cohort have had different opportunities to churn. Include lost customers rather than analysing only surviving accounts. Then examine concentration by ultimate commercial group, not only by billing identifier. Ten subsidiaries under one purchasing decision may behave like a single exposure. Interview selected retained and lost customers using a documented selection process. Those interviews can explain mechanisms, but a small convenient sample does not estimate population-wide churn reasons with statistical precision. Bridge earned revenue to customer cash Opening receivables GHS 1.2m Corrected credit revenue 11.0m; closing receivables 2.0m Collections = GHS 10.2m under stated assumptions Collections = 1.2 + 11.0 − 2.0 = GHS 10.2m Assume the eleven million corrected revenue consists entirely of credit sales and uses the same scope as receivables. Opening receivables are one point two million and closing receivables two million. With no other movements, collections equal opening receivables plus credit revenue minus closing receivables, giving ten point two million. The assumption matters. Taxes, foreign exchange, write-offs, reclassifications, unbilled balances or factoring can require separate bridge lines. Do not insert a balancing plug and call it cash. Reconcile the inferred collection amount to matched bank receipts and investigate differences. This exercise tests consistency among records; it does not prove every sale was genuine or every remaining debtor will pay. Those claims require their own evidence and review. Follow unpaid invoices through subsequent events Age from contractual due date Inspect disputes, credits and payment allocation Distinguish cash received from cash applied An ageing schedule becomes useful when you understand what starts its clock. Age invoices from the contractual due date when assessing lateness, and retain invoice age separately if needed. Trace large overdue items to correspondence, disputes, credit notes and subsequent payments. Check that the cash was applied to the same invoice rather than merely received from the same customer. An advance for a new contract does not settle an old disputed invoice unless the allocation is valid. Analyse ageing by customer and cohort as well as in aggregate; new growth can dilute an old delinquency problem. Record the review date so that subsequent collections are not accidentally presented as cash available at the historical reporting date. Use risk-based tests and preserve their limits Test material and unusual transactions Add a documented representative sample Report unresolved scope and quantified exposure A focused review should cover individually material accounts and transactions with unusual timing, terms or adjustments. Add a documented representative sample if you need evidence about the wider population. These purposes differ. A deliberately high-risk sample may discover problems efficiently, but its exception rate cannot be mechanically extrapolated to all revenue as though the sample were random. Retain the sampling frame, selection method and tested amount. Distinguish confirmed adjustments, plausible exposure ranges and matters awaiting evidence. Where several tests identify the same invoice, count the underlying exposure once. The purpose is an interpretable risk assessment, not a dramatic sum of overlapping flags. Ask what additional evidence could resolve each material uncertainty before recommending a transaction condition. Does higher receivables prove poor revenue quality? Receivables rose from GHS 1.2m to 2.0m Corrected credit revenue was GHS 11.0m Calculate collections and propose three follow-up tests Pause and assess a colleague's claim: receivables increased by eight hundred thousand cedis, therefore eight hundred thousand of revenue is fraudulent. First calculate collections under the clean bridge assumptions. Then explain why the claim goes beyond the available evidence. Propose three follow-up tests that could distinguish normal commercial timing from collection or recognition problems. Your answer should connect each test to a competing explanation, such as longer contractual terms, a late-period sales increase or a disputed delivery. Do not merely request more documents. Specify which records would change your conclusion and how you would reconcile them. This is the central discipline of diligence: move from a signal to a testable explanation before moving from an explanation to a finding. A receivables increase is a signal to investigate Implied collections: GHS 10.2m Test terms, ageing and matched subsequent receipts Review disputes, cutoff and other bridge movements The clean bridge gives ten point two million cedis of collections. An eight-hundred-thousand increase in receivables says that recognised credit sales exceeded collections by that amount under our assumptions. It does not identify the cause. Compare agreed payment terms and sales timing, inspect overdue ageing, and trace subsequent receipts to the actual invoices. Review disputed deliveries and period-end credits where recognition may be affected. Finally confirm that tax, currency and other balance movements have been treated consistently. Report what each test establishes. Subsequent payment supports collectibility of the matched invoice, but it does not automatically resolve whether the sale belonged in the original reporting period. Keep those conclusions separate even when the same document informs both investigations. Explain the finding, evidence and decision consequence Separate corrections from analytical adjustments Show cohort, concentration and cash evidence Attach owners and deadlines to open conditions Build the conclusion around the decision, not around a catalogue of documents reviewed. Present the reported-to-corrected revenue bridge, then the separate recurring-base bridge. Show gross and net retention for consistent cohorts, highlight concentration, and reconcile collections with clearly stated limitations. For each material finding, explain the amount, mechanism, evidence and commercial consequence. An unresolved customer renewal may affect a forecast scenario; a confirmed cutoff error affects historical recognition. They need different responses. Identify the professional responsible for technical sign-off and retain the actual review status rather than implying approval. Attach owners and deadlines to outstanding evidence requests. A concise conditional conclusion is more useful than a confident headline that hides an unfinished investigation inside an appendix. Reproduce the bridge before trusting the headline Audit identifiers and period definitions Recalculate retention on the opening cohort Complete the assessment and evidence-plan tasks For your next assignment, begin with a small reproducible slice of the data. Select one reporting month, reconcile it to the ledger, trace a few transactions through delivery and payment, and rebuild one customer cohort. This pilot exposes definition and join problems before they contaminate a much larger analysis. Then expand using documented rules and risk-based review. Complete the assessment below and the two applied tasks. Your work should show that a recurring label, a strong growth rate and a cash receipt answer different questions. Assess yourself on whether another practitioner can follow the calculation and identify exactly what remains unproven. The practical outcome is a decision supported by evidence, with commercial uncertainty expressed clearly enough to manage. Check your understanding 1. After removing GHS 0.6m wrong-period revenue and GHS 0.4m reversal from GHS 12m, corrected recognition is: A. GHS 10m B. GHS 11m C. GHS 11.4m D. GHS 12m 2. A valid GHS 1m one-off project excluded from the recurring base is: A. Automatically fraudulent B. An accounting error because it is unusual C. A commercial analytical adjustment, not necessarily an accounting correction D. Always recurring if already paid 3. What is gross revenue retention in the case? A. 85% B. 105% C. 135% D. 95% 4. What is net revenue retention? A. 135% B. 115% C. 85% D. 105% 5. With no other movements, opening AR 1.2m + credit revenue 11m − closing AR 2m implies: A. GHS 13m collections B. GHS 10.2m collections C. GHS 11.8m collections D. GHS 0.8m fraudulent sales 6. A risk-selected sample has a 20% exception rate. Which conclusion is defensible? A. Exactly 20% of all revenue is misstated B. No more testing is needed C. The selected exceptions require investigation; population extrapolation needs a justified sampling design D. All untested invoices are valid Applied task 1 Prepare two revenue bridges and a cohort dashboard using the case. Clearly label period, population and accounting versus commercial adjustments. Applied task 2 Design an evidence plan for the GHS 0.8m receivables increase. Include three competing explanations, targeted records and a conclusion limit for each. Answers and explanations 1. B. 12−0.6−0.4=11 million. The one-off exclusion is a separate analytical step. (Slide 5) 2. C. A valid sale remains recognised; recurrence is a separate forecasting definition. (Slide 5) 3. A. (100−10−5)/100=85%; expansion and new customers are excluded. (Slide 8) 4. D. (100−10−5+20)/100=105%. New-customer revenue is not part of the opening cohort. (Slide 8) 5. B. The receivables roll-forward gives collections of 10.2m; it does not establish fraud. (Slide 10) 6. C. A targeted high-risk sample is not automatically representative of the full population. (Slide 12) Applied task 1 Recognition: 12.0−0.6−0.4=11.0m. Recurring analytical base: 11.0−1.0=10.0m. Opening cohort monthly GRR=85%, NRR=105%; closing monthly run-rate including new customers=135,000. Annualising that snapshot does not reconstruct annual earned revenue. Two separate bridges with no double counting Correct opening-cohort retention denominators Explicit period and run-rate versus earned-revenue distinction Applied task 2 Compare contractual terms and sales timing; review due-date ageing and disputes; match subsequent bank receipts to invoices. Reconcile taxes, currency, write-offs and other movements before relying on the 10.2m clean bridge. Subsequent payment supports collection but does not independently prove correct original-period recognition. At least three explanations linked to specific tests Cash allocation and bridge-scope controls Explicit distinction between signal, confirmed error and unresolved exposure References IFRS Foundation — IFRS 15 overview https://www.ifrs.org/issued-standards/list-of-standards/ifrs-15-revenue-from-contracts-with-customers/ Revenue recognition based on delivery of promised goods or services; the clinic’s recurring-revenue adjustments are separate original analytical examples. US SEC — Beginners’ Guide to Financial Statements https://www.sec.gov/investor/pubs/begfinstmtguide.htm Distinction between income, balance-sheet receivables and cash flows; original worked reconciliation and diligence procedures.