This study develops and calibrates a revenue accruals model. Changes in accounts receivable and deferred revenues are modelled using the respective income statement and cash flow numbers (i.e., revenues and cash flows from sales) that relate directly to the accruals’ origination and reversal. Compared to existing models, the proposed specification explains more variation in the data and, in simulations with seeded revenue manipulation, exhibits greater detection power and less bias. Furthermore, the abnormal revenue estimates are positively associated with cases of revenue misstatements identified by the Securities and Exchange Commission. Results imply that researchers, auditors, and regulators interested to detect earnings management should focus on modelling specific accruals. As a practical matter, to detect revenue management, they should consider broadening their scope to examine not only accounts receivable but also current and long‐term deferred revenues.