Online coaching platform — ASC 606 in under a quarter
Multi-product subscription catalog across coaching programs and digital products, multiple payment processors
Function
Context
Engagement
Single projectThe challenge
The business had been running on cash-basis recognition long past the point where it could stay that way. Two payment processors were feeding the same customer record in different formats, with one of them silently overstating transaction amounts because of a units bug nobody had caught. A diagnostic of the recognition data layer turned up a significant volume of payment events with no subscription match — cash that had never entered the deferred revenue schedule — alongside subscriptions carrying collected cash that had gone missing from the engine entirely. The leadership team could not produce clean revenue or close the books on time. They could not answer the questions the next round of capital would ask.
System architecture
Input
Payment Processor A
Input
Payment Processor B
Validation
Data reconciliation & quality validation
Both processors treated as untrusted until proven
ASC 606
Bespoke revenue recognition system
Methodology aligned with CPA and tax professional
Coaching subscriptions
Deferred linearly over service period
Digital products
Recognized at point of sale
The approach
The engagement was structured as a leadership and architecture role. The accounting requirements — what ASC 606 demands of a multi-product subscription catalog, how coaching subscriptions differ from digital product sales, what "defensible" means to a next-round auditor — came from this side of the table. So did the data architecture: how the bespoke rev rec system needed to be structured, which exceptions to remediate first, how to rank the data quality work by materiality rather than by what was easiest to fix.
Before any build, the methodology was aligned with the client's CPA and tax professional. Recognition timing, deferral treatment, and the handling of the two product types were confirmed across all three parties. The system was built to a methodology everyone had signed off on — not a reasonable interpretation that would need to be relitigated later.
AI aided the execution of the data build. The diagnostic of the recognition data layer, the unit test pack, the remediation work — analytical tasks that would have taken a team weeks to scope and run moved in days. The governance held because the accounting logic and the architecture had a human owner throughout — reviewing every finding, directing what came next, and signing off on what was correct before it moved forward.
Two payment processors fed the system. Both were treated as untrusted sources until the data quality was proven. A units error in one of the feeds — silently overstating transaction amounts — was identified in the diagnostic, corrected, and reconciled across the full payment history before the system was validated.
AI aided the execution of the data build. The accounting judgment, the architecture, and the methodology alignment were mine. That is what compressed the timeline without compressing the standard.
The outcome
An AI-assisted program stood up an ASC 606-compliant recognition engine in under a quarter. The engine traces every dollar from contract through daily and monthly revenue schedules, with coaching subscriptions deferring cleanly and digital products recognizing at sale. The data quality exceptions were diagnosed systematically — the units bug identified, corrected, and reconciled across the full payment history. The engine was validated against a structured unit test pack covering all subscription strata and passed clean. A significant volume of previously collected cash is now correctly deferred and scheduled into future periods. The team closed the quarter with a defensible path off cash basis and the financial reporting the next stage of growth requires.
Timeline
Under a quarter
Full delivery: diagnostic, remediation, and validation
Standard
ASC 606
Bespoke system built to methodology agreed across CFO, CPA, and tax professional
Method
AI-aided
Data build accelerated without compressing the accounting standard
Working on something similar?
Get in touch