SaaS due diligence: reconcile ARR without confusing it with revenue

Annual Recurring Revenue generally describes an annualization of recurring revenue on a given date, according to a definition that must be specified. I would start with this definition: active contracts, contracted or invoiced revenue, rebate processing, usage, currencies and services. There is no formula that allows you to compare all publishers without adjustment. The ARR does not become accounting turnover because it is calculated from invoices.

QoE examines the outcome and its drivers. ARR is useful for understanding the business basis, but it must be linked to service costs, margin and customer loyalty. An increase in ARR may be accompanied by a deterioration in cash or significant discounts. I would keep separate bridges for ARR, accounting income and EBITDA, with links explained.

CRM, billing and accounting can tell different things without one source necessarily being false. The CRM tracks commercial opportunities and dates; invoices follow invoicing; accounts apply recognition rules. I would build a customer and contract accounting framework that allows discrepancies to be found. Duplicates, migrations and changes of identifiers must be documented.

Expired contracts, expected renewals and suspended billings deserve explicit treatment. A commercial intention does not constitute an active contract. Conversely, a late invoice does not prove churn. You must read the termination clauses, the status of the service and the dates chosen. The scope at the closing date must remain reproducible.

An ARR bridge can separate new customers, expansion, contraction and churn. The conventions must be stable: is a reactivated customer new or existing, is a currency change isolated, is an acquisition included? Gross and net retention metrics require a comparable data set. Growth due to the expansion of a few large customers can mask broader attrition.

I would supplement the metrics by focus, discounts, contract duration and service commitments. Multi-year commitments do not guarantee collections if the clauses allow an exit. The commercial analysis must meet the reality of the cost of service: infrastructure, support, onboarding and possible partners. Margin is an essential bridge to QoE.

A presentation announces 4,8 M€ from ARR. The professional review identifies 300 k€ non-recurring implementation services and 200 k€ expired contracts which do not meet the definition adopted. Under these assumptions, the corrected ARR is 4,3 M€. I would check that the two categories are separate and that the amounts are annualized in a comparable manner.

This restatement is not mechanically deducted from EBITDA. Benefits may have generated legitimate income and costs in the historical period. Expired contracts may contribute to the bottom line before they expire. Their effect on trading and prospects must be explained without erasing the past. A commercial bridge and an accounting bridge must therefore remain separate.

The case also raises questions: will the contracts be renewed, what implementation income usually accompanies sales, what margins are associated? These answers influence the investment, but they do not retrospectively change the definition of ARR on the date tested. The report must distinguish between established fact and scenario.

I would target the gaps between commercial growth, recognized income and collections. Acceleration of contracts with large discounts may produce a temporary effect on invoices. Free expansion can mask economic contraction. Commercial credits and early renewals must be placed back in the contracts. The precise questions allow us to better understand commercial arbitrations.

The restitution must retain definitions, populations, calculations and limits. Comparable cohorts and periods are particularly important if the data has been migrated. A clean series after migration is not necessarily comparable to historical data. I would present the ruptures rather than silently smooth them over.

AI can extract clauses, associate contracts and suggest classifications. I would test expired contracts, amendments and offers combining subscriptions and services. The system must be able to flag ambiguity rather than force a category. The calculation must remain auditable with identified input data.

For Solenor, as for other tools, value is measured by the net accounting reconciliation and professional review time. A summary of growth is not enough. The team must be able to defend the definition used and show how it relates, or does not, to the financial results.