Nobody wrote down what ARR means
Does it include usage overage? One-off professional services? A customer in their notice period? Three teams answer differently, each defensibly, and the numbers diverge by the time they reach the board.
Chargebee · Reporting & Revenue Analytics
Three teams, three ARR numbers, and a board meeting in an hour.
Subscription metrics are not hard to calculate. They are hard to agree on — because nobody wrote down whether a downgrade counts as churn, whether usage overage is recurring, or which day a renewal lands. We govern the definitions, model the data once, and deliver reporting the whole company reads the same way.
12+
500+
40+
250+
Trusted by 500+ organizations — including SaaS, subscription and usage-based businesses running their quote-to-revenue operations on Salesforce, HubSpot and Chargebee with Twopir Consulting.
Reporting Coverage
- Salesforce Partner
- HubSpot Partner
- MRR & ARR
- Net Revenue Retention
- Cohort Analysis
- RevenueStory
- Warehouse & BI
- Board Reporting
Why subscription reporting stops being believed
Almost none of these are calculation errors. They are definition failures, and a dashboard cannot fix a definition nobody agreed.
Retention is reported as one number
Net revenue retention hides the mechanism. Without expansion, contraction, churn and reactivation separated, a healthy NRR can conceal a serious downgrade problem masked by a few large upsells.
Reporting is assembled by hand
An analyst exports, pivots, reconciles and formats for three days a month. The output is late, unreproducible, and stops entirely when that person is on leave.
Billing and CRM report different customer counts
Different definitions of active, different treatment of trials and parent-child accounts, and no reconciliation — so pipeline reporting and revenue reporting cannot be laid side by side.
Usage revenue is not modelled
Hybrid pricing means a share of revenue is consumption-based and volatile. Reporting built for flat subscriptions treats it as noise instead of forecasting it.
Recognised revenue and billed revenue get confused
Invoiced, collected and recognised are three different numbers with three different timings. When a report does not say which one it is showing, someone will use the wrong one in a fundraise.
The definitions that have to be written down
This is the first deliverable of every reporting engagement we run: a metric dictionary, agreed by finance, signed off, and implemented once. The table below is the starting template, not the answer — the decisions column is where your business differs.
| Metric | What it measures | The decisions that must be made |
|---|---|---|
| MRR | Normalised recurring revenue per month across the active base. | Does usage overage count? Annual plans divided by twelve or recognised monthly? Are discounts netted? |
| ARR | Annualised run rate of recurring revenue at a point in time. | MRR times twelve, or contracted annual value? Are non-recurring charges and services excluded? |
| New, expansion, contraction, churn | The four movements that reconcile last period's MRR to this period's. | Is a plan downgrade contraction or partial churn? Where does a currency movement land? When is a pause churn? |
| Net revenue retention | Revenue from a cohort this period against the same cohort a year ago. | Cohort by first payment or contract start? Are reactivations included? Which customers count as a cohort at all? |
| Gross revenue retention | The same cohort excluding expansion — the floor under the business. | Same cohort definition as NRR, or a separate one? Consistency matters more than the choice. |
| Customer churn | Logo count lost over the period. | Is a parent with five child accounts one logo or six? When does a non-renewal become churn — notice date or end date? |
| Involuntary churn | Cancellations caused by payment failure rather than intent. | How long after the dunning sequence ends does it count? Are recoveries back-dated or treated as reactivation? |
| Billed vs collected vs recognised | Three different numbers at three different moments in the cycle. | Which one leads the board pack, and does every report label which it is showing? |
| Committed vs consumed usage | For hybrid models, what was contracted against what was actually used. | Is unconsumed commitment revenue? When does it expire, and does it roll over? |
What These Numbers Feed
Once the definitions are governed, they drive Chargebee revenue operations and AI — the retention, recovery and expansion motions Chargebee optimization services — and the audit that quantifies where revenue leaks Chargebee consulting services — the Chargebee practice these sit inside
Native, warehouse, or both with a boundary
The wrong answer here is expensive in both directions — a warehouse project for reporting the platform already does, or a native dashboard asked to join data it cannot see.
Report inside Chargebee
RevenueStory gives subscription analytics across revenue, signups, activations and churn without engineering. For a single-product business whose questions are about the billing data itself, this is the right and cheapest answer.
- No pipeline to build or maintain
- Subscription metrics available immediately
- Reveal for payment performance diagnostics
- Limited to data Chargebee holds
- Our default for early and single-product teams
Model it in Snowflake or BigQuery
When the question spans billing, CRM, product usage and the ledger, the join has to happen somewhere that can see all four. Metrics are defined once in the modelled layer and every tool reads from it.
- Billing, CRM, product and ledger joined
- Definitions implemented once, centrally
- Cohort and usage forecasting possible
- A pipeline and a modelling skill to maintain
- Our default once usage pricing is material
Native for operations, warehouse for truth
Most mid-market businesses land here. Operators use native reporting for day-to-day billing questions; the board pack, retention analysis and anything cross-system comes from the modelled layer — with the boundary written down.
- Operational questions answered natively
- Governed metrics from the warehouse
- One documented boundary between them
- Reconciliation between the two, scheduled
- Our most common recommendation
Reporting work in the order it should happen
The sequence matters. Building dashboards before governing definitions produces attractive disagreement.
Metric Dictionary & Governance
The written definition of every revenue metric, agreed with finance and signed off. Always the first deliverable, and regularly the one that ends the longest-running argument in the company.
- Definitions for MRR, ARR and the four movements
- Retention cohort rules, written down
- Treatment of usage, services and discounts
- Owner and change process per metric
- Sign-off by finance before implementation
Revenue Data Model
The modelled layer where the definitions are implemented once, so every dashboard and export is reading the same computation rather than its own.
- Billing, CRM, usage and ledger joined
- Movement and cohort tables built once
- Slowly-changing dimensions for plan history
- Currency and entity normalisation
- Tested against a known period
Operational Dashboards
Day-to-day reporting for the people who act on it — billing operations, CS, sales leadership — not a board pack that has been shrunk.
- Subscription and MRR movement views
- Renewal and at-risk pipelines
- Failed payment and recovery queues
- AR aging and collections status
- Usage against entitlement by account
Board & Investor Reporting
The pack assembled by the system and reviewed by a human, with every number traceable to a governed definition and a source.
- ARR bridge with the four movements
- Net and gross revenue retention by cohort
- Churn analysis split voluntary and involuntary
- Committed versus consumed usage
- Traceability from figure to source
Payment & Usage Analytics
The two areas most subscription reporting ignores, and where the operational levers actually are.
- Authorisation and decline rates by segment
- Recovery performance by failure type
- Usage distribution and overage patterns
- Entitlement consumption and expansion signals
- Forecasting for consumption revenue
Exports, Alerts & Delivery
Reporting nobody opens is not reporting. Getting the number to the person at the moment they can act on it is part of the build.
- Scheduled exports to finance and the ledger
- Threshold alerting on the metrics that move
- Reports embedded in the CRM record
- Warehouse feeds for downstream tools
- Access and permission model
The sources a revenue number actually depends on
An ARR figure that only reads the billing system is not wrong so much as incomplete. These are the feeds, and what each one contributes.
Chargebee billing data
Subscriptions, plan and price-point history, invoices, credit notes and payments — the base for MRR, ARR and the four movements. Plan history matters as much as current state: without it, you cannot explain why last quarter's number changed.
Salesforce pipeline
Open opportunities, renewal pipeline and forecast, joined to billing so committed revenue and pipeline sit in the same view. Without the join, forecast and actuals are compared by hand in a spreadsheet each month.
HubSpot lifecycle data
Lifecycle stage, campaign attribution and CS engagement joined to subscription revenue, so acquisition cost and expansion can be attributed to the motion that produced them.
Product usage events
Consumption against entitlement, which for hybrid pricing is both a revenue driver and the strongest expansion and churn signal available — usually earlier than anything in the CRM.
Payment and gateway data
Authorisation rates, decline reasons and recovery outcomes. Chargebee Reveal reports these at transaction and order level across gateways, which separates involuntary churn from real churn.
RevRec and the general ledger
Recognised revenue and journal entries, so billed, collected and recognised can be reconciled and reported as the three distinct numbers they are rather than used interchangeably.
Definitions first, dashboards last
Five stages. Reversing the first two is the most common and most expensive mistake in revenue reporting projects.
Metric Workshop
We put finance, RevOps and leadership in a room and force the definitions to be decided. Most of the value of the engagement is created here.
Source Audit
We trace each metric to the systems and fields it depends on, and find where the data does not actually support the definition that was just agreed.
Model & Implement
Definitions implemented once in the modelled layer or configured natively, then tested against a period whose real numbers are already known.
Deliver & Train
Dashboards, exports and the board pack, plus the walkthrough that explains what each figure includes — because a number without its definition gets misused.
Govern
An owner per metric and a change process, so the definitions survive the next pricing model rather than silently drifting again.
Reporting engagements from the same practice
Getting one operating truth out of several systems is what most Twopir engagements are ultimately for.
Sales Operations & Lead Management
Pipeline and lead reporting rebuilt on governed definitions so leadership read one operating picture instead of three.
- Pipeline and stage model redesign
- Data quality at the point of capture
- Operating dashboards for leadership
- One reporting definition across teams
Accounting Seed & Salesforce Integration
Financial and operational reporting brought onto one record, with reconciliation designed into the close rather than performed after it.
- Financial reporting on the CRM record
- Bi-directional data synchronisation
- Reconciliation built into month end
- Shared reporting across ops and finance
Salesforce & HubSpot Integration
Two platforms reconciled so marketing, sales and revenue reporting could finally be laid side by side without a manual mapping step.
- Cross-platform object and field mapping
- Lifecycle stage alignment
- Attribution data joined to revenue
- One reporting view across both systems
We fix the definitions before we build the dashboard
We help growing and mid-market companies solve complex CRM, integration and business system challenges, and we work with enterprise organizations on the same reporting problems across more entities and currencies.
The metric dictionary is the first deliverable
Not the last. Every reporting engagement starts by forcing the definitions to be decided and signed off by finance, because a correctly computed metric that nobody agreed the meaning of is still an argument.
We read billing and CRM as one dataset
As a Salesforce Partner and HubSpot Partner we build both ends, so pipeline, attribution and subscription revenue can be joined properly instead of compared side by side in a spreadsheet.
We separate billed, collected and recognised
Three different numbers at three different moments. Every report we build says which one it is showing, because using the wrong one in a fundraise is a genuinely expensive mistake.
We recommend native reporting when it is enough
Not every business needs a warehouse. If RevenueStory answers your questions we will say so and configure it rather than selling you a data platform project you will have to staff.
We build the governance, not just the report
An owner per metric and a change process, so the definitions survive the next pricing model. Reporting projects fail on the second pricing change, not the first launch.
Reporting questions worth settling early
Chargebee's native analytics, RevenueStory, covers subscription reporting across revenue, subscriptions, signups, activations and churn without engineering work, and Reveal reports payment performance — authorisation and decline rates segmented by gateway, currency and geography. That is genuinely enough for many single-product businesses. It becomes insufficient when a question needs data Chargebee does not hold, such as CRM pipeline, product usage outside billing, or ledger detail.
Almost always because the definition was never written down rather than because a calculation is wrong. Whether usage overage counts as recurring, whether an annual plan is divided by twelve, whether a customer in their notice period is still active, and how parent-child accounts are counted are all defensible choices — and teams make them differently. The fix is a governed metric dictionary, signed off by finance, implemented once.
Not always. If your questions are about the billing data itself, native reporting is cheaper and faster and has no pipeline to maintain. A warehouse earns its place when the question spans billing, CRM, product usage and the ledger, because the join has to happen somewhere that can see all four — which in practice means once usage-based pricing becomes material to revenue.
Revenue from a defined cohort in the current period against the same cohort's revenue a year earlier, including expansion and contraction within that cohort. The decisions that matter are the cohort rule — first payment or contract start — whether reactivations are included, and whether gross retention uses the same cohort. Consistency matters far more than which convention you pick, provided the choice is documented and applied everywhere.
Yes, and it needs modelling rather than a chart. Consumption revenue is volatile, so reporting has to separate committed from consumed, show the distribution rather than just the average, and treat unconsumed commitment explicitly — including whether it rolls over and when it expires. Usage against entitlement is also the earliest expansion and churn signal most subscription businesses have.
The metric workshop and dictionary run one to two weeks and deliver value immediately. Native configuration adds two to three weeks. A modelled warehouse layer with billing, CRM, usage and ledger joined is typically six to twelve weeks depending on source-data quality, which is nearly always the constraint.
Your team. The metric dictionary names an owner and a change process per metric, the modelled layer is documented and tested, and we walk the team through what every figure includes. Reporting that only the consultancy can explain will be quietly replaced by a spreadsheet within two quarters.
Next Step
Agree the definitions, and the numbers stop being an argument
We will run the metric workshop, write the dictionary, model the data once and deliver reporting the whole company reads the same way — across Chargebee, the CRM, your product and the ledger.
The metric dictionary is the first deliverable, not the last
