ProcessGrowth, Marketing & Sales
How to set up revenue operations, one foundation at a time
Revenue operations is less a team than a set of agreements: what each funnel stage means, which system owns each record, how accounts pass between teams and which numbers the business runs on. This guide sequences those agreements into seven steps, shows what each one produces, and explains where AI fits only once the foundations are in place.
On this page
- Symptoms that tell you revenue operations is overdue
- The seven steps to a working revenue operations function
- How an account moves between teams once handoffs are defined
- Choosing the system of record for each revenue object
- Revenue metrics to define before building any dashboard
- Where AI fits once the foundations exist
- Questions and answers
Symptoms that tell you revenue operations is overdue
Two or three of these together usually mean the problem is structural rather than a matter of effort.
The seven steps to a working revenue operations function
Agree funnel definitions and stage exit criteria in writing
Define each lifecycle stage for leads, accounts and opportunities with an observable exit criterion: what must be true, and recorded, before a record moves on. Marketing, sales, customer success and finance sign the same document, and it becomes the reference for every report built afterwards.
Build one account and contact data model
Decide how accounts, contacts, parent companies and buying groups relate, which fields are mandatory, how duplicates are detected and merged, and who owns an account when it moves from prospect to customer. Ownership rules prevent the turf disputes that otherwise follow every territory change.
Name systems of record and wire the integrations
Assign each object to the system that owns it and define which fields flow where, in which direction and how often. Integrations between CRM, marketing automation, billing and support are built against that map rather than tool by tool.
Set handoff service levels between teams
Each handoff gets a time limit, a required payload and a way back: sales accepts or rejects a marketing lead with a reason, customer success receives the deal context before kickoff, and finance receives contract terms at booking. Rejection reasons are reviewed jointly, because they are the cheapest feedback marketing will ever get.
Build a metrics tree from revenue down to activity
Start from bookings and net revenue retention and decompose them into the rates and volumes that drive them: pipeline created, stage conversion, win rate, deal size, cycle length, gross churn and expansion. Each metric gets one definition, one owner and one source.
Set forecasting and pipeline hygiene rules
Choose a forecast method, define forecast categories and set hygiene rules such as close dates in the past, missing next steps or stages that conflict with recorded activity. Weekly reviews look at exceptions rather than walking every deal.
Write the team charter and change-request process
Settle who revenue operations reports to, what it owns and what it does not, and how anyone requests a new field, workflow or report. A short intake with impact assessment keeps the CRM from accumulating one-off fields that nobody maintains.
How an account moves between teams once handoffs are defined
- Marketing
Generates and nurtures leads, then passes qualified ones with their history.
- Sales development
Accepts or rejects leads with a reason and converts accepted ones to opportunities.
- Account executive
Runs the opportunity through stages with recorded exit criteria.
- Customer success
Receives deal context before kickoff and owns adoption and renewal.
- Finance
Books revenue from contract terms and reconciles bookings with the CRM.
- Marketing to Sales developmentQualified lead + context
- Sales development to MarketingReject with reason
- Sales development to Account executiveOpportunity + notes
- Account executive to FinanceClosed-won + terms
- Account executive to Customer successHandover brief
- Customer success to Account executiveExpansion signal
- Customer success to MarketingAdvocate or churn reason
Choosing the system of record for each revenue object
A common pattern, not a rule. The point is that each object has exactly one owner system and everything else reads from it.
| Object | Usual system of record | Systems that read it | Watch out for |
|---|---|---|---|
| Account and contact | CRM | Marketing automation, support desk, billing | Marketing tools creating contacts the CRM never dedupes |
| Lead and campaign membership | Marketing automation, synced to CRM | CRM, attribution reporting | Two-way sync loops that overwrite sales edits |
| Opportunity and forecast | CRM | Finance planning, data warehouse | Spreadsheet forecasts that drift from the CRM |
| Subscription, invoice and payment | Billing system | CRM, customer success tooling, warehouse | Contract values typed into the CRM by hand |
| Product usage | Product analytics or warehouse | Customer success tooling, CRM summary fields | Raw event streams pushed into the CRM |
| Support case | Support desk | CRM, customer success tooling | Case history invisible to account owners |
Revenue metrics to define before building any dashboard
- Bookings
- The contract value signed in a period, with a written rule for multi-year deals, discounts and cancellations.
- Net revenue retention
- Recurring revenue from an existing customer cohort at the end of a period, including expansion and net of contraction and churn, divided by that cohort's recurring revenue at the start.
- Gross revenue retention
- The same ratio excluding expansion, so it shows only what was kept.
- Pipeline coverage
- Open qualified pipeline due to close in a period divided by the target for that period.
- Sales velocity
- Qualified opportunities multiplied by win rate and average deal size, divided by average cycle length.
- Stage conversion rate
- The share of records entering a stage that reach the next one within a defined window.
Where AI fits once the foundations exist
Enrichment, routing and qualification are the usual first AI uses, and each depends on the steps above. A routing model is only as good as the ownership rules it applies, and a qualification agent is only useful if sales has agreed what qualified means. The AI lead qualification use case shows that pattern for inbound demand.
Further along, models can flag forecast risk from activity patterns, summarize calls into CRM fields and suggest next steps. On a Salesforce estate, Agentforce is one way to run such agents inside the CRM. Whatever the platform, give every automated action an owner and log what it changed, so the hygiene rules from step six still apply.
Questions and answers
Who should revenue operations report to?
Ideally to the executive accountable for all revenue, such as a chief revenue officer or chief operating officer, so it can arbitrate between marketing, sales and customer success without belonging to any one of them. Reporting into sales alone tends to tilt priorities toward pipeline over retention. Some companies place it under finance, which helps forecast discipline but can slow tool changes.
When is a company big enough to need revenue operations?
Size matters less than the number of handoffs. Once marketing, sales and customer success are separate teams with their own tools, someone has to own definitions, data and process between them, even part-time. Small companies often start with one person holding the CRM and reporting, and grow the function as segments, products or regions multiply.
How is revenue operations different from sales operations?
Sales operations traditionally supports the sales team: territories, quotas, CRM administration and forecasting. Revenue operations takes the same disciplines across the whole customer lifecycle, covering marketing operations, sales operations and customer success operations under shared definitions and one data model. The practical difference is who arbitrates when two teams count the same thing differently.
Should we buy new tools when setting up revenue operations?
Usually not at the start. Definitions, data models and handoff rules can be put in place in the CRM and marketing automation you already run, and doing so shows which gaps are real. Buy tools afterwards for specific needs, such as forecasting, enrichment or conversation analysis, once you know which system of record each one will read from and write to.