The RevOps Audit Checklist: 40 Questions That Find Revenue Leaks

Most teams don’t run one until something forces the issue — a board question the dashboards can’t answer, a migration, a new revenue leader inheriting a stack nobody documented. That’s later than ideal but better than never. The checklist below is the one we work through with B2B SaaS companies; it’s designed so you can run it yourself in a focused week. One framing note before the questions: an audit is not a tool inventory. Listing your 23 martech subscriptions tells you what you pay for, not where revenue leaks. Every question below is about flow — can a record move from anonymous visitor to closed-won without being lost, duplicated, mis-routed, or mis-counted?

How to run this

• Score each question 0, 1 or 2. 0 = no / don’t know. 1 = partially / undocumented. 2 = yes, documented, and someone owns it. “Don’t know” scores 0 deliberately — in operations, undocumented and nonexistent are functionally the same.
• Pull real records as evidence. For at least ten questions, don’t answer from memory; open the CRM and trace ten actual leads from creation to current state. The gap between how the funnel is described and how records actually behave is the audit finding.
• Involve one person from marketing, sales and CS each. Half the value is discovering that three teams give three different answers to the same question.

Pillar 1 — CRM & data foundation (10 questions)

1. Is there a written data dictionary — fields, definitions, allowed values, and which system writes each?
2. Do you know your duplicate rate? (Not “do you have duplicates” — everyone does. Do you measure them?)
3. Is there dedupe logic at every intake point: forms, imports, integrations?
4. Are the fields used for routing and segmentation (country, employee band, industry) governed picklists rather than free text?
5. Do imports go through a gated template with a mandatory source tag?
6. For every integration writing to the CRM, is field precedence defined — who wins on conflict?
7. Is original source captured automatically and never overwritten?
8. Do you have a suppression policy (competitors, bounces, opt-outs) that actually excludes those records from scoring, sync and sends?
9. Are you tracking database size against platform record tiers, with an archive policy?
10. Could you restore from a backup/export if a mass update went wrong tomorrow? Why it’s first: every downstream pillar inherits this one. Routing distributes duplicates; reporting aggregates free text; automation acts on stale fields. (Our full maintenance system for this pillar: the CRM data hygiene guide.)

Pillar 2 — Lifecycle, routing & handoffs (10 questions)

1. Does every lifecycle stage have written entry criteria, an owner and an SLA — or just a name?
2. Does a Sales Accepted stage exist, so marketing’s MQL claim gets an explicit yes/no from sales?
3. Do backward paths (Recycled, Disqualified) exist with mandatory reason codes?
4. Is one system the source of truth for lifecycle stage, with one-directional sync?
5. What is your measured speed-to-lead for a hand-raise — measured from form submit, not from assignment?
6. Do hand-raises bypass scoring and nurture and take the fastest path to a rep?
7. Do routing rules check availability (working hours, PTO, capacity), not just names?
8. When the first-touch SLA is breached, does anything happen automatically?
9. Are departed reps fully removed from routing rules and record ownership?
10. Query it now: how many open leads have no owner, or a queue owner, older than 24 hours? Evidence exercise: take last month’s ten most recent demo requests and reconstruct the timeline from submit to first logged touch. This single exercise finds more revenue than any dashboard review.

Pillar 3 — Automation & integrations (10 questions)

1. Is there an inventory of every workflow/automation that writes to lifecycle, owner or scoring fields — and could you name the owner of each?
2. When did someone last retire an automation? (Stacks where nothing is ever deprecated are stacks nobody fully understands.)
3. For your core sync (e.g., HubSpot–Salesforce, Marketo–Salesforce), is there a field-mapping document a new hire could read?
4. Is the sync filtered by an inclusion rule, or does everything cross by default?
5. Who reviews sync errors, on what cadence, and what’s the current backlog count?
6. Are enrichment and dedupe sequenced before routing in the automation chain?
7. Do any two automations write the same field with different logic? (The classic silent war.)
8. Are API limits, batch windows and sync latency known numbers or mysteries?
9. Is there a sandbox/test path for automation changes, or does everything ship to production?
10. If your integration vendor or iPaaS went down for a day, would you know within an hour, and what breaks?

Pillar 4 — Reporting & attribution (10 questions)

1. Can marketing and sales produce the same funnel numbers from the same definitions — or does every QBR start with reconciliation?
2. Are stage conversion rates and stage velocity trended monthly, from CRM data rather than spreadsheets?
3. Do you know your SAL acceptance rate (the single best alignment metric)?
4. Is UTM capture enforced and standardized — governed values, not ad-hoc tagging?
5. Does source attribution survive the lead-to-opportunity conversion, or does it die at the handoff?
6. Can you connect closed-won revenue back to originating campaigns without manual stitching?
7. Is pipeline coverage reported against a documented definition of “pipeline”?
8. Do executives use the CRM dashboards, or do they ask for spreadsheets? (Behavior is the real trust metric.)
9. Are recycle/closed-lost reason codes reported and reviewed — is the funnel teaching you anything?
10. If the CEO asked “what did we get for last quarter’s marketing spend,” is the answer a report or a project?

Scoring and what to do with it

Total the pillars separately — the profile matters more than the grand total:

Pillar score	Reading
16–20	Healthy; maintain cadence
10–15	Functioning but fragile; prioritize the 0-scores
< 10	This pillar is actively leaking; fix before investing downstream

Three patterns we see constantly:

• Strong Pillar 3, weak Pillar 1: heavily automated mess — sophisticated workflows acting on data nobody trusts. Counterintuitively, pause automation expansion and fix the foundation; automation multiplies whatever quality it’s fed.
• Strong Pillars 1–3, weak Pillar 4: the machine works but can’t prove it. Usually a UTM governance and conversion-mapping problem, not a dashboard problem — fixable in weeks.
• Uniformly ~50%: nothing is on fire, everything is friction. This is the profile that most benefits from sequencing help, because every fix depends on two others. Sequence rule of thumb: data foundation → lifecycle/routing → integrations → reporting. Reporting is last not because it matters least but because it’s downstream of everything — building dashboards on unaudited pipes just publishes the leaks.

Self-audit vs. bringing someone in

This checklist is genuinely runnable internally, and for a first pass you should. Where outside help earns its fee is (a) the evidence work — tracing records, querying the anomalies, quantifying what each leak costs — and (b) political neutrality: “the audit found it” lands differently than “marketing says sales ignores leads.” If you’d rather have a second set of eyes do the tracing, that’s exactly what our free RevOps audit is — we run a version of this review against your actual stack and hand you the findings, whether or not you work with us afterwards.