RevOps Automation Audit Checklist

An automation audit turns “things feel slow” into a specific, ranked list of what to fix. Done manually, an audit relies on spreadsheets and inconsistent CRM exports, and whoever runs it usually ends up with a gut feeling rather than a decision anyone can act on. Done properly, it is a repeatable system that surfaces revenue leakage, duplicate workflows and integrations nobody remembers building, the same way a financial audit surfaces unreconciled accounts rather than a vague sense that the business “feels” profitable. This checklist covers what to check, the failure modes that show up most often, and how to turn the findings into a roadmap your team will actually work through.

How to Prepare Before You Start Auditing

Start with people and process, not tools. If you begin by pulling reports out of the CRM, you are already biased toward whatever that system happens to log well, and you will miss every manual workaround a rep built because the automation didn’t do what they needed. Talk to the people doing the work first: sales reps, SDRs, customer success managers, and whoever owns marketing operations. Ask them where they still export a spreadsheet, copy a number by hand, or keep a personal tracker outside the CRM. Those answers tell you more about where automation has actually failed than any dashboard will.

Map Every System That Touches Revenue Data

Build a simple inventory before you judge anything: CRM, marketing automation platform, any enrichment or intent-data vendor, billing or finance system, customer success platform, and the automation or integration layer connecting them (n8n, Zapier, Make, or custom scripts). For each system, note two things: which direction data flows, and whether that flow is a native integration, a middleware workflow, or a person manually exporting and importing a file. A meaningful share of the automation you find during this stage will turn out to be redundant, doing a job something else already does, often because two different teams built overlapping fixes for the same problem without knowing the other existed.

Assign an Owner Before You Find Problems

Audits stall when nobody has both the authority and the access to act on findings. Appoint a single named owner before you start, and give that person read access across every system on the map in advance, not partway through the audit once trust has already broken down between marketing ops and sales ops over whose data is “correct”. This matters because the two functions usually disagree about which system is the source of truth for a given field, and that disagreement is exactly the kind of thing an audit needs to resolve rather than sidestep.

Five Pillars to Check, and What Good Looks Like in Each

Once the map and the owner are in place, work through five areas in order. Each one has a distinct failure signature, and treating them as one undifferentiated “automation problem” is how audits end up producing a list of symptoms instead of causes.

Data Infrastructure: Is Data Moving Without Human Hands

Check whether every tool in your stack, CRM, marketing platform, customer success software, pushes and pulls data automatically, or whether someone still exports a CSV to close the gap. Where a native integration exists, check what happens when a field gets renamed or a custom property is added on one side. A renamed field in HubSpot, for instance, will silently stop mapping into the corresponding Salesforce field rather than throwing an error, which means the sync keeps “working” while quietly dropping data. Also check for last-write-wins conflicts: if two systems can both write to the same field, whichever syncs last overwrites the other, and nobody notices until a rep acts on stale information. The fix is either to designate one system as the single writer for that field, or to build explicit conflict-resolution logic into the middleware layer, which is exactly the kind of node-based logic n8n’s workflow documentation covers for merging and deduplicating records from multiple sources.

Workflow Automation: Finding Redundant and Conflicting Triggers

Export the full list of active workflows from each platform, not just the ones people remember, and check for overlapping enrollment triggers. It’s common to find two workflows in the same tool both firing off the same trigger event, built months apart by different people solving what they thought was a new problem. Also check re-enrollment settings specifically: a workflow that re-enrolls a contact into itself under the wrong condition can loop a lead through the same email sequence repeatedly, which looks like a bug but is usually a misconfigured re-enrollment trigger. HubSpot’s own guidance on building and configuring workflows is worth reviewing against your live workflows, not just your original build documentation, because settings drift over time as different admins make changes. Finally, check the lead-qualification handoff itself: if marketing’s scoring criteria for a marketing-qualified lead don’t match what sales actually treats as qualified, routing automation will hand over leads that sales ignores, and both teams will blame the automation rather than the mismatched definition underneath it.

Reporting and Forecasting: Testing Whether Numbers Are Trustworthy

Check whether forecasts refresh dynamically against live pipeline data or whether someone is still exporting a snapshot into a separate spreadsheet each week. Persistent forecast variance usually signals a model problem, not a market problem: if the gap between forecast and actual is consistent in direction, the stage-exit criteria in the CRM probably don’t match what reps actually do before moving a deal forward. Test this directly rather than guessing: pick ten recently closed-won deals, trace their stage history, and check whether each stage transition matches what genuinely happened with the buyer at that point. If deals are routinely sitting in “proposal sent” for weeks after the proposal was actually sent and negotiated, your stage definitions are decorative rather than functional, and no amount of dashboard polish will fix a forecast built on them.

Technology Stack: Spotting Duplicate and Underused Tools

Pull login and usage activity for every tool in the stack, not just licence counts, and look for tools nobody has opened in the past month. Then look specifically for two systems doing the same job in parallel, most often a legacy CRM still running alongside a newer one after a migration that was never fully completed. This happens more often than teams expect: a migration gets 80 percent finished, the old system still holds a handful of active records or a team that never moved over, and both platforms end up receiving live data indefinitely because switching off the old one feels riskier than leaving it running. An audit should name this explicitly rather than letting it persist as background cost.

Governance and Compliance: Checking Permissions and Retention

Check who currently holds admin-level permissions across the CRM and connected tools, and whether any of those accounts belong to contractors or former employees whose access was never revoked. Check whether you can actually trace who changed a deal stage, a contact’s data, or an automation’s configuration, and when, since an audit trail that only exists in theory is not an audit trail. Finally, check data retention: is there a defined policy for how long lead and prospect data is kept, and does it hold up against UK data protection obligations? The ICO’s UK GDPR guidance and resources sets out what regulators expect an organisation to be able to demonstrate, and the gov.uk guidance on data protection is a useful baseline to check your retention settings against, particularly if enrichment data or old marketing lists are still sitting in the CRM with no defined expiry.

Common Failure Modes an Audit Should Catch

Beyond the five pillars, certain failure patterns recur often enough that it’s worth checking for them by name rather than waiting to stumble across them.

• Silent webhook failures. A webhook can return a success status while delivering a malformed or incomplete payload, and without logging or alerting on the receiving end, the failure never surfaces until someone notices missing data weeks later. Fix this by adding explicit error handling and a dead-letter path for failed payloads, rather than assuming a 200 response means the data actually arrived intact.
• Orphaned automations after staff leave. Workflows and integrations are often built under a specific person’s login or personal API key. When that person leaves and their account is deactivated, the automation can fail without warning, sometimes weeks later once a token expires. Fix this by auditing which automations run under personal credentials versus a shared service account, and moving anything business-critical onto the latter.
• Duplicate contact creation. Multiple lead capture points, a web form, a chat widget, an event list import, often create separate records for the same person because matching rules only check one field, usually email, and miss variations like a typo or a different email address entered at a trade show. Fix this with matching rules that check multiple fields and a defined merge process, not just a dedupe tool running in the background.
• Race conditions in lead routing. If two systems can both claim ownership of a new lead, for example a routing workflow in the CRM and a separate one in the marketing platform, both can assign the lead within seconds of each other, and whichever writes last wins, sometimes to the wrong rep. Fix this by making one system the single routing engine and having every other tool defer to it.
• Field mapping drift. Platform updates and admin changes shift field mappings gradually rather than all at once, so a sync that worked correctly for a year can degrade one field at a time without a single obvious breaking change. Fix this by keeping workflow and mapping configurations in version control or a change log, so drift is visible rather than discovered by accident.

Scoring Findings and Building a Fix Roadmap

Score each of the five pillars, then rank the individual fixes you’ve found by impact and effort rather than tackling them in the order you happened to discover them. A fix that synchronises lead-scoring data between marketing and sales usually beats a dashboard redesign nobody was actually confused by, because the first changes what reps do every day and the second changes what they look at occasionally. Plotting fixes on an impact-versus-effort grid makes this ranking visible to the whole team rather than living in one person’s head: high-impact, low-effort fixes like syncing lead-scoring data become the obvious starting point, high-impact, high-effort work like consolidating two CRMs running in parallel gets scheduled as a proper project rather than squeezed in around other work, low-impact, low-effort items like a dashboard redesign get parked as fill-in work, and low-impact, high-effort ideas like rebuilding a forecasting model from scratch get reconsidered rather than started by default.

When to Re-run the Audit After the First Fix

An audit isn’t a once-a-year ritual you schedule alongside the annual budget review. Re-run at least the relevant pillar whenever something changes the underlying system map: a headcount change in sales or marketing ops, a new tool procured outside the usual process, a merger or acquisition that brings a second CRM into the business, or a major platform update from a vendor that changes default workflow behaviour. Between those triggers, revisiting the checklist roughly twice a year is a reasonable baseline for most teams, since drift in field mappings and permissions accumulates gradually rather than all at once, and catching it early is considerably cheaper than untangling a year’s worth of quiet workarounds.

Related Reading

This checklist feeds into a broader system: see RevOps Automation: The Complete Guide for how these individual fixes sequence into a full roadmap.

For more on this, see our automation and n8n coverage, including Data Governance Automation for RevOps Success in the UK Using N8N, Boost RevOps Efficiency with n8n Workflow Automation Strategies, and Advanced n8n Webhook Listeners for Real-time SaaS and RevOps Automation.

Book your free AI audit