Audit Your CRM Data Quality in One Afternoon

One afternoon is enough. Not to fix every broken record in your CRM — that's a multi-quarter project — but to find out exactly how broken it is, where the damage is concentrated, and what's costing your team the most pipeline. The audit below is designed to fit between lunch and end-of-day, and it produces a one-page memo your RevOps lead or VP can act on Monday morning.

Set the scope before you open a single report

The mistake most teams make is treating "data quality" as one thing. It isn't. A CRM has at least four distinct quality problems, and each one has different symptoms, different owners, and different fixes:

• Completeness — fields that should be filled but aren't (industry, employee count, lead source)
• Accuracy — fields that are filled but wrong (a "CTO" who left two years ago, a phone number that rings a deli)
• Consistency — the same value entered five different ways ("Series B", "series-b", "Round B", "B")
• Currency — records that were accurate once but have decayed (job changes, company moves, status drift)

Pick the two that matter most for the next quarter's plan. If you're about to launch an ABM motion, completeness and currency of account fields matter most. If forecasting accuracy is the boardroom issue, focus on opportunity-stage consistency and pipeline hygiene. Trying to audit all four in an afternoon produces a mess. Two is enough to be useful.

Write down your scope in one sentence before doing anything else. Example: "Audit completeness and currency of contact records on all open opportunities in the SMB segment." That sentence is your guardrail when you get tempted to chase a tangent at 3 p.m.

Build the sample, not the spreadsheet

You do not need to look at every record. You need a defensible sample.

Pull three slices and randomise within each:

1. Open opportunities forecast to close this quarter. These records have the highest cost-of-being-wrong. A bad contact here means a deal slips.
2. Closed-won deals from the last two quarters. These should be your cleanest data — they were touched repeatedly by AEs, CS, and finance. If they're messy, your whole instrumentation layer is suspect.
3. Leads created in the last 90 days that never converted to an opportunity. This is where rot accumulates fastest, and where SDR routing rules quietly break.

Take roughly thirty records from each slice. Ninety records is enough to see the patterns; more than that and you'll run out of afternoon. Drop them into a single sheet with columns for the fields in your scope, plus three audit columns: Status (clean / incomplete / wrong / stale), Evidence, and Likely cause.

The Evidence column is the discipline that makes this exercise worth anything. If you mark a contact "wrong," you write "LinkedIn shows she moved to Stripe in March." If you mark an opportunity "stale," you write "last activity 47 days ago, stage = Negotiation." Without evidence, you're just venting.

Run the checks in passes, not record-by-record

Going record-by-record across all fields is the slow way. Run the audit in passes — one field, ninety records, then the next field. Pattern recognition kicks in faster, and you'll spot systemic issues (everyone on the West team leaves "Next Step" blank, every record from the trade-show import has the same generic title) that would be invisible if you were jumping between fields.

Suggested passes for a contact-and-opportunity audit:

• Pass 1: Title and role. Cross-check against LinkedIn. Mark stale anything older than nine months without verification.
• Pass 2: Email validity. Run the addresses through whatever verification tool you have. Hard bounces and catch-alls both count as quality failures, for different reasons.
• Pass 3: Account firmographics. Industry, employee band, headquarters. These drive segmentation and routing; errors here misroute leads for months.
• Pass 4: Opportunity hygiene. Stage, close date, next step, amount. Flag any open opp where close date has already passed or next step is empty.

A worked example to ground the sampling math: say your team has 600 open opportunities and you sample 30. If 9 of them have a passed close date and no updated next step, that's a 30% defect rate in your sample. You don't have to claim that's the exact rate in the full population — you have to claim it's bad enough that forecast calls are running on fiction. That's enough to act on.

Translate findings into a one-page memo

The output of the afternoon is not a cleaned CRM. It's a memo with four sections:

What we audited. Scope sentence, sample size, slices.

What we found. Defect rates by field, by team, by record source. Lead with the worst three. Resist the urge to list everything.

What it's costing. Tie each major defect to a downstream consequence. Stale titles → SDRs opening with the wrong name and getting blocked at gatekeepers. Empty next-step fields → forecast roll-ups based on close-date guesses. Inconsistent stage definitions → velocity metrics that no one trusts.

What to fix first. Two categories: things a single admin can fix this week (required-field rules, picklist consolidation, a bulk re-verification job), and things that require a process change (SDR routing logic, manager inspection cadence, enrichment vendor swap). Be specific about the owner.

The memo is short on purpose. A leader who gets a 14-page audit will skim it. A leader who gets one page with three defect rates and a fix-it list will forward it to the person who can act.

The genuinely useful insight

Most CRM data problems are not data problems. They're incentive problems. Reps don't fill in "Next Step" because no one ever lost a deal review for leaving it blank. SDRs don't correct a stale title because the routing rule already gave them credit for the dispo. The audit's most valuable output is usually not the list of bad records — it's the map of fields that are required by the data model but not enforced by the management cadence. Fix the cadence and the data follows. Clean the data without fixing the cadence and you'll be running the same audit next quarter.

The takeaway

• Scope to two quality dimensions and one segment before you open the CRM. Write the scope sentence down and stick to it.
• Sample ninety records across three slices (open opps, recent closed-won, unconverted leads) and audit in field-by-field passes, not record-by-record.
• Ship a one-page memo that ties each top defect to a downstream cost and names an owner for the fix — and identify which management cadence needs to change, not just which records need scrubbing.