CRM Data Quality: How to Audit and Fix Bad Data in 2026
By Kushal Magar · November 4, 2026 · 13 min read
Key Takeaway
CRM data quality comes down to five measurable dimensions: completeness, accuracy, freshness, uniqueness, and consistency. Audit them on 8 to 12 critical fields, weight them into one 0 to 100 score, and fix the lowest dimension first. Email data alone loses about a quarter of its validity in a year, so a one-time cleanup does not hold. Validate at entry, re-audit quarterly, and re-enrich active records on a schedule.
CRM data quality is how complete, accurate, fresh, unique, and consistent your CRM records are. It decides whether routing works, whether emails land, and whether the forecast means anything.
Most teams know their data is messy but cannot say how messy. This guide fixes that with a concrete audit: five dimensions, a formula for each, one weighted score, and the exact places to pull the numbers in HubSpot and Salesforce.
It is written for RevOps, sales ops, and founders who own the CRM. You will leave with a sample audit, a validation checklist, and a fix order that stops the same problems from coming back.
TL;DR
- Five dimensions: completeness, accuracy/validity, freshness, uniqueness, and consistency. Each has a simple percentage formula.
- Audit critical fields only: 8 to 12 fields that drive routing, segmentation, and outreach. Not all 400 properties.
- One score: weight the five dimensions into a 0 to 100 score. 90+ is healthy, under 75 needs fixing before your next campaign.
- Decay is the real enemy: ZeroBounce measured at least 23% email list decay over 2025. A one-time cleanup starts rotting the day it ends.
- Fix order: dedupe, normalize, verify, enrich, then schedule re-enrichment. Doing it out of order wastes credits and creates new duplicates.
What Is CRM Data Quality?
CRM data quality measures whether the records in your CRM are fit for the jobs your team uses them for. A record is high quality when a rep can route, segment, and contact from it without re-checking anything.
"Quality" is too vague to manage. Break it into five dimensions you can count:
| Dimension | The question it answers | Formula | Suggested target |
|---|---|---|---|
| Completeness | Is the field filled in? | Records with field populated ÷ total records | 95%+ on critical fields |
| Accuracy / validity | Is the value true and usable? | Sampled records that check out ÷ records sampled (emails: valid ÷ verified) | 90%+ |
| Freshness | Was it confirmed recently? | Records verified or enriched in last 90 days ÷ active records | 80%+ of active records |
| Uniqueness | Does each person or company exist once? | 1 − (duplicate records ÷ total records) | 98%+ |
| Consistency | Is it in the standard format? | Values matching picklist or format rule ÷ populated values | 95%+ |
Targets are SyncGTM's working rules of thumb, not an industry standard. Adjust them to your motion.
Accuracy and validity are not the same thing. A validly formatted email can still belong to someone who left the company, so accuracy needs a check against a trusted source.
Freshness is the dimension most audits skip. A record can be complete, valid, and unique on the day you import it and still be wrong six months later.
Looking for the broader maintenance program rather than the audit? Our CRM data hygiene guide covers deduplication, standardization, decay, and governance as ongoing pillars.
What Bad CRM Data Costs: The Decay Math
Bad CRM data costs revenue, rep time, and trust in every report built on top of it. The surveys below are self-reported, so treat them as direction, not precision.
- Email decay: ZeroBounce's Email List Decay Report for 2026 says at least 23% of an email list decayed within one year, based on addresses processed from January to December 2025. Only 62% of the emails submitted to it in 2025 were valid.
- Accuracy gaps: In Validity's State of CRM Data Management in 2024 survey of 600+ CRM admins, 24% said less than half of their data is accurate and complete.
- Revenue impact: In the same 2024 survey, 31% said poor-quality data costs them at least 20% of annual revenue.
- AI readiness: Validity's 2026 report announcement(500 marketers, released August 2026) says just 21% rate their CRM data as "very well prepared" to support AI, and 62% report losing revenue directly to poor CRM data quality.
Decay math for a 12,000-contact CRM
At a 23% annual decay rate, roughly 2,760 email addresses go bad in a year. That is about 230 a month.
If you audit once a year, the CRM is wrong on hundreds of records for most of the year. That is why the fix section below ends with scheduled re-enrichment, not a bigger annual cleanup.
Emails are only the visible layer. Job titles change at promotions and job moves, and firmographics drift as companies hire, shrink, and swap tools, which is why technographic data goes stale too.
How to Audit CRM Data Quality in 6 Steps
A CRM data quality audit scores a short list of critical fields on five dimensions, then ranks fixes by business impact. Plan on a day for the first run and an hour a month after that.
Step 1: Pick your critical fields
Audit the 8 to 12 fields that drive routing, segmentation, scoring, and outreach. Everything else can wait.
- Contacts: email, first and last name, job title, seniority, phone, LinkedIn URL, country.
- Companies: company domain, industry, employee count, country, lifecycle or account owner.
- System fields:lead source, and a "last enriched date" property if you do not already track one.
Scope the audit to active records, such as contacts touched in the last 12 months or in open pipeline. Scoring dead records drags the numbers down without changing any decision.
Step 2: Measure each dimension
- Completeness: count records where each critical field is known. Report it per field, because a 90% average can hide a 40% fill rate on job title.
- Accuracy: pull a random sample of 100 records and check each against LinkedIn or the company website. For emails, run the populated set through verification and count the valid share.
- Freshness: count active records with a last enriched or verified date inside 90 days. No date means not fresh.
- Uniqueness: count duplicate records, matching contacts on email and companies on domain. Our data deduplication guide covers the matching rules.
- Consistency: count values that match your picklists and formats, such as country codes, phone format, and industry labels.
Step 3: Roll it into one score
Leadership will not read a 40-cell table, so roll it into one number. Here is the weighting we use:
Completeness and accuracy get the most weight because they decide whether a record can be used at all. Shift weight to freshness if you run high-volume outbound, where stale emails hurt deliverability fastest.
| Band | Score | What to do |
|---|---|---|
| Healthy | 90–100 | Keep monthly monitoring and scheduled re-enrichment |
| Needs work | 75–89 | Fix the lowest dimension this quarter; block new bad data at entry |
| Problem | Below 75 | Clean before the next campaign, territory plan, or forecast |
Sample CRM data quality audit
Here is an illustrative audit of a 12,000-contact CRM. The numbers are an example, not benchmark data.
| Field | Dimension | Result | Status |
|---|---|---|---|
| Completeness | 11,280 / 12,000 = 94% | Amber | |
| Validity (verified) | 9,700 valid / 11,280 = 86% | Red | |
| Job title | Accuracy (100-record sample) | 78 / 100 = 78% | Red |
| Company domain | Completeness | 11,640 / 12,000 = 97% | Green |
| Industry | Consistency | 8,900 / 10,100 = 88% | Amber |
| Country | Consistency | 10,900 / 11,500 = 95% | Green |
| Last enriched date | Freshness (90 days) | 4,200 / 12,000 = 35% | Red |
| Contact record | Uniqueness | 1 − (540 / 12,000) = 95.5% | Amber |
Rolled up: completeness about 95, accuracy about 82 (average of email validity and title accuracy), freshness 35, uniqueness about 96, consistency about 91. The weighted score lands near 79, in the "needs work" band.
The diagnosis is clear: this CRM is full but stale. More manual cleanup will not fix it, because freshness is the problem.
Step 4: Pull the numbers in HubSpot
HubSpot's data quality tools sit under Data Management > Data Quality, a page HubSpot also calls the data quality command center. HubSpot marks the article as beta, so menu names may shift.
- Property insights: flags properties with no data, duplicate, or unused issues. Use it to spot empty critical fields fast.
- Formatting issues tab: surfaces formatting problems in property values. That feeds your consistency score.
- Manage duplicates tab: lists suspected contact and company duplicates for your uniqueness score.
- Data enrichment tab: shows enrichment coverage and gaps.
Per HubSpot, the tools are available on Starter, Professional, and Enterprise tiers of Marketing, Sales, Service, Data, and Content Hub. Some features, such as duplicate alerts and property anomalies, need Data Hub Professional or Enterprise.
For exact completeness counts, filter a contact view by "[property] is unknown" and divide by the total. For freshness, filter on your last enriched date property.
Step 5: Pull the numbers in Salesforce
Salesforce has no single quality dashboard, so build the audit from reports. One report per critical field is enough.
- Completeness:a Contacts or Leads report filtered on "[field] equals" blank, compared with the unfiltered row count.
- Consistency:group a report by Country, Industry, or Lead Source and look for variants such as "US", "USA", and "United States" sitting side by side.
- Freshness: filter on a custom last enriched date field, or on Last Modified Date as a weaker proxy.
- Uniqueness: duplicate jobs scan accounts, contacts, or leads across the whole org using standard or custom matching rules. They are available in Performance and Unlimited editions.
On other editions, duplicate rules still warn or block at entry. Pair them with a report grouped by email or website to find existing duplicates.
Step 6: Rank the fixes
Fix the dimension with the lowest score on the fields with the highest business impact. A 70% fill rate on industry matters less than 86% email validity if your team runs outbound.
Write each fix as an owner, a date, and a target score for the next audit. Without that, audits turn into slide decks nobody acts on.
CRM Data Validation Checklist
Validation stops bad data at the door, which is cheaper than finding it in the next audit. Use this checklist for forms, imports, integrations, and enrichment writes.
At entry
- Required fields enforced on forms and imports: email, company domain, country.
- Email checked for syntax, disposable domains, and role addresses such as info@.
- Company domain stored bare (acme.com), not as a full URL or with www.
- Picklists instead of free text for country, industry, seniority, and lead source.
- Phone numbers stored in one format, ideally E.164.
- Duplicate check on email and domain before a new record is created.
- Every import tagged with a source and date, so bad batches can be traced and rolled back.
On a schedule
- Emails re-verified before every major send and at least quarterly on active records.
- Job titles and companies re-checked for job changes on open pipeline and key accounts.
- Firmographics refreshed on target accounts: headcount, industry, tech stack.
- Duplicate scan run monthly and merged under written survivorship rules.
- Data quality score logged monthly with the owner of each red dimension.
For the email-specific side of this list, see our email hygiene guide.
How to Fix Bad CRM Data, in Order
Fix CRM data in this order: dedupe, normalize, verify, enrich, then schedule re-enrichment. Each step makes the next one cheaper.
- Dedupe first. Merging after enrichment means you paid twice for the same person. Match on normalized email and domain, then review fuzzy matches; the deduplication guide walks through merge rules for HubSpot and Salesforce.
- Normalize formats. Standardize countries, phone formats, job title seniority, and domains so filters and matching work. Our data normalization guide has the rules and mappings.
- Verify emails. Remove or flag invalid and risky addresses before they hit a sequence and damage sender reputation.
- Enrich the gaps. Backfill missing titles, phones, and firmographics. A single provider rarely covers everything, which is why waterfall enrichment tries several in sequence until a field is found.
- Schedule re-enrichment. Put active records on a refresh cadence so freshness stays up without a project. This is the step that changes the next audit.
Comparing vendors for steps 3 and 4? See our roundup of the best CRM data enrichment tools and the playbook on enriching CRM data for B2B sales teams.
How Often to Re-Audit
Run a full CRM data quality audit quarterly and a light scorecard monthly. Validation at entry runs every day.
- Daily: entry validation, duplicate blocking, email verification on new records.
- Monthly: recompute the five dimension scores on critical fields and log the overall score.
- Quarterly: full audit with a fresh 100-record accuracy sample, plus a review of which sources create the most bad records.
- Before big moves: audit before a migration, a territory redesign, a new scoring model, or an AI rollout.
Clean data pays off downstream too. Nurture programs in lead nurturing tools only personalize well when titles, industries, and emails are right.
How SyncGTM Keeps HubSpot and Salesforce Data Fresh
SyncGTM fixes the dimension most audits flag and few teams solve: freshness. It connects to HubSpot and Salesforce with 2-way sync, so enriched values write back to the records your team already uses.
- Backfill, dedupe, re-enrich. The GTM automation platform is built to "backfill missing fields on every existing record, dedupe on entry, and re-enrich on a schedule so contact and account data in HubSpot or Salesforce stays current without manual scrubbing."
- Pull records in for checks. The HubSpot and Salesforce lead sources bring CRM records into SyncGTM for enrichment and quality checks.
- Completeness and accuracy. Waterfall enrichment across 50+ providers fills missing emails, phones, and firmographics, and Verify Email confirms addresses before they reach a sequence.
- Uniqueness and consistency. Company Domain resolves messy company names to a domain, the key that dedupe and account matching depend on.
- Freshness on people. Job change signals flag when a contact moves, so the stale title and email get replaced instead of bouncing.
You can start free and see per-credit costs on the pricing page.
Final Thoughts
CRM data quality is measurable, and what gets measured gets fixed. Score five dimensions on your critical fields, roll them into one number, and fix the lowest one first.
Then make the fix permanent. Validate at entry, re-audit quarterly, and put active records on scheduled re-enrichment so your next audit shows a better number instead of the same problems.