Data Quality · January 2026

Data Quality Metrics That Matter

You could track a hundred data quality metrics. Field completion percentages, format compliance scores, freshness indicators, consistency ratios. Most of them won't tell you anything useful.

The metrics that matter are the ones that predict operational problems. The ones that, when they go bad, mean something breaks. Here are those metrics, how to calculate them, and what targets to aim for.

The Problem with Most Data Quality Metrics

Traditional data quality frameworks give you metrics like "correctness," "completeness," "consistency," and "timeliness" (the standard dimensions defined by DAMA International's DMBOK). These sound reasonable but often miss the point.

A single completeness percentage does not tell you what to fix. Report completeness by field and workflow, then identify which missing values affect routing, reporting, or outreach.

Good metrics connect directly to operational outcomes. When this metric gets worse, something specific breaks. When it improves, something specific works better. That's the test.

The Core Metrics

How to measure: Run your contact emails through a verification service. Count valid/total. Segment by marketing status since active contacts matter more than dormant ones.

When to check: Monthly for active marketing contacts, quarterly for full database.

How to measure: Run duplicate detection using email exact match, name + company fuzzy match, and phone number match. Count unique records flagged as having duplicates. Divide by total records.

When to check: Monthly. Track trend over time to see if duplicates are being created faster than they're being merged.

How to measure: Identify the specific fields your routing logic uses (typically: company size, industry, country, maybe title or revenue). Calculate completion rate for each on your active lead population.

When to check: Weekly if you have high lead volume. At minimum, before any routing rule changes.

How to measure: Count contacts with no account/company association. Divide by total contacts. For HubSpot, this is contacts without company associations. For Salesforce, contacts without an account relationship.

When to check: Monthly. Spikes often indicate integration issues or import problems.

Why These Metrics Drift: Data Decay

A clean database changes over time as people move, companies are acquired, domains lapse, and phone numbers are reassigned. Measure field-level change between dated samples, then set the review cadence from changes observed in your own system.

That's why each metric above comes with a cadence. A single audit gives you a snapshot. The trend is what tells you whether your processes are keeping up with decay or losing to it.

Decay also concentrates in specific segments. Measure change against a dated baseline instead of applying a general decay benchmark, then set the review cadence from the changes observed in your own system. Contacts from purchased lists decay faster than inbound signups because the data was already stale when you bought it. When you run your quarterly deep dive, segment by record age and source. The averages hide where the rot is. We covered the ongoing maintenance side of this in our CRM hygiene guide .

The Operational Metrics

Beyond the core data metrics, track metrics that show data quality impact on operations.

Metric Summary: Targets and Cadence

Metric	Target	Check
Email validity (active contacts)	95%+	Monthly
Email validity (full database)	85%+	Quarterly
Duplicate rate	Under 5%	Monthly
Routing field completion	95%+	Weekly at high volume
Contact-account association	95%+	Monthly
Hard bounce rate	Under 2%	Per campaign
Routing exception rate	Under 5%	Weekly

Building a Dashboard

Put these metrics on a single dashboard that gets reviewed regularly. Here's a structure that works:

Weekly Review (Ops Team)

• Campaign bounce rates from past week
• Routing exceptions from past week
• New duplicates created (if trackable)
• Any sudden changes in key metrics

Monthly Review (Leadership)

• Email validity rate (trend)
• Duplicate rate (trend)
• Routing field completion (current)
• Contact-account association rate (current)
• Month-over-month changes with context

Quarterly Deep Dive

• Full data quality audit with detailed metrics
• Decay rate calculation for the quarter
• Segment-specific analysis (by source, by age, by segment)
• Recommendations for next quarter's priorities

Setting Targets

The targets above are reasonable benchmarks, but your specific targets should consider:

• Where you are now: Use the dated baseline to set an incremental target that the current workflow can measure.
• Your tolerance for operational issues: Some businesses can tolerate more routing exceptions than others.
• The effort to improve: Review the exceptions remaining after each pass and stop when the next increment costs more than the workflow benefit.

What Not to Measure

Avoid metrics that have no owner, no decision threshold, or no connection to a workflow your team can change.

Some commonly tracked metrics aren't worth the effort:

• Overall "completeness score": Too vague. Completion of what fields? Weight fields by their operational importance.
• Format compliance: Unless bad formatting breaks something, it's not worth tracking.
• Data age: Age alone doesn't mean bad. A 3-year-old contact could be perfectly accurate.
• Record counts: Duplicate or invalid records can make a larger database less useful.

Focus on metrics tied to outcomes. If you can't explain how a metric affects operations, you probably don't need it.

Using Metrics to Prioritize

Once you have baseline metrics, use them to decide what to fix first:

1. Fix what's broken: Prioritize the metric tied to an observed workflow failure, such as rejected mail or unresolved routing.
2. Fix what blocks operations: Repair missing routing fields before lower-impact formatting inconsistencies.
3. Maintain what's working: Keep dated baselines and alert when a stable metric begins to drift.
4. Improve for optimization: Once basics are stable, improve metrics that drive better performance, like enrichment coverage and segmentation correctness.