How to Measure CRM Data Decay

You verified those email addresses two years ago. They were good then. Marketing automation was humming, deliverability was solid, sales had accurate contact info for their accounts.

Now bounce rates are climbing. Reps are calling disconnected numbers. Half the contacts in your key accounts have moved on, and you're not sure which half.

This is data decay, and it happens to each database regardless of process. People change jobs, companies get acquired, email addresses stop working. A dated validation run can reveal which addresses, employers, phones, and classifications changed since the prior snapshot. That measured difference belongs to the cohort and interval tested.

Use two dated snapshots to identify what changed by field and segment. Report the observed interval and denominator with the result; a change measured in one cohort should not be presented as a universal annual decay rate.

What Does CRM Data Decay Look Like?

Decay shows up in three concrete workflows. Measure each one from observed activity and exceptions.

Job Changes

Every job change means:

• The email address you have stops working
• The phone number (if it was a work number) no longer reaches them
• The job title is wrong
• The company association is wrong
• The buyer persona mapping is wrong

A CMO who leaves their company becomes a different opportunity entirely. They might be at a better-fit company now. Or a worse one. You won't know unless you update the data.

Company Changes

Companies don't stay static either. Each year, thousands of businesses:

• Get acquired (new parent company, often new domain)
• Merge (which company survives? which domain?)
• Rebrand (domain changes, company name changes)
• Go out of business (everyone's email stops working)
• Get restructured (divisions spin off, get absorbed)

When a company you sell into gets acquired, its email domain often gets folded into the parent's over the following year. Your contacts don't send you a notice. The data just silently goes wrong.

Technical Changes

Even when people stay at their jobs and companies stay stable, technical infrastructure changes:

• Email domains change (company rebrands, consolidates)
• Phone systems change (new provider, new numbers)
• Office locations close (addresses become invalid)
• LinkedIn URLs change (people update their profile slugs)

The Compound Problem

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.

Cohort	Measure	Decision
Recently created records	Field validity at capture and review date	Confirm source and intake controls
Older active records	Changes since the dated baseline	Set the next field-level review
Inactive records	Current business purpose and reachable fields	Retain, suppress, archive, or remove under policy
High-value exceptions	Conflicts and missing evidence	Route to an accountable reviewer

A CRM can contain records from different periods, sources, and workflows. Age alone does not establish validity. Test the fields that matter for the next action, retain the check date and method, and keep uncertain records separate from confirmed failures.

The Hidden Costs

Bad data quietly makes everything downstream worse.

Marketing Waste

Email to invalid addresses bounces. That's obvious. But the less obvious spend is mailbox acceptance. High bounce rates damage your sender reputation. Your emails start landing in spam, even for valid addresses. You're paying for email marketing that nobody sees.

Marketing automation sequences run against people who left the company years ago. Lead scoring gives points for engagement that will never happen. Campaign metrics are inflated by unreachable contacts.

Sales Waste

Reps call numbers that don't connect. They send emails that bounce. They research accounts based on outdated information.

Worse, they lose trust in the CRM. When data is wrong often enough, reps stop trusting any of it. They maintain their own spreadsheets. They don't log activities. The system becomes a compliance exercise instead of a useful tool.

Strategic Waste

Your ICP analysis, market sizing, and account scoring depend on CRM fields. Measure the incorrect or stale share by field before trusting an analysis built on those records.

You might be targeting the wrong segments because the data told you those segments converted well, when data quality issues were masking the truth.

How to Measure Your Decay Rate

Before you can fix the problem, you need to understand how bad it is.

Run an Email Validation

Take your email list and run it through a validation service like NeverBounce, ZeroBounce, or Kickbox. They'll tell you what percentage of your emails are:

• Valid: Deliverable, working addresses
• Invalid: Hard bounces, dead addresses
• Risky: Domains with domain-wide acceptance, temporary addresses
• Unknown: Servers that didn't respond

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.

Check Last Activity Dates

In Salesforce, report on contacts grouped by their last activity date:

• No activity inside the team's first review window
• No activity inside the team's second review window
• No activity inside the team's archive-review window
• Never any activity

Last activity is a useful sampling field, but inactivity alone does not prove a record is wrong. Validate representative records in each dated cohort and report the observed invalid share with the cohort definition.

Sample and Verify

Take a representative sample from the record-age cohort you want to test. Verify the email domain, current company, and phone status, then record the result and source date for each field.

This is tedious, but it gives you ground truth. 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.

What to Do About It

Decay itself can't be stopped. What happens to your database after it starts is up to you.

Establish a Baseline

Run the assessments above. Know your current state: email validity rate, percentage of stale records, estimated decay by age cohort. You can't improve what you don't measure.

Clean the Backlog

If you haven't done a serious data cleaning in years, start there. Validate emails. Remove or archive dead records. Enrich gaps in firmographic data. This is a one-time project to get back to a usable baseline.

Implement Ongoing Maintenance

Quarterly data quality audits at minimum. Check email validity before major campaigns. Re-enrich key accounts annually. Set up processes to catch decay before it compounds.

Consider Automated Enrichment

Some tools can continuously monitor your CRM and update records when they detect changes (job changes on LinkedIn, company acquisitions, etc.). If you have budget for it, automated enrichment can materially slow decay.

Accept What You Can't Control

Define field-level acceptance rules for the workflow and measure the submitted cohort. The useful target is an agreed, dated quality threshold with a review path for exceptions.

Turn the baseline into a queue

After sampling, convert each observed failure type into a named review queue with an owner and disposition field. Keep corrected, unresolved, and intentionally retained records separate. That makes the next measurement comparable and prevents the headline decay rate from hiding a concentrated problem in one source, territory, or workflow.

Common Questions

Related: How to Clean Salesforce Data | Data Enrichment Services | Email Enrichment