Part of Lusha’s ongoing measurement of B2B data decay. Every finding in one place: How fast B2B contact data decays.
Every rate in this report comes from measurements Lusha has already published, linked where they appear. Every competitor statement comes from the vendor’s own page, linked, checked September 13, 2026. No new contact data was pulled for this piece.
Every data platform says it refreshes continuously. None of them tells you what that means: how many records change in a month, which fields change, how long it takes a change to be detected, and how often a refresh has to run before the CRM is wrong again. This report answers those four questions with numbers, then shows what each platform actually publishes about refresh.
TL;DR
- How fast records go stale: 12.6% of senior US sales contacts changed roles in twelve months, 25.7% at 24 months. About 1% of a working list a month. UK lists move about 1.7 times faster.
- What goes stale: the job title and the employer, not the email or the phone. In Lusha’s US test, every case where a record was out of date was a title or company change; the contact details were still correct.
- Why “continuous” is not a number: changes surface with a lag. Lusha’s own three month measurement returned half the monthly rate of the longer windows, because the newest changes had not been detected yet. Any vendor’s refresh, including Lusha’s, is behind the real world by weeks.
- How often to refresh: monthly for contacts on active deals, quarterly for territory lists, at send time for campaign lists. At 1% a month, 1,000 active contacts produce about 31 wrong records a quarter.
- What the platforms publish: Lusha publishes a measured decay rate and a weekly signal refresh. Cognism publishes a monthly enrichment share. ZoomInfo publishes a monthly job change processing volume. Apollo and UpLead describe real time verification and publish no rate or cadence.
Published: September 2026.
Data sources: Lusha decay measurement, 148,000 records, rerun September 12, 2026. Lusha Q2 2026 Contact Change Report, 1,470,414 company changes. Lusha US accuracy tests, 322 contacts, July 28, 2026. 290M+ verified contacts in the live database.
Part one: how fast CRM records go stale
The figure that circulates is 30% a year. We could not find a primary source for it, so we measured our own records instead.
Source: How fast does B2B contact data decay?, rerun September 12, 2026. Method, cohorts and limitations on that page.
At the database level the movement is larger in absolute terms. Lusha’s Q2 2026 Contact Change Report counted 1,470,414 company changes and 194,165 promotions between January 1 and June 1, 2026: 13,600 contacts changing jobs every working day.
For a CRM, the cohort rate is the one to plan with. Nobody’s CRM looks like the whole database. It looks like a few functions, a few seniority bands and a few countries, and those are the segments that move fastest. At about 1% a month, a team holding 1,000 contacts on active deals should expect roughly 31 of them to become wrong each quarter.
Part two: what actually goes stale
Refresh is usually sold as re verifying emails and phone numbers. That is the part of the record that holds up best.
In Lusha’s US accuracy test, 322 contacts, every case where the search result and the enrichment call disagreed was a job title or a current employer. A VP at a SaaS company was now a partner at a venture firm. A director of operations was now a founder and CEO somewhere else. In each case the work email and the mobile phone Lusha returned were correct and current. The contact details had survived the job change. The snapshot of where the person sat had not.
The change report adds a second kind of staleness that no bounce will ever reveal. For every contact promoted in place, 7.6 changed companies, and a promotion does not break the email address. A VP of Sales who is now CRO is still reachable. The personalization, the score and the routing are all wrong, and the record looks fine.
A refresh that only re verifies contact details is checking the fields that were already right.
Part three: why “refreshed continuously” is not a number
Every vendor in this category, Lusha included, describes its data as continuously updated. That describes a process. It does not tell you how far behind the world the data is.
Lusha’s own measurement shows the gap. The three month window in the decay study returned about half the monthly rate of the twelve and twenty four month windows. That is not because fewer people changed jobs in the most recent quarter. It is because the newest changes are the ones least likely to have surfaced yet: the person has not updated a profile, the company has not announced the hire, the old email has not started bouncing. Detection lags the event, and the lag is weeks, not hours.
Two consequences for a CRM owner.
First, a recent refresh date is weaker evidence than it looks. A record refreshed last week against a source that has not yet detected last month’s move is still wrong, with a fresh timestamp on it.
Second, no vendor can honestly promise that a record is current. What a vendor can do is publish the rate at which its records go stale and the cadence at which it re checks them. That is what Part four compares.
Part four: what each platform publishes about refresh
Checked on each vendor’s own site, September 13, 2026. Where a vendor publishes no figure for a row, the table says so.
Three things to read from the table.
Only one of the five publishes how fast its own records go stale. That is the number a CRM owner needs to set a cadence, and four vendors leave you to guess it.
Two vendors publish a volume or a share instead of a rate. ZoomInfo’s 15 to 20 million monthly job change events and Cognism’s 5% monthly enrichment are real commitments, and neither tells you what share of your list will be wrong in a quarter.
Two vendors describe refresh as real time verification of contact details. That is the field that held up in the test. Neither statement covers the title and employer changes that Part two shows are the ones that actually go stale.
Part five: how often a refresh has to run
Three tiers, derived from the 1% a month rate and the finding that titles go stale before contact details.
Active deal contacts: monthly. A departed champion discovered mid deal costs more than twelve checks a year. Check the title and the employer, not only the email.
Territory lists: quarterly, adjusted by region. A quarterly pass catches roughly 3% of a US list. A UK list moves about 1.7 times faster, so the same cadence leaves it in worse shape at every point in the cycle.
Campaign lists: at send time. A list that was accurate 60 days ago is about 2% wrong. At 10,000 contacts that is roughly 200 invalid sends, and the ones that arrive with a stale title in the first line do more damage than the ones that bounce.
Whatever the vendor’s cadence, the detection lag in Part three means the campaign tier belongs at send time and not on a schedule.
Part six: measure your own CRM’s stale rate
Take your active contact list size. Multiply by 0.0105, the measured monthly rate for senior sales contacts. Multiply by the months since your last full verification pass.
50,000 contacts, six months since the last pass: 50,000 × 0.0105 × 6 = about 3,150 contacts likely to have changed companies. Multiply by 1.7 for a UK list. If your list skews junior, the figure will overstate it, because the measurement covers VP and C suite.
Then check it against reality rather than trusting the estimate. Three plays run the check live with the Lusha connector:
- Refresh stale CRM records before you reach out
- Find every account where the main contact has gone dark or left
- Build the quarterly account contact change log
- For a comparison of the enrichment tools themselves, by CRM integration, see the guides for Salesforce, HubSpot and Pipedrive users.
Run these live: add the Lusha plugin to Claude →
Methodology and limitations
Decay rate. Lusha cohort measurement, US sales leaders at VP and C suite, 148,000 records, tracked at 3, 12 and 24 months, rerun September 12, 2026. Published with cohort sizes and limitations at How fast does B2B contact data decay?
Change volumes. Lusha live signal database, queried June 1, 2026, changes detected January 1 to June 1, 2026. Published at the Q2 2026 Contact Change Report.
What goes stale. Two US tests, 322 contacts, July 28, 2026, hit defined as a real enrichment call returning a usable value. Title and employer discrepancies observed in the name brand sample of the metros test. This is a directional finding from a small number of cases, published as such.
Competitor statements. Each vendor’s own site, linked in the table, checked September 13, 2026. Where a vendor publishes no figure for a row, the table says Not stated as of that date. Statements were not converted or compared across different definitions.
What this cannot tell you. The decay rate covers one seniority band in two countries and will overstate movement for junior lists. Detection lag means every recent figure, including Lusha’s, is a floor. This report does not measure how quickly any vendor, including Lusha, detects a change after it happens; no vendor publishes that, and it is the number this category should be asked for.
Privacy. Aggregate figures only. No individual contact is named or identifiable. Data accessed under Lusha’s certified compliance framework; certifications are listed on the trust center.
FAQ
Which platforms refresh stale CRM data?
Lusha, ZoomInfo, Apollo, Cognism and UpLead all describe continuous or real time refresh. Only Lusha publishes the rate at which its own records go stale, 12.6% a year for senior US sales contacts, with the method. ZoomInfo publishes a monthly job change processing volume, Cognism a monthly enrichment share, and Apollo and UpLead describe real time verification without a rate or cadence.
How often should CRM data be refreshed?
Monthly for contacts on active deals, quarterly for territory lists, at send time for campaign lists. The cadence follows from a measured rate of about 1% of senior contacts changing roles each month, and from the finding that titles and employers go stale before emails and phones do.
What does CRM data refresh actually mean?
It depends on the vendor. Some mean re verifying email deliverability at export. Some mean processing job change events and updating the record. The field that goes stale first is the job title and the current employer, so a refresh that covers only contact details is re checking the part of the record that was already right.
How fast does CRM data go stale?
Lusha measured 12.6% of senior US sales contacts changing roles in twelve months and 25.7% in twenty four, about 1% a month. UK sales leaders changed at 21.7%, about 1.7 times the US rate. The widely repeated 30% a year figure has no primary source that Lusha could find.
Why does a recent refresh date not guarantee a record is current?
Because detection lags the event. A person who changed jobs last month may not have updated a profile, and the company may not have announced it. Lusha’s three month measurement returned half the monthly rate of the longer windows for this reason. A record refreshed last week against a source that has not yet detected the change is still wrong.
How do I find out how stale my own CRM is?
Multiply your active contact count by 0.0105 and by the months since your last full verification pass for an estimate, then run a real check on a sample. Take 100 records, enrich them, and count how many came back with a different title or employer. Your own result is worth more than any vendor’s average, including the one in this report.
Related resources
- How fast does B2B contact data decay?: the 12.6% measurement behind this report
- Q2 2026 Contact Change Report: 1,470,414 company changes, 13,600 a working day
- How accurate is B2B contact data in the US?: the test that found titles go stale before contact details
- Best sales intelligence software for US data accuracy and coverage: what each vendor publishes about accuracy
- 10 best data enrichment tools in 2026: which tools do the enrichment, compared
- What keeps CRM data from going stale: the operating practices
Every rate in this report is time sensitive. Run a live check on your own list for the current state.
