Blog · crm-hygiene
How fast does CRM contact data go stale? Measure your own rate
No measured decay rate fits every CRM. Official job data: 14.6% of German and 20.6% of US employees were new to their employer within a year. Measure yours.
of US wage and salary workers had been with their employer a year or less (January 2026)
U.S. Bureau of Labor Statistics, 2026of employees aged 20–64 in Germany had been in their job 12 months or less in 2025
Eurostat, 2026of US workers aged 55–64 were new to their employer, against 25.5% at 25–34
U.S. Bureau of Labor Statistics, 2026of 10-person buying-group maps lose someone within a year at a 15% rate (Illustrative)
Panelhop, 2026How fast does B2B contact data go stale?
No measured rate fits every CRM. Official job data give a sourced base rate for a single cause: people changing employer. In January 2026, 20.6% of US wage and salary workers had been with their current employer a year or less.↗ In 2025, 14.6% of employees aged 20–64 in Germany had been in their current job 12 months or less, against 13.7% across the EU.↗
The one direct measurement of B2B contacts we found comes from a data vendor, so read it as a vendor’s count. Lusha checked its own database: 12.25% of 140,284 US sales leaders at VP or C-suite level had an employment change recorded within 12 months.↗ So did 17.2% of a German cohort of 4,204 and 21.40% of a UK cohort of 4,214.↗ Lusha says of its own figure: “This is a job change rate, not a full decay rate.”↗
Treat all of these as proxies. Tenure counts people starting their first job or returning to work, so it overstates job switching. It also misses new roles at the same employer and people leaving work altogether. Renamed or acquired companies and changed phone numbers or email formats add more decay on top. The calculations below use 15% a year, close to the German figure.↗ Anything built on that rate is Illustrative.
Where do the decay rates AI quotes come from?
They come from vendor pages and a 2022 survey forecast, not from a measurement. We captured 5 AI answers on 6 October 2026.↗ Only Google AI Mode gives a rate: “B2B data naturally decays by 22%–34% every year due to job changes, company renames, and acquisitions.”↗ The AI Overview, Gemini and Perplexity give hygiene checklists without one.↗
AI Mode cites Default and ZoomInfo. Default’s page gives no decay figure.↗ ZoomInfo’s says “B2B contact data decays at 22-30% annually” without a source.↗ It also puts the annual rate at about 34%, citing a Validity survey.↗ That survey, published in February 2022, covered 1,241 CRM administrators in the US, UK and Australia.↗ They estimated their data quality would degrade by 34% by the end of 2022 if their company didn’t invest.↗ So the 34% is a forecast by respondents about their own data.↗
The other common figure, 22.5%, comes from HubSpot’s Database Decay Simulation.↗ It compounds 2.1% a month, credited to Marketing Sherpa research with no year or method.↗ Our trace in the CRM cleanup post follows each figure back, and none reaches a measured yearly rate.↗ Even a measured rate would date. The share of US wage and salary workers new to their employer fell from 22.2% in January 2024 to 20.6% in January 2026.↗
| Figure | Where it comes from | What it measures |
|---|---|---|
| 22.5% a year | HubSpot’s Database Decay Simulation: 2.1% a month, credited to Marketing Sherpa research with no year | Not stated |
| 22–30% a year | ZoomInfo’s CRM hygiene page, updated July 2026 | No source given |
| 34% a year | Validity survey of 1,241 CRM administrators, published February 2022 | Respondents’ forecast of their own data’s decline by the end of 2022 without investment |
| 22–34% a year | Google AI Mode, 6 October 2026, citing Default and ZoomInfo | Spans ZoomInfo’s 2 figures; Default’s page gives none |
| 12.25% US, 17.2% Germany, 21.40% UK | Lusha, September 2026 | Employment changes within 12 months among sales leaders at VP or C-suite level, counted in Lusha’s own database |
| 20.6% US, 14.6% Germany | BLS, January 2026; Eurostat, 2025 | Employees with 12 months or less at their current employer, from national labour force surveys |
What does official job data show by country and age?
It shows a base rate that can nearly double between European countries and falls steeply with age.↗↗ In 2025, 18.8% of employees aged 20–64 in the Netherlands had been in their current job 12 months or less, against 10.5% in Poland.↗ Switzerland stood at 16.9%, Austria at 14.8% and Germany at 14.6%.↗ A database weighted to the Netherlands should decay faster than one weighted to Poland, whatever a vendor’s global rate says.
About 1 in 7 employees in Germany had started their job in the past 12 months; in the Netherlands, nearly 1 in 5
Data behind this chart
| Item | Value |
|---|---|
| Netherlands | 18.8% |
| Switzerland | 16.9% |
| Sweden | 15.5% |
| Austria | 14.8% |
| Germany | 14.6% |
| EU-27 | 13.7% |
| Poland | 10.5% |
Age moves the rate even more. In January 2026, 25.5% of US wage and salary workers aged 25–34 had been with their employer a year or less.↗ The share falls to 15.4% at 35–44, 12.2% at 45–54 and 9.2% at 55–64.↗ Weighting those groups by their size gives about 16.4% for ages 25–64, between the German and Dutch figures.↗ Among the major occupation groups, management, professional and related occupations had the longest median tenure: 4.9 years, against 4.1 for all workers.↗
In the US, 1 in 4 workers aged 25–34 had been with their employer a year or less, against 1 in 11 at 55–64
Data behind this chart
| Item | Value |
|---|---|
| All aged 16+ | 20.6% |
| 25–34 | 25.5% |
| 35–44 | 15.4% |
| 45–54 | 12.2% |
| 55–64 | 9.2% |
The US and European figures aren’t strictly comparable. BLS counts wage and salary workers aged 16 and over; Eurostat counts employees aged 20–64.↗↗ Both measure time with the current employer, though. Expect a CRM of senior buyers to decay more slowly than one of junior evaluators.
What does decay do to a 10-person buying-group map?
At the rates official job data suggest, most maps lose someone within a year. Assume each person in a 10-person map leaves with a 15% chance a year, independently of the others.↗ Then about 80% of maps lose at least 1 person within 12 months (Illustrative).↗ The average map loses 1.5 people a year, and after 2 years about 96% are missing someone.↗ At the US all-worker rate of 20.6%, the 1-year share rises to about 90%; at 10%, it is about 65%.↗
We use 10 people because 6sense puts the typical buying group at around 10 members.↗ Our post on buying-group size argues for mapping at least that many at €25k and above. A bigger map loses someone sooner.
A rep works from the map. It tells them whom to call, and a departed CFO in it keeps drawing tasks, sequence steps and ads until someone notices. The move can help too: a contact who changes employer takes what they know about you to a new account.
How do you measure your own decay rate?
Take a random sample of 100 contacts and check each against a current source. Count how many changed employer or role since they were last verified, then annualise the share. Draw the sample from target accounts and from the roles your routing, scoring and buying-group maps read. That is where a stale record does damage. A person does the checking: the company’s own website, a bounced email, a call to reception or a public profile all count as evidence.
- 01Step 1
Draw 100 contacts
Pick them at random from target accounts, in the roles your routing, scoring and buying-group maps read. Note when each was last verified.
- 02Step 2
Check each one by hand
Use a current source: the company’s website, a bounce, a call to reception or a public profile. Mark each as same role, new role, left or can’t tell.
- 03Step 3
Count and annualise
Divide new roles plus leavers by the contacts you could check. If they were last verified m months ago on average, the yearly rate is 1 − (1 − share)^(12 ÷ m).
- 04Step 4
Split it and repeat
Split the rate by country and seniority. Repeat on a fresh sample each quarter: the first result is your baseline.
An example: if 6 of 100 contacts last verified 4 months ago have moved, the yearly rate is 1 − 0.94³, about 17% (Illustrative).↗ With 100 contacts, a rate near 15% is known to within about 7 points either way.↗ That is precise enough to plan on.
Count new roles at the same employer as changes. The official data miss them, and in a buying-group map a new role can change who decides. Leave out the contacts you can’t check, and note how many there were: a high share is a finding in itself.
Re-verify the contacts your routing and scoring read first
We think the first contacts to re-verify are the ones an automation reads, and the rate to plan on is your own. A borrowed rate can’t tell you which records are wrong. A scoring rule, a routing rule or a sequence acts on specific fields. A departed contact in those fields does damage at machine speed. The account score still counts a buyer who has left, and the sequence keeps writing to an inbox nobody reads.
One fact sits behind that view: in Validity’s 2025 survey, 76% of CRM users said less than half of their organisation’s CRM data is accurate and complete.↗ We don’t think hygiene has to be perfect before anything else runs. In our view, only the fields a specific automation depends on need to be reliable first. For routing, that means email and owner; for scoring, title and seniority; for the buying-group map, employer and role.
We also think your measured rate belongs in the baseline before you buy enrichment. Without it, a vendor’s freshness claim can’t be checked. With it, you can run the vendor on the same 100 contacts you checked by hand and count how many it gets right. We’d call a results claim without a baseline a guess, whoever makes it.
In practice
How we do it at Panelhop
In a Panel Check (GTM audit · 2–3 weeks), check C1 measures the stale contacts in your CRM export, those that bounced or show no activity and no update for 12 months, next to the duplicate rates. The result goes into your baseline as part of the CRM hygiene score.
Where routing, scoring or buying-group maps read stale records, the CRM hygiene module of a Leak Fix (we build the fixes) settles with your team what counts as stale and what happens to it: verify, archive or keep. It adds a scheduled job that flags stale records to their owner, and Panel Ops (we run it monthly) reports the score against the baseline every month. In Signal Desk (in-market accounts, weekly), a champion’s job change counts as a signal, for contacts already in your CRM. The 100-contact check above works with us or without us.
Questions buyers ask about this
What counts as a stale contact?
A record that no longer points to the right person in the right role. The person may have left the company or changed role, or their email or phone may have changed. Official job data only cover people changing employer, so count role changes too when you measure your own rate.
Does our email bounce rate show how fast contacts go stale?
Only in part. A bounce shows that a mailbox has closed, but a new role at the same company never bounces, and some companies keep old mailboxes open for a while. Use bounces as one piece of evidence in the sample check, not as the rate.
How often should we re-verify CRM contacts?
Set the cadence from your measured rate and from what each contact feeds. We’d check contacts in open deals and Tier 1 buying-group maps every quarter and sample the rest once a year. Re-measure the rate each quarter, so the cadence follows your data.
Should we buy an enrichment tool to fix decay?
Measure your own rate first. Then run the tool on the same 100 contacts you checked by hand: the share it gets right is its accuracy on your data. Compare that with what checking by hand costs you before you decide.
