How fast B2B contact data decays: what the research shows

B2B contact data decays at roughly 22 to 40 percent per year, with email-specific decay running about 2.1 percent per month (HubSpot Database Decay) and accelerating to 3.6 percent per month in fast-moving sectors (RevenueBase analysis of B2B email decay). That means a database you bought in January is missing or wrong on roughly one in three records by December, and the loss compounds while you keep paying for it. The mechanism is simple: people change jobs, companies merge or close, offices move, mailboxes get deactivated 30 to 90 days after departure, and any snapshot taken before those events is wrong the moment they happen.

TL;DR

• Industry research puts B2B contact decay at about 22.5 percent per year on average, climbing past 35 percent in tech and senior roles (HubSpot, Cognism C-suite decay data).
• The U.S. Bureau of Labor Statistics reports median job tenure at 3.9 years in January 2024, the lowest in two decades, with workers ages 25 to 34 sitting at just 2.7 years (BLS Employee Tenure 2024).
• ZoomInfo's marketed 95 percent accuracy is a contact-to-company match rate, not an email deliverability rate; independent reviews put real-world email accuracy at 75 to 85 percent (Prospeo independent review).
• 44 percent of CRM users say their company loses more than 10 percent of annual revenue to poor data quality, according to a Validity survey of 1,200+ users (VentureBeat coverage of the Validity report).
• The fix is not refreshing the snapshot faster, it is replacing the snapshot with the underlying events: hiring posts, review velocity, location changes, and re-verifications at delivery time.

Where the decay number comes from

Three independent lines of evidence converge on roughly the same range.

The first is direct database measurement. HubSpot's long-running database decay simulation models contact lists at 22.5 percent annual decay, derived from average B2B contact churn observed across HubSpot CRM tenants (HubSpot Database Decay). RevenueBase's 2024 analysis tracked email-specific decay and found it accelerating to 3.6 percent per month, which compounds above 35 percent annually in fast-moving sectors (RevenueBase).

The second is labor-market data. The U.S. Bureau of Labor Statistics reports median job tenure dropped to 3.9 years in January 2024, with workers ages 25 to 34 at just 2.7 years (BLS). JOLTS quits data shows a sustained 3.1 to 3.3 million workers per month leaving their jobs in mid-2024 (BLS JOLTS archives). If a worker who leaves makes their old email a hard bounce within 30 to 90 days, those quit numbers translate directly into contact-record decay.

The third is the vendors' own admissions. Cognism's research on C-suite records reported 35 percent annual decay for CMOs, 34 percent for CROs, 32 percent for CFOs, and 26 percent for CEOs (Cognism newsroom). ZoomInfo markets a 95 percent contact-to-company match accuracy, but independent testing by Prospeo and others puts real-world email deliverability at roughly 75 to 85 percent, which is consistent with a database that decays faster than its refresh cycle (Prospeo). For a deeper side-by-side, see our Apollo vs ZoomInfo breakdown.

The honest summary: 22 percent on the optimistic end, 40 percent on the realistic end, with senior roles and high-turnover industries running above that.

Why static databases decay so fast

Decay is not a database hygiene problem. It is a category problem. The world keeps moving while the snapshot stays still. Four mechanisms drive most of it.

Job changes. With median tenure at 3.9 years, roughly a quarter of workers move in a given year (BLS). Most B2B prospect databases tag a person to a company at the time of last enrichment. The day they leave, the email begins forwarding, then bounces, and the company-affiliation field is wrong even when the email still resolves.

Email rotation and deactivation. Corporate IT typically deactivates a departed employee's mailbox within 30 to 90 days, sometimes routing to a manager during the interim (RevenueBase). After deactivation, sends produce hard bounces, which damage sender reputation past a 2 percent threshold and risk blacklisting past 5 percent.

M&A and corporate restructuring. S&P Global recorded $1.7 trillion in North American M&A deal value in 2024, a 9 percent year-over-year increase (S&P Global Market Intelligence). Each deal triggers domain consolidation, role eliminations, and rebrands, all of which invalidate prior contact records.

Role and seniority changes. Promotions and lateral moves inside the same company are not captured by simple email-validity checks. The address still resolves, but the title and decision-making authority have shifted, which is exactly what makes a contact actionable for B2B sales.

The common thread: every one of those mechanisms starts with an observable event in the world (a hiring post, a press release, a Maps listing edit, an MX-record change). The static database is always reading the world late. That is also why understanding what makes a buying signal high-intent matters: the same events that decay your contact list are the ones that flag your next prospect.

What gets stale first

Not every field decays at the same rate. The faster a field changes in the real world, the faster the snapshot fails.

Data type	Approximate annual decay	Example mechanism
Personal email at company	25 to 40 percent	Job change deactivates the mailbox within 30 to 90 days (RevenueBase)
Job title at current employer	30 to 35 percent	Internal promotions and lateral moves; not visible to email-validity checks (Cognism)
C-suite identity (CMO, CRO, CFO)	26 to 35 percent	Senior turnover runs above the average (Cognism)
Direct dial phone number	15 to 25 percent	Number reassignment, role change, switch to mobile
Company name and domain	5 to 10 percent	M&A and rebrands (S&P Global)
Company headquarters address	5 to 10 percent	Office relocation, hybrid consolidation
Industry / NAICS classification	Under 5 percent	Mostly stable unless the business pivots

Email and title decay together account for the majority of the operational pain. They are also the two fields static databases are worst at refreshing, because both require an event observation, not a record lookup.

The signal-based alternative

If decay is caused by events, the only durable fix is to read the events directly instead of buying a snapshot of their last consequence.

A signal-based approach inverts the workflow. Rather than store a list of contacts and hope it stays fresh, you watch the world for the events that produce qualified prospects (a multi-location SMB adding a new branch on Google Maps, a sustained spike in review velocity, a new hire announcement, a domain change), then resolve the contact at the moment you act on the signal. The contact record is generated when it is needed, verified at delivery time, and discarded if the signal goes cold. There is no list to decay because there is no list.

This is the thesis behind Keendai and the architecture of how Keendai works: mine live signals from Google Maps and similar live sources, enrich the matched company with Apollo and Hunter at retrieval time, re-verify email deliverability with Hunter on the day of export, and attach a plain-English score that cites the actual signal. Every record carries a source URL and a timestamp so you can see exactly how recent the underlying observation is.

The trade-off is real and worth being honest about: signal-based prospecting produces fewer contacts than a 200-million-record static database. It also produces contacts that are current, sourced, and explainable. For a wide-net cold blast at the lowest possible cost per record, a static database still wins on volume. For follow-up-able, deliverable, in-context prospects, the math flips fast once you account for the bounce rate, the wasted SDR time, and the sender-reputation damage from sending into decayed addresses.

The static-vs-signal trade-off, head to head

Dimension	Static database (Apollo, ZoomInfo, etc.)	Signal-based prospecting
Freshness mechanism	Periodic re-crawl of public sources	Event observation at retrieval time
Annual decay impact	22 to 40 percent of records go bad (HubSpot)	Records are generated on demand, no backlog to decay
Email accuracy in practice	75 to 85 percent (Prospeo)	Re-verified at delivery; rejects flagged before export
Price model	Per-seat or per-credit, paid up front	Per-lead, paid only on what you keep
Best for	Wide ICP, low signal sensitivity, high volume	Narrow ICP, high signal sensitivity, deliverability matters
Hidden cost	Bounce rate, SDR time, sender reputation	Smaller raw volume per query

For Keendai's per-lead model and free tier, see Keendai pricing.

Frequently asked questions

What is a realistic B2B contact data decay rate per year?

Industry research puts the average between 22 and 30 percent per year, with senior roles, tech, and high-turnover industries running 35 to 40 percent or higher (HubSpot, Cognism). A 30 percent annual figure is a reasonable single-number benchmark for planning, with the caveat that your specific ICP could be materially higher.

Why do databases like Apollo and ZoomInfo still claim high accuracy?

Their published accuracy numbers are usually contact-to-company match rates, not email deliverability rates. ZoomInfo's marketed 95 percent figure is a match accuracy claim (Prospeo). Independent testing puts real-world email deliverability lower, which is consistent with the rest of the industry research on decay. Both numbers can be true at the same time.

How much revenue does bad contact data actually cost?

Gartner's cross-industry estimate is $12.9 million per year per organization in losses tied to poor data quality (Gartner via LinkedIn). A separate Validity survey of more than 1,200 CRM users found 44 percent of companies lose over 10 percent of annual revenue specifically to poor CRM data quality (VentureBeat).

Does a higher refresh frequency fix decay?

It helps but does not solve it. The faster you refresh, the closer your snapshot is to current, but you are still paying for snapshots. If a contact changed jobs the day after your refresh, you carry the wrong record for the entire next refresh window. Live signal mining sidesteps the refresh-cycle question entirely by reading the events as they happen.

What signals are most useful for replacing decayed contact data?

The most useful signals are the ones that anticipate or coincide with the changes that cause decay: new hire announcements, location additions on Google Maps, sustained review-velocity spikes, ad activity changes, and verified email re-checks at the moment of send. See anatomy of a buying signal for a longer breakdown of what separates a real signal from a vanity metric.

Should I keep my static database subscription?

It depends on your motion. If you run high-volume, low-touch outbound to a wide ICP, a static database may still be the cheapest source of raw contact records, with the understanding that you will pay for the bounce rate. If your motion is narrower or your sender reputation matters (it does), the math usually favors a signal-based source for primary discovery and a static database for backfill, not the other way around.

---

Ready to test signal-based prospecting against your current database? Start with 50 free leads.

Last updated: May 2026