Contact Data Decay: Why Some Lists Go Stale Faster
B2B contact data decays 22% to 30% a year, but not evenly. Learn which segments go stale fastest, why standard refresh methods miss them, and how to test your list.
Most conversations about B2B data quality start with coverage. Can the vendor find your buyers? How many records do they hold? Coverage is the easiest thing to compare across vendors, so it is what gets compared.
It is also largely settled. You can buy a list of almost any audience now, including several that used to be genuinely hard to reach.
The expensive problem is what happens to that list afterwards. And the part that catches teams out is that it does not happen at the same rate to every list you own.
The cost is decay, not coverage
Published estimates for B2B contact decay do not agree with each other, and the spread tells you more than the midpoint does.
- 22.5% a year: the most commonly cited figure, derived from a monthly rate of roughly 2.1% compounded across twelve months.
- Over 30% a year: a common estimate for B2B databases specifically, attributed to the pace of job changes and corporate restructuring.
- 70% a year: circulates widely, but tends to count more fields than contact detail alone, and usually traces back to a vendor selling the remedy.
Take 25% as a working assumption. The important property is that it compounds on whatever is still accurate, so a file of 100,000 good records is down to something like 56,000 inside two years without anyone doing anything wrong.
That is the number most teams plan around. It is also the wrong number to plan around, because it is an average across the whole file, and your file is not one population.
Why some segments decay faster than others
Ask how a contact record actually gets corrected and the answer is usually a version of the same thing. Somebody updated a professional profile, and that change was picked up.
That mechanism carries a dependency almost nobody examines. It requires the person whose record has gone stale to perform an act of self-maintenance, reasonably promptly, in public, for reasons entirely of their own.
Where the mechanism works well
In industries where a current profile is part of doing the job, it works very well. Sales, marketing, agencies, technology, consulting, anywhere the next role arrives through the network rather than through an internal posting. Keeping a profile current is career maintenance, so the signal fires early and often, and the vendor looks accurate because the population is doing the work.
Where it quietly fails
In sectors where people stay a long time, move up internally, and find the next role through a professional association or a word from a colleague, the profile is a formality. It gets created once, during a job search years ago, and is never touched again.
Nothing has gone wrong with the vendor. The input the refresh depends on simply never arrives.
| What drives the update | Where records stay current | Where records go quietly stale |
|---|---|---|
| How the next role is found | Through the network, so a current profile is the shop window | Through internal promotion, or a word from a professional association |
| Typical tenure in role | Two to four years | Five years and longer |
| Reason to maintain a profile | Career maintenance, done promptly | No practical reason, so often never |
| What a refresh job can see | A change event, usually within weeks | No event at all |
What that looks like in practice
Public sector and education buyers are the clearest worked example, because both halves of the problem are measurable.
- Turnover is measurable. ILO Group’s tracking of the 500 largest US school districts found 87 of them, or 17.4%, changed superintendents in 2025-26. That was down from a pandemic-era peak above 23%. A broader analysis covering more than 12,500 districts put the rate at 14.2% in 2019-20, rising to 17.1% by 2022-23.
- Profile upkeep is not. One analysis of senior public servants in Australia found that 65% had a professional profile, but 45% of those profiles were dormant and only 36% had posted anything original in six months. That study is small, single-country and published by a communications consultancy rather than peer-reviewed, so treat it as indicative rather than settled.
Put the two together. Roughly one senior buyer in six changes each year, and the event that would tell you has a meaningful chance of never being announced anywhere a vendor can see it.
The failure is silent
This is what makes decay expensive rather than merely annoying. A stale row looks exactly like a fresh one. Same fields, same formatting, no flag, no confidence score that moves. Nothing in the file distinguishes the record that is three weeks old from the one that has been wrong for two years.
So you do not discover it during a data review. You discover it at send time, from a bounce, or later and far more expensively, in a meeting with somebody who left the role last year.
Three symptoms that your average is hiding a split
- Your bounce rate looks fine but your meeting-held rate does not. Bounces catch dead mailboxes. They do not catch a live mailbox belonging to somebody who changed roles internally six months ago, so the damage surfaces further down the funnel and usually gets attributed to messaging.
- One territory keeps missing with a clean-looking list. When error rates vary by segment, and territories are drawn along segment lines, a data problem reads as a performance problem for a long time before anybody thinks to check the file.
- Re-verification keeps returning the same records. If the same rows fail cycle after cycle, the refresh is not failing at random. It is failing against a population the source cannot observe, and running it more often will not change that.
Prefer sources that update by obligation
The answer is not refreshing the same source more often. If the input never arrives, running the job weekly instead of quarterly changes nothing.
The answer is to anchor the segments that decay fastest to a source that updates because somebody is required to update it.
- Statutory registers: entity name, status and officers of record, maintained because filing is a legal obligation rather than a preference.
- Published board agendas and minutes: released on a meeting cycle, and they name the people actually in the room rather than the people on an org chart.
- Official staff directories: usually maintained by an administrator whose job includes maintaining them, so the trigger does not sit with the individual.
- Budget documents and awarded contracts: dated, attributable, and they tell you which department is holding the money this year.
None of these depend on an individual deciding it is time to tidy up a profile. They move on a schedule somebody else is accountable for, which is a far stronger guarantee than goodwill.
Martech Zone published a walkthrough of building an account list from source records that covers where those records live and how to reconcile them with each other.
Two caveats worth having in advance. These sources are slower and messier than a vendor export, and they are better at telling you a role exists and who now holds it than at handing you a working email address. So use them for the entity and the role, use enrichment for the contact detail, and let the obligated source be the thing that triggers a recheck rather than the thing that supplies every field.
How to audit your own list
You cannot manage decay you have not measured, and a single file-level accuracy figure will hide the one thing you need to see.
- Split the file the way you sell. By industry, buyer role or organization type, not by source or purchase date.
- Sample 50 records per segment. Smaller than feels rigorous, large enough to tell a 10% error rate from a 40% one.
- Verify against the primary source, not another database. Two vendors agreeing often means they share an upstream supplier, not that the record is right.
- Record error rate and record age separately. A segment can be old and accurate, or recent and wrong, and those two need different fixes.
- Set refresh frequency per segment. One cadence across everything over-serves the stable segments and under-serves the volatile ones.
Teams running this for the first time usually find the average was concealing a split. Several segments come in comfortably better than the headline number, and one or two come in far worse.
That is the practical point. Decay is not really a property of the vendor you bought from. It is a property of the people in the list, which makes it something you can predict, test for and plan around before it costs you a quarter.
Frequently Asked Questions (FAQs):
1. What is a realistic B2B contact data decay rate?
Published estimates mostly fall between 22% and 30% a year, with 22.5% the most commonly cited. Treat any single figure as a planning assumption rather than a measurement. If the number is going to drive a budget decision, measure your own file instead.
2. Why do some contact lists decay faster than others?
Because the correction mechanism depends on people updating their own professional profiles. Where that is a career habit, records get corrected quickly. Where tenures are long and promotion is internal, the update frequently never happens, so the record ages with nothing flagging it.
3. How often should I re-verify contact data?
Set the cadence per segment rather than across the whole file. A segment with high turnover and low profile upkeep may need verifying quarterly, while a stable segment can run annually without much loss.
4. Is it better to buy a new list or clean the one I have?
Cleaning is usually cheaper, and rebuying from the same source tends to reproduce the same gaps. If a segment is still poor after verification, the problem is the source for that segment rather than the age of the file.
5. A contact bounced. Does that mean they left the role?
Not reliably, and the reverse is the more expensive case. A bounce tells you the mailbox is gone, which is the easy failure to catch. The costly one is the mailbox that still works because the person moved to a different team at the same organization. The email delivers, the record looks healthy, and the opportunity goes to whoever replaced them.
