Data Decay: How Candidate Records Go Stale and What to Do
Which candidate and client fields go out of date first, how to measure decay in your own database, and where refreshing pays off

Reopen a search from two years ago, and you'll notice data decay first. Work emails bounce, a phone number reaches someone who has never heard of the candidate, and profiles describe jobs people left long ago. Those records were accurate once, but the people behind them moved on.
Data decay is the gradual loss of accuracy in records as people and companies change jobs, move, and switch contact details. In recruiting, it affects the candidate and client records in your applicant tracking system (ATS) and customer relationship management (CRM) database. Some fields go stale much faster than others, so knowing which ones to check lets you refresh the records you rely on without re-verifying everything.
Key takeaways
- Work emails, pay expectations, phone numbers, and current employer go stale fastest, so confirm them before reaching out from an old record.
- A small random sample from one segment of your database shows how far that part of it has drifted.
- A "last confirmed" date on each record or key field makes the age of every detail visible at a glance.
- Refresh at the point of use, keep candidate-confirmed details over bought or scraped data, and archive records you can't confirm.
What is data decay?
In practical terms, data decay is the widening gap between what a record says and what is true today. The file stays the same while the candidate changes employer, city, or priorities.
That makes decay different from the other problems a recruiting database collects. A mistyped email address or a duplicate created during an import is wrong from the start, while a decayed record was accurate once and drifted. The difference matters when you clean up: entry errors call for better intake habits, and decay calls for regular confirmation. Both belong in a CRM hygiene routine.
Search results also mix in a second meaning from IT, where data degradation, sometimes called bit rot, means files corrupting on storage media. That's a hardware and backup concern, separate from the recruiting meaning covered here.
Why candidate records go out of date
A candidate record holds details that depend on parts of a person's life, and each of those changes on its own schedule. Employment is the clearest example, and the Bureau of Labor Statistics (BLS) measures it as tenure, meaning how long people have been with their current employer. BLS tenure data puts median tenure for wage and salary workers at 4.1 years in January 2026 and 3.0 years for ages 25 to 34.
Those figures measure time in the current job, not job changes per year, so they can't be converted into a decay rate for your database. They do show that an employer and title field can outlast the job it describes.
Location changes as well. The Census Bureau's American Community Survey found that 11.6 percent of people in the U.S. moved to a different residence in 2025. That figure covers the whole population, not just candidates, but it's a good reason to confirm an older city field before you match someone on location.
Contact details decay on their own timelines. A work email stops working the day someone leaves the company, and a mobile number can be disconnected and later reassigned to a stranger. Pay expectations, availability, and openness to relocation can change between conversations without leaving a trace in the record.
Client and hiring-manager contacts drift in the same way, which is the side of the problem that B2B data decay describes. Hiring managers change roles, and the person who approved your last search may have left. For agency recruiters, whose business depends on those relationships, keeping B2B contact data current is a separate job alongside the candidate database.
Which fields go stale first
Fields tied to a current employer go stale first, while skills and personal details change more slowly. The table below orders common candidate fields from most to least exposed, with what changes each one, the early warning sign, and how to refresh it. The order is a way to reason about risk, not a set of measured rates.
| Field | What changes it | Early warning sign | How to refresh |
|---|---|---|---|
| Work email | Leaving or changing employer | Hard bounce | Ask for a personal email |
| Pay expectations and availability | New job or new circumstances | Last discussed in an earlier search | Ask at each conversation |
| Mobile number | Disconnection or a new number | Wrong person answers or number unavailable | Confirm by email before calling |
| Employer and title | Job change or promotion | Resume or profile shows a newer role | Ask the candidate to confirm |
| Contact preferences and consent | Candidate asks to change or stop contact | Unsubscribe or no recorded consent | Record the request and respect it |
| Location | Moving home or relocating for work | New city mentioned in a reply | Ask before matching on location |
| Personal email | Switching email providers | Soft or hard bounce | Confirm through a second channel |
| Skills and certifications | New training or expired credentials | Certification past its renewal date | Ask for an updated resume |
Pay expectations and availability rank second even though they never bounce. They change quietly, and the first sign is often a candidate turning down a role pitched at the wrong level or start date. Skills sit near the bottom because people add to them more often than they lose them, with certifications that carry renewal dates as the exception.
When a record check turns up a new employer, treat it as a reason to correct the record, not as a sign the person is ready to move. Whether they want to hear about a role is something only a conversation with them can tell you.
How to measure data decay in your own database
A general decay figure says little about your own records, because databases differ in age and sourcing. A more useful approach is to pull a small random sample from one segment and try to confirm each record.
1. Choose a segment: pick one group, such as candidates for a specific role family or records untouched since a certain date.
2. Pull a random sample: select records at random rather than the ones you remember, since the records you know well may not represent the rest.
3. Check each record: look for bounces, wrong or disconnected numbers, and employer changes you can confirm with the candidate.
4. Record what you find: note which fields were wrong, so you learn which parts of the segment have gone stale.
The sample only needs to show a direction, such as whether a segment is mostly current or mostly stale, so keep it small enough to finish. A "last confirmed" date makes the next round much quicker. Added per record, or per key field such as email and phone, it separates a record updated last month from one imported years ago. Without it, both look equally trustworthy in search results.
Regular outreach produces measurement data too. If you track bounces and wrong numbers by segment alongside your other sourcing metrics, you can see where decay concentrates without running a separate project. Useful comparisons include active pipeline records against dormant ones, and early-career records against senior ones. The shorter median tenure for younger workers suggests their records may change sooner.
Consider an agency recruiter reopening a search similar to one filled two years earlier. A quick sample of the old shortlist shows several people have changed employers and some work emails bounce. So they confirm each person's details by personal email or phone before adding anyone to the new pipeline.
How to refresh the records that matter
Refreshing a record at the point of use, just before you contact or submit someone, puts the effort where you'll rely on the data. Re-verifying an entire database on a schedule takes far longer and spends effort on records you may never search again. A full refresh can still make sense at specific moments, such as before migrating to a new ATS or contacting a large dormant segment.
Point-of-use checks work well alongside a light cadence by segment. Confirm candidates in an active pipeline at every touch, run periodic checks on recent shortlists, and wait on dormant records until a search makes them relevant. An agency database may hold large dormant segments plus client contacts, while an in-house team's may hold years of past applicants. Grouping those records through talent pool management makes it clearer which ones deserve attention now.
Making that a habit within candidate relationship management keeps your most-used records current. For the rest, a few practices help:
- Treat enrichment as a lead: contact data enrichment can surface a newer email or title, so confirm it before it overwrites a field.
- Protect candidate-supplied fields: bulk updates through ATS data enrichment should fill gaps rather than overwrite details the candidate gave you.
- Merge duplicates: when one person has two records, at least one is probably out of date, so deduplication belongs in any refresh.
- Rank details by who supplied them: information the candidate confirmed reflects how they want to be contacted, so it outranks bought or scraped data when the two disagree.
- Archive what you can't confirm: a record you can't reach adds noise to every search, and your retention policy decides how long archived records stay.
Check old phone numbers before you call
A number in an old record may now belong to someone unrelated to the candidate. The Federal Communications Commission (FCC) runs a Reassigned Numbers Database that lets callers check whether a number has been permanently disconnected. The FCC's public notice says providers have reported disconnections to the database monthly since April 15, 2021. Checking an older number there, or confirming it by email first, helps you avoid calling a stranger.
Start from contact details candidates shared themselves
Book a demoStart with a sample before you refresh everything
Data decay is continuous, so no recruiting database stays fully current for long. You can control what you know: which segments have drifted and which records you can confirm before you rely on them. For your next search, pull a sample from the segment you plan to contact and confirm it before sending anything. What you find will show how much of that segment needs refreshing.
Use a shortlist built on opted-in candidate data
Book a demoFrequently asked questions
What is the difference between data rot and data degradation?
Both are IT terms for files becoming corrupted on storage media over time. Data rot is the informal name and data degradation the formal one. Neither refers to recruiting records going out of date.
What causes digital decay?
In the storage sense, digital decay comes from physical media wearing out and from file formats becoming unreadable. In a recruiting database, the cause is human: people change jobs, move, switch numbers, and update their preferences while the record stays the same.
Is it okay to contact someone from an old record?
Check what consent and contact preferences the record holds first, and respect any earlier request to stop contact. If the details are old, confirm them through a channel the person gave you. Open by explaining why you're reaching out rather than assuming they want to hear about roles.
About the author

Founder & CEO of Rotto, building tools that help tech recruiters source better candidates, faster.






