How Much of a Recruitment CRM Is Out of Date? The Two-Year Figure

A candidate record is out of date when the person it describes no longer holds the job it says they hold. Across recruitment CRMs analysed by Illumini, 34% of candidate records show a changed employer or a changed job title over a two-year period.

That figure is the answer to a question most agency owners have asked and almost none have measured. It is also lower than the numbers circulating in recruitment software marketing, which tend to sit closer to two thirds, and higher than most consultants assume when they open a record from 2024 and dial the number on it.

Key findings

• 34% of candidate records show a changed employer or changed job title across a two-year window (Illumini)

• Illumini enriches between 50% and 70% of the records in a typical recruitment database

• The gap between 50% and 70% is driven by the quality of the identifying data the CRM already holds, not by the size of the database

• UK labour market movement data independently supports a decay rate in this range, which means an agency can sanity check the figure without taking a vendor's word for it

What does "out of date" mean in a candidate database?

Out of date is not one thing, and the choice of definition is where most published statistics on this subject fall apart. A record can be wrong in several separate ways. The employer can have changed. The job title can have changed while the employer stayed the same. The direct dial can have been reassigned. The work email can bounce. The mobile number can still work while everything around it is wrong.

The 34% figure counts one thing: a change of employer or a change of job title. It does not count contact detail decay, which runs on its own clock and is usually worse, because a work email dies the moment somebody leaves and a switchboard extension can be reassigned within weeks.

Counting only employer and title keeps the figure conservative and keeps it verifiable. Somebody's current employer is a matter of public record in a way their current mobile number is not.

Does the labour market data support a 34% two-year decay rate?

It does, roughly, and the arithmetic is worth running because it means the number does not rest on any single provider's claim.

The Office for National Statistics publishes labour market flows, including the volume of people moving directly from one job to another each quarter. Job-to-job moves have run at low single-digit percentages of total employment per quarter in recent years, and moves driven by resignation fell substantially from their 2021 peak as the market cooled. Compound a quarterly rate in that range across eight quarters and a fifth to a quarter of any given population will have changed employer.

Illumini's observed rate sits above that, which is what should be expected for two reasons. Candidate databases are not a random sample of the working population. They are weighted towards people who have already demonstrated a willingness to move, since that is how they entered the CRM. And the 34% includes internal promotions and title changes at the same employer, which do not appear in job-to-job flow data at all.

The ONS has said it plans to publish experimental labour market flows statistics built from payroll records later in 2026, <cite index="9-1">capturing movements in and out of payrolled employment including job-to-job flows between and within industries</cite>. That will make this kind of cross-check considerably easier for anyone who wants to run it themselves.

Why can only 50% to 70% of a database be enriched?

This is the number that matters operationally, and it is the one vendors are least willing to publish, because a range that starts at 50% looks worse in a slide than a single figure of 98%.

Enrichment works by matching a record held in the CRM to a current, publicly verifiable version of the same person. That match is only as good as the identifying information the record carries. A record with a full name, a former employer, a job title and a location matches reliably. A record with a first initial, a surname and an email address does not, and no amount of processing power fixes it.

The spread between 50% and 70% is almost entirely a function of how the database was built. Agencies that imported from a previous system, ran a bulk CV upload years ago, or allowed free-text entry without validation sit at the lower end. Agencies with consistent registration processes sit at the upper end.

There is a partial answer to this. Illumini parses the CVs attached to candidate records to extract identifying fields the CRM itself never captured. A record with almost nothing in its structured fields often has a CV sitting against it containing an employment history, a location and a qualification set. Reading the document recovers records that would otherwise be unmatchable, which is what moves an agency up the range rather than leaving it stuck at the bottom of it.

The remaining unmatched portion is worth being honest about. Some records describe people who have left the workforce, moved abroad, or maintain no public professional presence. Some are duplicates. Some were never accurate when they were entered.

How can an agency measure its own decay rate?

The exercise takes an afternoon and produces a number that carries far more weight internally than anything in this article, because it came from the agency's own database.

Pull a random sample of 100 candidate records last updated two years ago. Check where each of those people works now. The proportion whose employer or title has changed is the decay rate for that database.

Then run the second half, which almost nobody does. Take the job board, licence and advertising spend for the same two-year period, and estimate how much of it went on sourcing people who were already in the CRM. The first number is interesting. The second one is the one that gets a budget conversation started.

Methodology

Figures are drawn from Illumini's enrichment runs against recruitment CRM databases, covering perm, contract and temp agencies operating in the UK and internationally. Decay is measured as a change of employer or a change of job title observed across a two-year window on records held at the start of that window. Enrichment rate is measured as the proportion of records in a database successfully matched to a current verified profile. Contact detail accuracy is measured separately and is not included in the 34% figure.

Labour market context is drawn from Office for National Statistics labour market flows and PAYE Real Time Information statistics.