A head of recruitment is preparing for her executive committee. She wants to present the conversion rate from interview to hire, and her ATS gives her one, to one decimal place. That figure is computed only on the records whose final stage was entered. In her team, some recruiters stop at the « offer » stage and never move the candidate to « hired ». The rate on screen describes the data entry of a few people.
A dashboard always computes something. It does not say how many records it relies on, or how many it leaves out. That is why a recruitment audit starts with the data, before anyone comments on a result.
A recruitment audit reviews how a team hires: timelines, application sources, how stages run, the candidate experience, costs. It ends with findings and an action plan. The team can run it itself, or bring in a firm or a consultant.
Nearly all of those findings come out of the recruiting system. Average time to hire, the share of each source, the conversion rate per stage: each depends on a field someone filled in, or did not. The data audit is therefore the first part.
An ATS data audit is a review of the data in a recruiting system to establish what it can measure, what is missing and who has to fix it.
An audit that starts from the available data ends up commenting on what is easy to count. Start from the questions management and the team are asking, and keep five at most. For example: where the candidates we hire come from, at which stage we lose the most, how long a record waits for an answer, why we turn candidates down, which roles stay open with no movement.
Each question points to the fields it needs. The application source for the first, the stage history for the next two, the rejection reason for the fourth. That list of fields sets the scope of the audit.
For each field on the list, four checks are enough:
- Completeness. On how many records the field is filled in. A rejection reason present on a third of rejections supports no conclusion about why candidates are turned down.
- Consistency. Whether everyone writes the same information the same way. A source entered as « Referral », « ref » and « internal reco » gives three lines instead of one.
- Freshness. Whether the stage shown is the one the record is in. A hired candidate left at « interview » skews both timelines and conversions.
- Uniqueness. Whether a person exists only once. Two profiles for the same candidate double the applications and cut their history in two.
Run this check on the whole database when you can. A sample of fifty records gives an impression, but it misses the gaps between recruiters, between roles and between periods, which are the most telling.
These are the points that come up in most teams, with the metric each one makes wrong.
| What to check | What goes wrong when it is missing |
|---|---|
| A source on every application | Channel performance and the budget each one gets |
| A reason entered on every rejection | The analysis of why candidates are turned down |
| The final stage entered (hire, rejection, withdrawal) | Conversion rates and time to hire |
| A date on every stage change | Time per stage and records left waiting |
| A scorecard filed after the interview | Comparing candidates and following up with managers |
| A CV attached to the profile | Searching the talent pool |
| Location and contract type on every role | Comparisons between roles and between sites |
| A recruiter named on every open role | Workload per person and roles nobody follows |
| One profile per candidate | The number of applications and a person's history |
| Few tags, each one defined | Talent pool filters |
The data says what was entered. It does not say why a field stays empty. Half an hour with two or three recruiters and a manager is usually enough to find out.
The reasons are almost always practical. The field is badly placed on the screen, the list of reasons does not contain the right one, the final stage is validated in another tool, or nobody said what the information was for. Each of these reasons calls for a different fix, and none is solved by a reminder.
A useful audit ends with a short list. For each gap, write down the action, the person doing it and the date.
Separate two families. Quick fixes are done within the week: making a field mandatory, closing a list of values, merging tags. Projects take time: going back over a quarter of rejection reasons, deduplicating the database, training managers to file their scorecards. Start with the fixes that make a metric readable the following month. Our data entry rules describe these actions in detail.
Set the pace for what comes next as well. A short check every month, on the fields behind the questions from step 1 only, is enough to see whether completeness is improving.
Hirify connects to your ATS and measures how complete the fields behind your metrics are, on every record from the past twenty-four months. The Reliability report shows which metrics your data can compute, which are partial and which are waiting for data.
Every week, it lists the records to complete on open roles: those missing a rejection reason, a source, a scorecard after the interview or a CV. Each list opens the records concerned, with a direct link to your recruiting system. The action plan sorts these queues into quick wins and projects, and a recap goes out by email every Monday to the people you choose.
Consistent values and duplicates still have to be checked in your ATS, with the checklist above. Hirify changes nothing in your ATS while reading it. Fixes are still made by the team, in its own tool.
Four steps. First, write down the questions the team wants answered. Then measure how complete the ATS fields behind those questions are. Compare what is entered with what happens in practice. Finish by giving each gap an action, an owner and a deadline.
An HR audit covers the whole function: payroll, personnel administration, training, employee relations, hiring. A recruitment audit only looks at how the company attracts, assesses and hires. It can be one chapter of an HR audit, or run on its own when the difficulties are about hiring.
It is how well the data can support a measurement. It rests on four points: fields are filled in, the whole team fills them in the same way, they are up to date, and each record exists only once. GDPR compliance is a neighbouring but separate subject, about what you are allowed to keep.
No. An ATS stores what people enter. Some flag duplicates or let you make a field mandatory, but none fills in a forgotten rejection reason or a skipped stage after the fact. The fix goes through the team, which is why knowing where to look matters.
One full audit first, then a short check every month on the fields behind the metrics you follow. Three moments call for a full audit again: changing ATS, a new head of recruitment arriving, and two teams merging.
- A dashboard always computes a figure, even on half-filled fields.
- A recruitment audit starts with its data: check what it can measure before commenting on results.
- Start from the team's questions, which point to the fields to check.
- Four checks per field: completeness, consistency, freshness, uniqueness.
- An empty field almost always has a practical reason, which you learn by talking to the people who enter the data.
- Each gap gets an action, an owner and a date.