Recruiter Performance: Measure the Team, Not the Person

Why most recruitment teams fly blind

Ask a sales manager about the team’s conversion rates and you get an exact answer. Ask a marketing manager about cost per lead and you get a dashboard. Ask a recruitment manager how the team is doing and you get… an opinion, and a spreadsheet somebody rebuilds every quarter.

That is the problem, and it is rarely a problem of willingness. The numbers exist. They are sitting in an applicant record, a job record, a placement record. What is missing is a way to ask a question of the whole set without either exporting a CSV or handing someone the keys to every candidate file in the business.

Those are usually presented as the only two options: fly blind, or watch everybody. There is a third, and it is the one worth building a reporting habit on. A manager can have the trend. That is not the same thing as having the records.

What a recruitment dashboard actually returns

A reporting dashboard in Simply returns aggregates. Counts, rates and trends over the objects in your own data model — jobs, applications, placements, and the object you added last month because your desk needed one the template never had.

The aggregation does two things and does them properly: bucket a measure over time, or break it down by a dimension. Between them, that covers most of what a partner actually asks on a Monday morning.

What that looks like in practice:

• Where the pipeline is stuck: applications by stage, and how long they have been sitting in the one they are in.
• What converts: stage-to-stage conversion per job, per client or per recruitment type, counted from the applications themselves rather than from a status field somebody remembers to update.
• Where placements come from: source, client and job type behind the hires, taken from the placement records.
• Workload per desk: open jobs, live applications and outstanding tasks per recruiter — without opening anybody’s records.
• Anything you added: a field you defined yourself is a dimension you can group by the day you define it.

And one thing it does not return: the rows behind the number. There is no drill-through from a chart to the records underneath it. That is the constraint the rest of this article is really about, because it is what decides whether reporting makes a team better or makes a team defensive.

Three levels, and the line between them

Level 1: The desk

At desk level you get aggregates about one person’s work. How many jobs are open on that desk, how many applications are live, how much is ageing, what the stage-to-stage conversion looks like.

That is a number about somebody’s work. It is not a file about somebody. The difference sounds like semantics until you try to act on it — and the system enforces it rather than trusting you to.

A recruiter carrying eighteen open jobs and converting badly is not a performance problem yet. It might be a capacity problem, a client-mix problem, or a job-quality problem. The number tells you to go and ask. It does not tell you the answer, and it does not hand you a list of the candidates involved.

Level 2: The team

At team level you see patterns. Which type of vacancy moves fastest through the pipeline, where the whole team loses candidates, whether one desk is carrying twice the load of the desk next to it.

That is where redistribution decisions and training decisions come from. If everybody stalls at the same stage, that is a process problem, not five coaching conversations.

Level 3: The business

At business level you connect the pipeline to commercial reality. Which clients produce placements and which produce work. Which recruitment type converts. Which sources are worth the money.

This is the level where recruitment stops being an operational function and starts being one the board can plan against.

The level that does not exist

There is no fourth level where you click a bar and get the candidates behind it. Permission is re-derived per request across four layers — workspace, object, record and field — and the check fails closed by default. If authority for a slice of data cannot be established, that slice is left out of the total instead of quietly included on the assumption that it is probably fine.

One consequence surprises people: two colleagues can open the same dashboard and legitimately see different totals, because they are counting different sets. That is the system working, not a bug, and the dashboard says what it is counting so nobody has to guess where the difference came from.

From data to action: how to use the numbers

Weekly team review

Schedule 30 minutes a week to look at the dashboard together. Not as a performance review. As a planning session.

• Walk the pipeline: what moved, what did not, what is ageing.
• Pick the one stage where the team loses the most and ask why out loud.
• Agree one thing to change for the coming week.
• Let people share what worked, in their own words. The dashboard cannot tell you that part.

One-to-one coaching

Use the number to start the conversation and the conversation to find the cause. The dashboard is not the evidence — it is the reason to go looking.

Instead of: ‘I feel like your conversations could be better.’

Say: ‘Nine of your applications have been sitting in screening for over two weeks. Walk me through what is holding them.’

If you need detail from an actual conversation, that comes from the conversation itself, not from a chart. Every transcript line is anchored to the moment in the audio it came from, so checking what a candidate actually said is a click rather than a re-listen — and it happens on the record, where permissions apply in full.

Setting goals

Set goals the dashboard can actually settle:

• ‘No application sits in screening longer than ten days.’
• ‘Interview-to-offer conversion moves from 22% to 30% this quarter.’
• ‘Every open job has at least three live candidates by the end of week two.’

Specific, countable, and derived from data you already generate by doing the work.

The impact on team culture

Reporting does more than improve planning. It changes what a team argues about.

From opinion to counted: ‘I think John is a good recruiter’ becomes ‘John’s desk carries the most open jobs and still converts above the team median.’ That is a fairer sentence, and a checkable one.

From private to shared: in many teams nobody knows what anybody else’s week looks like, so workload complaints turn into personality conflicts. A shared view of who is carrying what removes most of that in one meeting.

From surveillance to accountability: these are not the same thing, and the difference is architectural rather than cultural. Accountability is a number about the work. Surveillance is a list of the people. A reporting layer that cannot produce the second one is much easier to introduce than a promise that nobody will look.

Specific metrics that matter

Not everything countable is worth counting. These are the ones that tend to change a decision.

Pipeline

• Applications per stage, right now
• Ageing in stage — how long things have been sitting where they are
• Stage-to-stage conversion, per job, client or recruitment type
• Drop-off per stage: where you lose the most candidates

Desk

• Open jobs per recruiter
• Live applications per recruiter
• Outstanding tasks per recruiter

Outcome

• Placements by source, client and job type
• Time to hire, computed from stage transitions
• Anything you have chosen to record as a field — quality of hire at six months, for instance, is reportable the day you add the field and start filling it

Cost

• What the AI spent, per model and per action, in euros rather than tokens or credits

One honest caveat on time-based metrics. Anything computed from stage transitions only knows about transitions that actually happened in Simply. If you migrate history from an older system, you get the dates your export contained and nothing finer, so the first months of a trend line are exactly as good as your previous system was.

And a note on what is not in that list. Per-recruiter behavioural scoring — talk ratio, question depth, a rating of how somebody handled a candidate — is not something a dashboard here produces, and that is deliberate. Those numbers require reading individual conversations at scale, which is the drill-through problem wearing a different hat.

Resistance and how to overcome it

Not everybody cheers when reporting arrives. The common objections, and what actually answers them:

‘It feels like surveillance’

Reasonable reaction, and the honest answer is structural rather than reassuring words. The dashboard returns aggregates and cannot return the list behind them. There is no view where a manager reads through somebody’s candidates from a chart, because that path does not exist.

‘My conversations are private’

Recording a conversation is a separate decision from reporting on the pipeline, and it should stay separate. Consent, notice and retention belong to the conversation. A dashboard counting applications by stage does not read transcripts to do it.

‘Two of us see different numbers’

That one is real, and it is worth explaining before somebody discovers it in a meeting. Authority is re-derived per request, so a recruiter with team-level access and a partner with workspace-level access are counting different sets. Neither number is wrong.

‘Data doesn’t tell the whole story’

True. It is not meant to. Data narrows the question. The answer still comes from talking to the person.

Tip: start with the pipeline view, not the desk view. A team that first sees reporting as a way to argue about workload adopts it much faster than a team that first sees it as a scoreboard.

Dashboards for agencies vs. corporate teams

The questions differ more than the tooling does.

Agencies

At an agency the link to revenue is direct, so the useful reports are about throughput and client mix. Which clients convert. Where time-to-fill is going. Whether a desk is loaded or drowning.

That only works if the underlying data is actually there, which is the unglamorous prerequisite. When conversations, messages and documents propose the field changes and a human approves them, the record stays current as a by-product of the work — and a dashboard over stale records is worse than no dashboard, because it is confidently wrong.

Corporate

In corporate teams volume matters less than quality and predictability. The reports that earn their place are the ones a hiring manager can be shown: where their vacancy actually is, how long each stage took, what the funnel looked like.

If your reporting has to end up somewhere else — a BI tool, a warehouse, a board pack — the same objects are available over the public API under exactly the same permission rules. There is no separate export path that quietly bypasses them, which is the point.

From individual numbers to team performance

Desk-level numbers are useful, but the largest gains come from patterns you can only see across the team. Which office spends the most time on administration versus candidate contact. Which recruitment type has the healthiest conversion. Where the pipeline consistently narrows.

Trends over weeks and months are what make those visible. If average time in first-stage screening starts climbing, that could be volume, a hiring freeze at a client, or one person quietly underwater. Context decides the action. Without a trend line you notice none of it until it shows up in the placement numbers a quarter later.

The part people underestimate is that reporting exposes good practice as readily as it exposes problems. When one desk converts consistently above the rest, that is a signal to go and ask what they are doing — and the answer comes from asking, not from reading their candidates. Usually it is something small and copyable: a longer intake with the client, more follow-up on motivation, better timing on the salary conversation. The dashboard finds the desk worth asking. The team turns it into something everybody can use.