Skip to content
CogniYukti

Insight

An HR dashboard is trusted exactly until the first time it is confidently wrong.

Most people analytics oversells: six fields become an attrition explanation, a headcount chart becomes a workforce strategy. One bad inference in a board meeting and nobody opens it again.

What is there

Six views, computed as at a date you choose.

Headcount by department and by work location, as at any date rather than only today. Attrition across a window you set, as exits per month with a headline rate. Tenure distribution. Gender split. Leave utilisation per leave type for an accrual year.

Each is a question a founder or an HR head asks monthly, and each is answerable from the record rather than from a survey.

ViewWhat it answersWhere it stops
  • Headcount by departmentwhere people are, as at any datenot why a team grew
  • Headcount by locationthe same, geographically
  • Attritionexits per month, and a headline ratenot broken down by reason
  • Tenurehow long people have stayed
  • Gender splitthe composition as it standsnot a pay-gap analysis
  • Leave utilisationper leave type, per year
Six views, and the third column is the reason to trust the second. A dashboard that claims to explain attrition from six fields is one you stop believing the first time it is confidently wrong.
Illustrative
  • As at a date, not just now

    Headcount and tenure take a date, so a question about last March is a question rather than an archaeology project.

  • It reads the record, not a survey

    Everything here comes from the employment record itself — joins, exits, assignments, leave. Nothing depends on anybody filling in a form.

  • Gated separately

    Reporting access is its own permission, because who may read company-level composition is not the same question as who may open an employee.

What is deliberately not there

It measures attrition. It does not explain it.

Attrition is counted — exits per month, a rate against headcount at the start of the window. What it does not do is break that down by reason or exit type, and nothing analyses exit-interview notes.

That is a real limit and worth stating rather than implying. The exit type is recorded on the settlement and the interview notes are on the exit, so the raw material exists; the roll-up does not. If attrition-by-reason is what you need on a board pack, ask for it directly rather than assuming a dashboard called attrition contains it.

What a founder asks

Mostly about what they can safely quote.

How is the attrition rate calculated?

Exits inside the window against headcount at the start of it. It is a simple, stateable formula rather than an industry-standard annualised one, and you should know which you are quoting before it reaches a board pack.

Can we see why people are leaving?

Not as a chart. Exit type is recorded on the settlement, and interview notes are on the exit — both readable, neither rolled up.

Can we compare departments on attrition?

Headcount splits by department; attrition does not. That asymmetry is real and we would rather name it than let a demo imply otherwise.

Does the gender view do pay-gap analysis?

No. It reports composition. Pay equity is a different analysis with different obligations, and a composition chart is not it.

Who can see this?

Whoever you give reporting access to — a separate permission from opening an employee record.

Can we export it?

The views are on screen. If a board pack is the destination, tell us what it has to contain and we will show you what actually comes out rather than promising a format.

Book a demo

Run your own numbers.

The last one is the honest boundary of this page.

  • Pull headcount as at a date twelve months ago
  • Set the attrition window and check the formula against your own
  • Open leave utilisation for last year
  • Ask where attrition-by-reason would have to come from