Case Study · Retail

Defacto HR Analytics: from dashboard to alert, from alert to decision model

A second period with Defacto after the DSO: consulting for Human Resources' own analytics team. Turnover, demographics, academy and HRBP dashboards; time-and-attendance and bonus automation; an alert setup connected to the company's internal assistant; and the Norm Staffing model at the backbone. Ongoing since December 2025.

2nd period
Period: 4
Layers: 2026 plan
Roadmap: Norm Staffing
Backbone

Context

Our first four years with Defacto were the founding of the Data Science Office and retail analytics. At the end of 2025 a second period began: this time with Human Resources’ own analytics team. HR had rich sources — payroll, time and attendance, academy, surveys, employee master data — but they lived in periodic presentations, could not be drilled to an actionable level, and could not answer “what is happening in my store this month” on the spot.

The approach is the one we learned at the DSO: first make the data visible, then turn it into alerts, and only then build a decision model. Each layer rests on the one before.

Visibility — the dashboards

Turnover. Joins and leaves; voluntary / involuntary split, tenure band, region, store and position breakdown; period and year-on-year comparison. Where attrition rose — which role, which tenure, which region — so the question reaches a level where action can be taken.

Demographics. The structure of the workforce: age, gender, tenure, education, position and location distribution, and how it changes over time. The ground for hiring and talent management decisions.

Academy and training. Training hours, attendance, completion and distribution by programme, together with the international dashboards. These were designed around the questions of the period they were built in; their metric definitions, breakdowns and layout are being restructured around the questions the department asks today.

Head-office HRBP. Bringing the visibility that stores have had for years to the head office: each HRBP sees, for the departments in their own portfolio, headcount, staffing structure, joins and leaves, open positions and absence on one screen. Completed.

All dashboards are fed from the same sources; a turnover figure is calculated by the same definition whichever dashboard shows it. Writing those definitions down — how headcount is counted, who is an active employee, which formula measures turnover — is the subject of the “global data dictionary” item on the roadmap.

Alerts and automation

A dashboard waits to be looked at; an alert arrives. Three pieces of work sit in the second layer:

Time-and-attendance alert system. Automatic detection of missing clock-ins, unmatched shifts, irregular overtime and leave inconsistencies in attendance records, delivered to the person responsible. The source of payroll errors is mostly record quality, and the error is usually noticed after payroll closes; the alert brings it forward.
Quarterly bonus automation. End-to-end automation of the calculate–approve–distribute chain based on performance data; a process that required many manual steps and data collected from several sources.
Assistant integration. A three-tier setup: alerts produced by the analytical systems delivered to the right person through the company’s internal assistant; natural-language Q&A — a store manager typing “how is absence in my store this month” and getting an answer; a weekly summary by role.

Alongside these, the maintenance of every dashboard, report and automation in production is a permanent item on the plan: data quality control, tracking changes in source systems, meeting user requests. If it is not written into the plan, new work swallows it.

Decision model — Norm Staffing

The backbone of the third layer is Norm Staffing: a model that calculates the staff each store × role should have from measured workload and an efficient peer in the same segment, and produces an increase / reduce / keep decision. Phase 1 has been presented and Phase 2 is under way; its continuations — per-country calibration for international stores and a cold-start staffing estimate for new stores — are on the roadmap.

Roadmap — 2026

Part of the engagement is building HR analytics’ annual plan together with the team: every project with impact, complexity, priority, integration need and status; current work and proposals kept apart. Highlights of the 2026 plan:

Area	Work	Status
Global HR performance	Trade data (revenue, gross margin, conversion, m² productivity, footfall) and HR data (headcount, hours, labour cost, turnover, absence) on one screen	New
Employee experience	Bringing NPS and recognition / feedback platform data onto dashboards	Awaiting data
International	Norm Staffing adaptation and a staffing–cost budget simulation	Proposal
AI	Exit interview and survey text split into topics with NLP; GenAI support for talent and performance management	Proposal
Infrastructure	Global data dictionary; hiring funnel and quality-of-hire dashboard	Proposal
Team	Application development training for the AI team within HR	Proposal

For every proposed item, “where will it be used, is the impact profitability or efficiency” was written into the plan; the decision is made together with the team.

Outcome

HR data is moving from periodic presentations to continuously monitored dashboards, from dashboards to alerts that reach the person responsible, and to a model that makes the staffing decision from data. The engagement continues; this page will be updated as it progresses.

How the system fits together

All our work