An operations director at a textile factory in the Vale do Ave region recently asked us for something simple: "I want to know why weaving line 3 produces 14% more pieces per shift than line 2, when the equipment is identical and the operators have similar experience." A week went by. The answer didn't come from the production manager, nor from the maintenance technician. It came from the HR manager, who cross-referenced data on attendance, time on site, breaks and operator rotation between lines. No BI tool, no production KPI dashboard had spotted it. The data was in the HR system, lying dormant.
The conclusion he drew is uncomfortable: a line's efficiency is not determined by the machine alone. It is determined by the people who operate it, by their condition, by their continuity, by the quality of their integration. And this means the operations director is making decisions about capacity and cost with half the information he needs.
The mistake we still make: separating production from people
Even today we still see many Portuguese industrial companies using two completely watertight data universes. On one side, the ERP: volumes produced, cycle times, scrap, machine efficiency. On the other, the HR system: attendance, payroll, annual performance appraisal. The operations director looks at one. The HR director looks at the other. They never cross paths.
Five years ago, we still believed this was normal — that each function had its own data and that was that. We were wrong.
A garment factory in Guimarães with 120 operators discovered, by cross-referencing HR data with production, that 67% of the variation in daily OEE didn't come from machine stoppages — it came from staff rotation, unrecorded breaks and absenteeism concentrated on certain shifts. The line supervisor already knew this intuitively. But when the number appeared on the screen, the conversation changed: it went from "the operators have to work harder" to "we need team stability and a predictable shift pattern".
This is not HR invading operations. It is operations recognising that its most important metric — capacity and efficiency — depends on variables sitting in a different silo.
Which HR data really matters to the operations director
Attendance and presence patterns. It's not "X% of general absenteeism". It's who is absent and when, with granularity by shift, line and period. An absence concentrated on a Tuesday at the end of the month is different from a random absence. A chronic absence of an operator with critically specialised machine skills is a capacity-planning problem the operations director needs to solve, not ignore.
Real time on site versus contracted hours. An operator clocked in as present for 8 hours, but who only works 6.5 (uncounted breaks, late starts, early finishes), affects theoretical capacity. If an HR ETL cross-references the time clock with production records, planning stops being fiction — it becomes forecast.
Rotation within the factory. Who works at which station, how often, and for how long. A new operator on a specialised line has 40% lower productivity in the first 3 months — this is predictable, but it has to be in the capacity model. If the HR system doesn't have this mapped, the production forecast is optimistic.
Qualification and versatility. How many operators can run machine X? How many can be reassigned to line Y if there's a stoppage? This is an HR data point that should be in the production ERP, and frequently isn't — or it is, but out of date. A factory that can't answer this question in 10 minutes is leaving money on the table.
Labour costs per line and product. A line with senior operators (3 people) has a different unit cost from one with junior operators (5 people). An operations director who wants to optimise resource allocation needs to see this in the model — it's not an HR curiosity, it's a production decision.
What changes when you cross-reference HR with operations
Capacity planning stops being an exercise in wishful thinking. A factory that knows it has, on average, 85 operators present on a winter Tuesday (not 95), that 12 of them are qualified for the automatic cutting machine, and that 3 have scheduled holiday in March, can produce a realistic production forecast. Without this, it promises the customer 10,000 pieces and delivers 8,500 — and then blames the machine.
You can identify people bottlenecks before they become production bottlenecks. If turnover at a critical station is 35% a year, you know you'll continuously lose expertise. You can decide: do I invest in retention? Automate? Change the process? But the decision is informed, not reactive.
And third — this is what shocks operations directors most when they first see it — you can correlate absenteeism, health, psychological stress, and productivity. Not because you're being paternalistic. Because it's operationally more efficient. Stress and workplace psychological health problems cost companies in Portugal up to €5.3 billion a year (around 1.4% of turnover), adding together absenteeism and presenteeism, according to the Ordem dos Psicólogos Portugueses. An operator who is present but unwell costs as much as one who is absent — and is harder to see.
What the HR system has to do for this to work
The pplPortal — the HR system we support — has a layer that many people don't use: the integration of production data with attendance and individual performance data. It's not a pretty "BI module". It's a structured database that allows simple questions: "What is the correlation between real time on site and productivity on this line?" or "Who are the 5 operators that most affect OEE variance?".
This requires three things from the HR system. First: granular time recording — not just "present/absent", but time in, breaks, time out, by shift, by day. Second: integration with production data — the HR system has to receive, via API or ETL, the volumes produced, cycle times, and quality, per operator or per station. Third: the ability to answer ad-hoc questions — don't expect the HR manager to have prepared the dashboard the operations director wants. The system has to allow quick cross-referencing without code.
If the HR system doesn't do this, it's a payroll system. It's not a people analytics system.
Why this isn't already everywhere
Because it requires the operations director and the HR manager to work together. And that is rare. Often, the HR manager sees this as an invasion of privacy — "I don't want the operations director to know when the operators step out for a smoke". The operations director sees it as a waste of time — "what I want to know is whether the line works; nobody's personal life interests me". Neither of them is wrong, but they're speaking different languages.
The solution is not philosophy. It's making clear that the goal is operational, not punitive. The operations director doesn't want to know names — he wants to know patterns. "Line 3 has 2.3 hours of uncounted breaks per shift" is useful information. "João left 4 times on Tuesday" is noise.
Second: many HR systems aren't integrated with the production ERP. The information stays segregated. Cross-referencing data manually is possible, but it's a monthly Excel exercise nobody wants to do — and one that's out of date the next day.
Third: there's still a belief that this is "advanced BI" and requires expensive consultants. That's not true. It's simple data engineering — but it has to be done well, and early, in the design of the system. If you leave it for later, it becomes a patch nobody can maintain.
A question for the next meeting
Ask your HR manager: "If I wanted to know tomorrow what the correlation is between absenteeism and OEE in this factory, how long would it take?" If the answer is "a week", your system doesn't have people analytics. If the answer is "I don't know if it's possible", your system is an archive, not a decision-making tool. And if the answer is "15 minutes", then your operations director is already winning — and probably already knows it.
Frequently asked questions
Why is HR data important to the operations director?
Because the efficiency of a production line depends not only on the equipment, but also on the people who operate it. Data on attendance, real time on site, operator rotation and qualifications directly affect capacity and OEE. An operations director who ignores this data makes decisions with half the necessary information.
What is the most common data mistake in Portuguese factories?
Completely separating production data (in the ERP) from HR data (in the human resources system). The operations director looks at one system, the HR director looks at another, and they never cross paths. This prevents the true causes of variation in productivity from being identified.
What does "real time on site" versus contracted hours mean?
An operator clocked in as present for 8 hours, but who only works 6.5 hours (uncounted breaks, late starts, early finishes), affects the factory's theoretical capacity. Production planning based on contractual hours, without considering real time on site, becomes fiction, not forecast.
How does operator rotation affect capacity planning?
A new operator on a specialised line has roughly 40% lower productivity in the first 3 months. If the HR system doesn't map who works at which station and for how long, the capacity model becomes optimistic and production forecasts fail.
What qualification information is critical for operations?
How many operators can run each machine, how many are versatile and can be reassigned in the event of a stoppage. This data should be in the production ERP, but it's frequently out of date or missing. A factory that can't answer this in 10 minutes loses optimisation opportunities.
How do you identify a people bottleneck before it becomes a production bottleneck?
By monitoring the turnover rate at critical stations. If a station has 35% annual turnover, you know you'll continuously lose expertise. You can then decide, informed: invest in retention, automate, or change the process.
What is the impact of absenteeism and psychological stress on production?
An operator who is present but unwell (stressed or with psychological health problems) costs, operationally, as much as one who is absent, but is harder to identify. Correlating absenteeism with productivity enables more efficient decisions based on data, not intuition.