Production KPIs: Building an Operational Dashboard

How to define the right production KPIs, build a useful operational dashboard, and connect field metrics to management decisions.

Why Production Dashboards Fail

The most expensive production dashboards often fail for the same reasons. Too many indicators: the operator does not know where to focus attention. Unreliable data: who trusts a figure that changes daily without reason? Dashboards disconnected from decisions: managers continue to direct based on intuition, not the dashboard. Common mistakes also include confusing design, slow updates, and no owner.

The solution is not to add more indicators, it is to remove some.

The Fundamental KPIs of Industrial Production

Everything starts with must-have indicators. They exist in all industries, even if their formula varies. Understanding these four pillars will keep you from reinventing the wheel.

OEE (Overall Equipment Effectiveness)

OEE is the king metric of production. Formula: OEE = Availability × Performance × Quality. Availability is the % of time the machine can produce (excluding planned maintenance). Performance is the ratio of actual speed / nominal speed. Quality is the ratio of good parts / parts produced.

An OEE of 85% is world-class. Industry average is 60–70%. If your OEE is below 50%, you are losing money with every production run.

Machine Availability

Availability is measured with MTBF (Mean Time Between Failures) and MTTR (Mean Time To Repair). The larger the MTBF, the longer between failures. The smaller the MTTR, the faster repairs happen. Improving availability comes down to two paths: predict failures to avoid them (see predictive maintenance), or repair them faster (tools, training, spare parts).

Link this indicator to real-time supervision for continuous visibility.

Quality Rate and Rejection Rate

PPM (parts per million) measures defects. A PPM of 1000 = 0.1% scrap. FPY (First Pass Yield) measures the % of good parts on first production, with no rework. An FPY of 95% means you produce scrap or rework on 5% of production.

Link this indicator to traceability to know exactly which batch, which machine, which hour produced the defect.

Throughput and Process Performance

The ratio of actual throughput / nominal throughput shows if your machine is running at its rated speed. Less than 100% = speed loss. Micro-stops (stops of a few seconds) that accumulate nibble away at throughput. A real-time dashboard lets you capture these losses invisible to the naked eye.

Advanced KPIs for Operational Management

Once fundamental KPIs are mastered, add advanced indicators suited to your context. Maintenance cost relative to production revenue: real profitability. TEEP (Total Effective Equipment Performance): like OEE but including planned maintenance stops. Compliance Rate: % compliance with production procedures.

From Data to KPI: Collection Architecture

A reliable KPI requires a chain of trust. Sensors measure, SCADA/IoT collect, data platform stores and calculates, dashboard displays. See Real-time industrial supervision and IIoT for SMEs.

For an implementation, check Azymuth monitoring solution and our IoT platform.

Designing an Effective Dashboard

Rules for a good dashboard design: 1 screen = 1 goal = 1 audience. One screen should serve only one decision. The operator needs one view, the team leader another, the director a third.

View hierarchy: summary at the top (traffic light), details below. Adapted refresh rates: a safety KPI updates in real time, a trend KPI updates hourly. Readable colors: avoid red/green for the colorblind.

Pitfalls to Avoid in KPI Construction

Pitfall 1: non-real-time-calculable metrics. You define a KPI based on data that arrives 24 hours late. The operator cannot use it to decide now.

Pitfall 2: ambiguous definitions between sites. Site A's OEE is not comparable to Site B's because calculation methods differ. Standardize before deploying.

Pitfall 3: non-actionable metrics. You display "production 5% behind schedule," but the operator does not know what to do to correct it.

Pitfall 4: dashboards with no owner. No one is responsible for updating definitions, verifying data quality, training new users.

Progressive Implementation

Implement in four phases. Phase 1: identify the 3 most impactful KPIs. Those that, if they improve by 10%, make a real financial difference. Phase 2: ensure source reliability. Audit sensors, validate calculation formulas. Phase 3: build views. Dashboards, alerts, reports. Phase 4: connect to management rituals. Managers must look at the dashboard daily and make decisions on it.

Tracking and Evolving KPIs Over Time

Your KPIs are not fixed. As your processes improve, you must re-evaluate thresholds and targets. A KPI that had a median of 70% last year should have a target of 75% this year. Establishing an annual or quarterly review cycle for indicators ensures they remain relevant and motivating.

KPI evolution also tells a story: steady improvement or plateau? Degradation following a team change? Visualizing trends over 12 or 24 months in management reports helps with strategic decision-making.

Conclusion: KPIs as Common Language

A well-built KPI dashboard becomes the common language between field and management. No more debate on "how much did we produce?": the data speaks. No more gut-feeling decisions: the number guides.

Discover our monitoring solution and services to implement your dashboard.

---

FAQ

What is a good OEE in manufacturing? OEE > 85% is considered world-class. The industry average is between 60 and 70%. If you are below 50%, there is a significant problem.

Should OEE be measured in real time or retroactively? Both. In real time so operators make decisions now. Retroactively (daily, monthly) for trend analysis and continuous improvement.

How to manage consistent KPIs across multiple production sites? Before deploying, define a shared metric dictionary and standardized calculation rules. Each site executes the same formula, even if local data differs.