Most industrial companies have KPI frameworks. Very few have KPI frameworks that work.
The typical situation: a scorecard that has grown organically over five or more years, with metrics added whenever a new initiative launched or a new leader joined, and almost nothing ever retired. The result is a dashboard with 60 KPIs, three different definitions of “on-time delivery”, and no clear connection between what gets measured and what the organisation is actually trying to achieve.
The problem is not a lack of measurement capability. Industrial companies generate enormous amounts of operational data. The problem is design — or more precisely, the absence of it.
Building a KPI framework that actually works requires starting from a different place than most organisations start.
Step 1: Start With Strategy, Not With Metrics
The most common error in KPI framework design is beginning with the question “what should we measure?” The right question is “what are we trying to achieve?”
Before a single metric is defined, the organisation needs a clear articulation of its strategic objectives — the three to five things that, if accomplished, would constitute real strategic success over the next one to three years. These are not operational goals (“improve OEE”) or financial targets (“hit revenue plan”). They are breakthrough objectives: the changes in capability, position, or performance that define what the business is becoming.
Every KPI in the framework should trace back to one of these objectives. If a metric cannot be connected to a strategic objective, it is measuring something — but not something that matters for this framework.
This sounds obvious. It is surprisingly rarely done. Most KPI design starts from existing reports, existing systems, or existing team habits. Starting from strategy requires a conversation that many management teams find uncomfortable, because it forces explicit choices about what matters and what does not.
Step 2: Define Your Outcome KPIs
With strategic objectives defined, the next step is to identify the outcome KPIs for each one — the measures that will tell you, unambiguously, whether you are achieving the objective.
Outcome KPIs measure results, not effort. “Customer satisfaction score” is an outcome. “Number of customer calls logged” is an activity. “Yield rate” is an outcome. “Number of quality improvement projects completed” is an activity. “Revenue from new markets” is an outcome. “Number of sales pitches delivered” is an activity.
The distinction matters because activity metrics are almost always improvable — you can always do more things — while outcome metrics tell you whether the doing is working.
A well-designed framework at this stage typically has no more than two or three outcome KPIs per strategic objective. If you have five strategic objectives, you should have somewhere between ten and fifteen outcome KPIs in total. If you have more than that, you have almost certainly included activity metrics or duplicated outcomes under different names.
Step 3: Add Leading Indicators
Outcome KPIs tell you what has happened. Leading indicators give you early warning of what is likely to happen.
For each outcome KPI, identify one or two operational metrics that tend to move ahead of the outcome. For a yield rate outcome, the leading indicator might be first-pass inspection rate or raw material quality scores. For a delivery performance outcome, it might be production schedule adherence or supplier lead time variance.
Leading indicators are particularly valuable in industrial operations because outcome metrics often have a lag built into them — a change in process today may not show up in yield figures for two or three weeks. Leading indicators allow intervention before the outcome is affected.
Two important caveats. First, a leading indicator is only useful if the causal relationship between it and the outcome is understood and verified. Correlation is not enough — if you change the leading indicator, the outcome should follow. Second, leading indicators belong in operational team reviews, not in leadership strategy reviews. Their role is to enable early operational intervention, not to occupy leadership agenda time.
Step 4: Assign Owners and Define Thresholds
A KPI without an owner is an observation. A KPI with an owner is an accountability.
Every metric in the framework should have a single named owner — not a team, not a function, but a person. That person is accountable for performance against the target, for understanding the drivers of variance, and for presenting a credible response when performance is off track.
Alongside ownership, every KPI requires a threshold — the specific value at which performance moves from green to amber, and from amber to red. This is not a detail. RAG status that is calculated automatically against pre-agreed thresholds is a management tool. RAG status that is assigned by whoever is presenting the data is a communication tool — and a much less reliable one.
Setting thresholds requires a conversation about what acceptable variance looks like. What level of deviation from target is within normal operating range? What level signals a genuine problem requiring management intervention? These conversations are often avoided because they feel precise in an environment where precision feels uncomfortable. They should be had anyway.
Common Mistakes to Avoid
Measuring activities instead of outcomes. Covered above, but worth repeating: if your KPI framework is full of metrics that describe effort rather than results, you will always be able to show progress even when nothing is improving.
RAG without thresholds. Red/Amber/Green status that is manually assigned rather than threshold-calculated is a subjective signal, not an objective one. It creates variability in how performance is reported and reduces the reliability of the review as a management tool.
Too many KPIs. A KPI framework is not a data catalogue. Its purpose is to focus management attention on what matters most. More metrics dilute that focus. If your leadership scorecard has more than fifteen to twenty KPIs, it is almost certainly including things that should be handled at an operational level rather than a strategic one.
Defining KPIs without agreeing on data sources. A KPI is only as reliable as the data behind it. Before any metric is finalised, the calculation method and data source should be agreed and documented. This removes the “my numbers don’t match your numbers” problem that derails so many reviews.
What “Done” Looks Like
A well-built KPI framework is not a large document or a complex dashboard. It is a lean set of strategic KPIs — typically fifteen or fewer — each with a clear link to a strategic objective, a named owner, a defined target, an agreed threshold, and an automated data connection that means the number updates without anyone compiling it.
When the monthly review happens, the framework surfaces the metrics that are off track, with the right people accountable for them, discussing what is being done rather than debating what the number is.
That is a very different meeting from the one most organisations are running today. Getting there does not require a major transformation programme. It requires a deliberate design process — starting from strategy, working down to outcomes, adding leading indicators where they add genuine early-warning value, and committing to the discipline of ownership and thresholds.
The framework is not the hard part. The hard part is the organisational willingness to retire the metrics that do not make the cut.
The free KPI scorecard template gives you a practical canvas for working through the design process described here — from strategic objectives down to leading indicators and ownership assignments.