Most manufacturing dashboards suffer from abundance. Forty metrics, updated monthly, reviewed by people who look at three of them and ignore the rest. The problem isn’t measurement — it’s that nothing has been prioritised, so nothing gets acted on.
A Useful Set is Small. These Eight Cover the Territory.
Service
On-time in-full, measured against the date you originally promised rather than the revised one. This distinction is where most reported OTIF figures lose their meaning — if the date resets whenever you’re late, you can hit ninety-eight percent while consistently disappointing customers.
Order cycle time, from receipt to delivery, tracked as a distribution rather than an average. The average tells you little, the spread tells you how predictable you are, which is what customers actually experience.
Inventory
Inventory turns, split between raw material, work in progress, and finished goods. A single blended figure hides where the problem is, and the three respond to completely different interventions.
Excess and obsolete stock as a percentage of total, with an ageing profile. This one tends to be quietly avoided because the answer is uncomfortable, which is precisely why it belongs on the list.
Production
Schedule adherence — jobs completed in the planned period, in sequence. This is a leading indicator for almost everything else. Poor adherence precedes late deliveries, expedite costs, and inventory imbalance by several weeks.
Overall equipment effectiveness on the constraint only. OEE across all equipment is a number that improves without output improving. On the bottleneck it’s directly meaningful.
Supply
Supplier on-time in-full, by supplier, with a trend line. The trend matters more than the level, because deterioration is your earliest available warning of a supply problem.
Forecast accuracy, weighted by value, measured at your planning lead time. Include bias direction alongside the error figure — they lead to different actions.
How to Make Them Work
Trend, don’t snapshot. A single month’s figure invites explanation. Six months of direction invites action.
Pair opposing metrics. Inventory turns alone drives stock down until service breaks. Service alone drives stock up. Reviewed together, the trade-off stays visible and neither gets optimised at the other’s expense.
Give every metric an owner. Not a department — a person, who presents it and explains movement. Unowned metrics get reported and never acted on.
Segment where averages mislead. Aggregate OTIF across all customers hides that your three largest accounts are the ones being missed. Break out what matters.
What to Leave Out
Anything nobody can influence, anything measured because it’s easy rather than because it matters, and anything that hasn’t prompted a decision in the last year. Cutting a dashboard is usually more valuable than adding to it.
A reasonable test: for each metric, ask what you would do differently if it moved five percent in either direction. If nobody can answer, it’s decoration. Run that test annually and the dashboard stays useful instead of accumulating.
ticktick.ai produces these from live operational data with trend and segmentation built in, so review time goes on the movement rather than on assembling the numbers.
