No items found.
blog
Schedule Adherence vs OEE: Which Metric Should Run the Shift?
By: Guidewheel
Updated:
11 min read
Use both, because they can move in opposite directions and that divergence is the most useful thing either one tells you. OEE measures how effectively an asset converted its available time into good output. Schedule adherence measures whether you made what you promised, in the quantity and order you promised it.
A plant can run one product beautifully all shift, post excellent OEE, and still miss three customer orders because it ran the wrong thing. The reverse also happens. Reading them together turns two arguments into one decision, and the trap to avoid is treating a high number in either as permission to stop looking.
Schedule adherence vs OEE at a glance
- OEE answers "did the machine convert its time well". It is asset-centric and says nothing about what you were supposed to make.
- Schedule adherence answers "did we make the right things, in the right quantity, on time". It is plan-centric.
- High OEE with poor adherence usually means long runs of the wrong product, often to protect the OEE number.
- Poor OEE with good adherence usually means a padded schedule that hides available capacity.
- Run both in one shift review, each mapped to a specific decision, and never average them together.
| Dimension | OEE | Schedule adherence |
|---|---|---|
| What it measures | Availability, performance and quality of an asset | Conformance of output to plan |
| Inputs it needs | Machine state, counts, ideal cycle time, reject counts | The plan from ERP or MES, plus actual completions |
| Whose behaviour it drives | Operators and maintenance | Planning and supervision |
| Typical gaming failure | Long runs of easy product | Padded plan that is always met |
| Decision it should trigger | Which loss to attack on this asset | Whether to resequence, expedite or escalate |
| When it misleads alone | You made the wrong thing efficiently | You met a plan that under-used the plant |
What each metric actually measures
They share almost no inputs, which is the root of the divergence.
OEE is availability multiplied by performance multiplied by quality. Availability is run time over planned production time. Performance is actual output against what the ideal cycle time says was possible in that run time. Quality is good units over total units. Every input is a property of the asset and the shift: machine state, counts, cycle time, rejects. Nothing in the calculation knows what you were supposed to make.
Schedule adherence compares what you produced against what the plan required, usually as a ratio of scheduled quantity delivered on time. Its inputs come from the plan in the ERP or MES, plus completions. Nothing in the calculation knows how efficiently the machine ran.
That is the whole asymmetry. OEE is asset-centric and plan-blind. Schedule adherence is plan-centric and asset-blind.
Two definitional traps worth avoiding before you start comparing them. First, schedule adherence measured against a plan that is revised during the shift measures nothing; freeze the plan at a defined point or you are grading against a moving target. Second, adherence measured only on quantity ignores sequence, and on a plant serving dated customer orders the sequence is frequently the thing that matters. Decide whether yours is a quantity metric or a quantity-and-sequence metric, and say which in the definition.
With the inputs separated, the divergence stops being mysterious.
How they diverge
Two worked cases, both common.
Case one: high OEE, missed schedule.
A line is scheduled to run four products across a shift: 400 of A, 300 of B, 200 of C, 100 of D. Each changeover costs roughly 40 minutes. The supervisor, measured on OEE, does the arithmetic and reaches an obvious conclusion: three changeovers consume two hours of a shift, and running one product all shift eliminates them.
So the line runs 1,100 of product A. Availability is excellent because there were no changeovers. Performance is excellent because a long stable run is where a machine performs best. Quality is excellent for the same reason. OEE comes in around 88%, the best figure that month.
Schedule adherence is 25%. Products B, C and D did not get made. Three customers are short, and the plant now needs an expedited changeover-heavy shift tomorrow to recover, which will post terrible OEE.
The metric did not fail. It measured exactly what it measures, and the behaviour it rewarded was locally rational and globally wrong.
Case two: poor OEE, perfect adherence.
A plant meets its schedule every day for a quarter. Adherence sits at 99%. OEE runs at 45%, and nobody is alarmed, because the orders are going out.
What that pair usually indicates is a padded schedule. The plan was built around the plant's demonstrated output rather than its capability, so hitting it requires no improvement and reveals no constraint. The plant has capacity it cannot see, and the adherence metric is actively concealing it. Guidewheel's downtime analysis, with staffing events averaging roughly 197 minutes and material or supply delays roughly 119 minutes, describes the kind of loss that hides comfortably inside a generous plan.
✅
The most dangerous pattern is when both metrics look acceptable but the plant is still underperforming. An adherence figure that never dips below 98% for a quarter usually signals a padded plan, while OEE that rises as adherence falls suggests someone is running long batches of the easy product. Both gaming patterns are visible only when you read the two metrics side by side — and both are more common than anyone admits.
Case three: both look fine, and the plant is still losing.
The subtler pattern. Adherence sits at 94% and OEE at 71%, and neither number is alarming enough to investigate. What the pair conceals is that the plant is meeting a schedule that was built around 71% OEE. The planner, reasonably, plans what the line has historically delivered. The line delivers it. Both metrics report success and the recoverable capacity never surfaces, because nothing in either number asks whether the plan was ambitious.
This is the case for reading the loss breakdown underneath OEE rather than the headline figure. A 71% OEE with 18% of available time in changeover is a different problem from a 71% OEE with 18% in unplanned stops, and the metric alone does not distinguish them. Neither shows up in adherence at all, because the plan already absorbed both.
The tell is an adherence figure that never moves. A metric that has not dipped in six months is usually measuring the plan rather than the plant.
All three cases are invisible to a single metric and obvious to the pair.
Which decision each one drives
Assign each metric to specific decisions, or both will be read as general performance scores and neither will change anything.
OEE should trigger:
- Which loss to attack on a given asset this week, by ranking availability against performance against quality.
- Whether a changeover improvement programme is worth funding, from the availability component.
- Whether an ideal cycle time is still realistic, from a performance factor that has drifted.
- Whether an asset is a genuine constraint or merely the noisiest.
Schedule adherence should trigger:
- Whether to resequence the remaining shift.
- Whether to expedite, and at what cost.
- Whether to escalate to planning, because the plan itself is wrong.
- Whether the plan is systematically padded, from adherence that never moves.
Neither should trigger an individual performance conversation on its own. Both are heavily influenced by decisions taken above the shift team, and using either as a scorecard is the fastest way to get the numbers gamed rather than improved.
The pairing rule that makes this work in practice: read adherence first, then OEE. Adherence tells you whether the shift did its job. OEE tells you how the time was spent getting there. Reading them in the other order invites the case-one failure, where a supervisor optimises the efficiency number and discovers the schedule consequence too late to fix it.
For choosing the unit these metrics apply to, see machine-level versus line-level OEE. For the neighbouring metric definitions this article deliberately does not restate, see OEE versus utilization versus capacity.
Running one shift review with both
One table, one order, ten minutes.
| Read | Question it answers | If it is bad |
|---|---|---|
| 1. Schedule adherence | Did we make what we promised? | Resequence, expedite, or escalate to planning |
| 2. Availability | Was the machine available to run? | Attack the largest downtime family |
| 3. Performance | Did it run at the expected rate while running? | Check the ideal cycle time before assuming a problem |
| 4. Quality | Did what it made conform? | Quality escalation, and check whether rework is being counted |
| 5. Top three losses by minutes | Where did the time actually go? | This is the improvement backlog |
Read adherence first. It is the outward-facing number and it frames everything after it.
Never average the two. A combined "operational score" blending adherence and OEE is a number that cannot be acted on, because a fall in it does not tell you which decision to take. Report them side by side, always.
Watch for the two gaming patterns explicitly. Adherence that never drops below 98% for a quarter suggests a padded plan. OEE that rises while adherence falls suggests someone is running long batches of the easy product. Both are visible only in the pair, and both are more common than anyone admits.
Benchmark context, carefully. Guidewheel's 2026 Factory Uptime Report puts U.S. factory uptime at 54.5%, with a 31-point gap between the typical machine and the top quartile. Useful for calibrating whether your availability is unusual. Not useful as a target, because the report's basis and yours are unlikely to match.
To build the shift view around both numbers, see designing an OEE dashboard, and for the formula itself the complete guide to OEE.
To set this up against your own plan data and machine data, talk to the Guidewheel team.
Frequently asked questions
What is the difference between schedule adherence and OEE?
OEE measures how effectively an asset converted its available time into good output. Schedule adherence measures whether you produced what the plan required, in the quantity and often the sequence it required. They share almost no inputs: OEE is asset-centric and plan-blind, adherence is plan-centric and asset-blind, which is why they can move in opposite directions.
Can OEE be high while the schedule is missed?
Routinely, and it is usually a sign the metric is driving the behaviour. A supervisor measured on OEE has an obvious incentive to run one product all shift and avoid changeovers, which produces excellent availability, performance and quality while three customer orders go unmade. The efficiency number was earned honestly and the plant still failed.
Which metric should a plant manager lead the shift review with?
Schedule adherence, then OEE. Adherence is the outward-facing question of whether the shift did its job, and it frames everything after it. OEE then explains how the time was spent getting there. Reading them in the other order invites the failure above, where the efficiency number is optimised and the schedule consequence surfaces too late.
Does OEE include whether we made the right product?
No. Nothing in the OEE calculation knows what you were supposed to make. Availability, performance and quality are all properties of the asset and the run, so a line producing entirely the wrong product efficiently will post an excellent OEE. That blind spot is precisely what schedule adherence exists to cover.
What inputs does schedule adherence need?
The plan, from your ERP or MES, plus actual completions. Two definitional decisions matter: freeze the plan at a defined point, because adherence measured against a plan revised mid-shift measures nothing, and decide whether yours is a quantity metric or a quantity-and-sequence metric. On a plant serving dated customer orders, sequence is frequently the part that counts.
About the author
