The Real Problem
Most companies spend weeks discussing the quality of production planning. S&OP reviews, MPS adjustments, consensus meetings. But few actually measure what happened on the shop floor after the plan was published.
Planning exists. Scheduling exists. But the factory does not execute as planned.
The reason is simple and uncomfortable: execution is not deterministic. It is probabilistic.
"The more detailed the plan, the lower the chance of executing it exactly as defined."
On the shop floor, reality asserts itself in predictable ways:
The operator does the easiest first, not the most urgent
The operator follows weight or volume targets, ignoring the mix
The operator changes the sequence to optimize setup
This is not indiscipline. It is the rational response of the shop floor to a plan that does not capture the real complexity of operations. Understanding this reality completely changes how a company should think about production planning and scheduling.
If there is no adherence measurement, planning loses relevance. The plan becomes a suggestion that no one needs to follow.
What Is Adherence
Adherence is the ability to execute what was planned.
It is the indicator that connects planning and execution. Without it, the plan is merely an intention and the factory operates on improvisation. With it, it is possible to identify where the system fails, why it fails, and how to improve.
There are two ways to measure adherence, and choosing the right one depends on the company's context, production type, and planning maturity level.
Build to Plan vs Build to Schedule
Before measuring adherence, you need to define adherence to what. The confusion between these two levels is one of the reasons many companies measure poorly, or simply give up measuring.
Build to Plan (BTP)
Plan Adherence
BTP = Mix
- Measures plan adherence per product
- Focus on product mix
- Simpler to measure
- Typical of MTS (Make to Stock) environments
- Associated with S&OP processes
Detail level
Aggregate: day, week or month
Horizon
Medium term (weeks to months)
Decision type
Volume and mix of finished goods
Accountability level
Management / Leadership
"Did we produce the correct volume and mix this week?"
Companies with structured S&OP and make-to-stock production primarily use Build to Plan.
"In this article, we will focus primarily on Build to Schedule, where execution becomes significantly more challenging."
How to Calculate
Build to Plan combines volume and mix. The table below lets you simulate the calculation with actual products.
Build to Plan Calculator
Change the production values to simulate different scenarios.
Formula
BTP = Mix
| Product | Planned | Produced | Mix % |
|---|---|---|---|
| Product A | 100 | 90% | |
| Product B | 200 | 100% | |
| Product C | 300 | 93.3% | |
| Total | 600 | 590 | 98.3% vol |
Mix
95%
Σ min(produced, planned) ÷ Σ planned
BTP
95%
= Mix
Mix: sum of min(produced, planned) for each product, divided by total planned. If a product was produced beyond planned, only the planned amount counts. Overproduction does not compensate for shortfall on another item.
Producing more than planned does not improve adherence. Neither volume nor mix exceeds 100%. Overproduction typically generates unnecessary inventory.
Build to Schedule adds the sequence dimension. With three multiplied factors, adherence drops faster and reveals where the main bottleneck is.
Build to Schedule Calculator
BTS multiplies three factors: planned vs produced quantity, mix percentage, and sequence percentage.
Formula
BTS = Volume × Mix × Sequence
Total quantity produced relative to planned (max 100%)
Percentage of produced items matching the planned mix
Percentage of orders executed in the defined sequence
Planned
Produced
Of total produced, how much matches the planned mix
Of executed orders, how many followed the planned sequence
Volume
92%
Mix
80%
Sequence
70%
Build to Schedule
52%
92% × 80% × 70% = 52%
Main adherence impact: Sequence at 70%. Improving this factor will have the greatest effect on the final result.
About sequence: if a single order at the beginning of the queue was not executed, it does not make sense to penalize all the others. The evaluation should be relative: skipping that order, was the rest of the sequence maintained? If so, the deviation is isolated, not systemic.
Tip: visualize indicators separately before consolidating. The multiplied indicator can mask which dimension is pulling adherence down.
Why Execution Fails
Operational variability
Operations are never stable. Each shift brings different conditions, and no plan can anticipate all the variations that occur on the shop floor.
Material unavailable
Delayed delivery, out-of-spec lot, insufficient quantity
Machine breakdown
Unplanned failure, corrective maintenance consuming scheduling hours
Performance loss
Below-standard speed, micro-stops, lower-than-expected yield
Operator absence
Absenteeism, shift change with different skills
Quality issues
Scrap above forecast, rework consuming capacity
Setup longer than expected
Underestimated changeover time, unavailable tooling
Variability is not the exception. It is the normal behavior of operations.
System complexity
Operations are dependent. Each step depends on the previous one. If one fails, the effect propagates to all subsequent operations. Uncertainty accumulates.
Execution is not deterministic. It is probabilistic. The more chained operations, the lower the chance of executing exactly as planned. Long sequencing is inherently fragile.
Probability Chain Simulator
Reliability by factor
Probability of full execution
42.1%
With a combined reliability of 84.09% per operation, full execution of 5 operations drops to 42.1%. Every factor below 100% contributes to cumulative degradation.
Time Degradation
Beyond the chain of operations, there is another dimension that erodes adherence: time. Tomorrow's plan has a reasonable chance of being executed. Next week's plan, less so. The plan two weeks out is, in practice, a projection.
Short term is executable. Medium term is uncertain. Long term is intention. Treating all three with the same level of accountability is costly in rework, rescheduling, and operational frustration.
Degradation Over Time
With 90% daily adherence, after 10 days the probability of maintaining the original plan drops to 34.87%
Planning limitations
Not every execution failure is the operation's fault. In many cases, the problem lies in the plan.
Infeasible sequence
The plan requires changeovers that don't respect real setup logic
Incorrect capacity
The plan assumes capacity the resource doesn't have
Unconsidered constraints
Tools, molds, certifications, or qualifications ignored
Inadequate lead times
Underestimated processing or movement times
Ignored setups
Changeover time between products is not in the model
Resource conflicts
Two orders scheduled for the same resource at the same time
When the plan is not executable, the error is not in execution. It is in scheduling.
Every Deviation Is an Error
If execution did not follow the plan, there was an error. There is no neutral deviation. Every difference between planned and executed represents a failure at some point in the system.
But this error can have completely different origins:
Operational failure
The operator did not follow the schedule. Lack of discipline, training, or motivation.
Unmet constraint
The material didn't arrive. The machine broke down. The tool wasn't available.
Local operational decision
The supervisor changed the sequence to avoid a larger stoppage. The decision was correct, but generated deviation.
Planning error
The plan assumed capacity that didn't exist, ignored constraints, or generated an impossible sequence.
"Adherence measures not only execution, but also the quality of the plan."
The error may be in scheduling
Many deviations occur because the plan was not executable from the start. Operations receive a schedule containing conflicts, overloads, or unrealistic premises. The inevitable result is deviation.
Improving adherence is not just about demanding more from operations. It involves:
Improve planning and sequencing rules
Review capacity and lead time parameters
Consider real constraints (tools, molds, qualifications)
Use more advanced tools (APS) to generate executable plans
Demanding execution without improving the plan does not increase adherence. It only increases conflict.
Adherence by Horizon
If adherence deteriorates over time, it cannot be treated as a uniform indicator. Accountability must vary according to the horizon:
Short term (0-24h)
Volume + Mix + Sequence
Medium term (24-48h)
Volume + Mix
Long term (48h+)
Volume only or out of scope
The longer the horizon, the greater the uncertainty. Demanding sequence adherence for a week ahead is unrealistic and counterproductive. Define below the rigor level that makes sense for each time band.
Define Rigor by Time Band
Current configuration: 0 to 24h: Volume + Mix + Sequence → 24 to 48h: Volume + Mix → 48h+: Volume only
Don't Measure Against Old Plans
Adherence should be measured against the last valid plan, not against the original plan.
Rescheduling happens. It is expected and healthy. The plan changes because conditions changed: a rush order came in, equipment became unavailable, material was delayed.
Measuring execution against a plan that has already been replaced creates distortion. The factory correctly executed the current plan, but the indicator shows failure because it is comparing with an obsolete version.
Wrong
Plan published Monday
Rescheduled on Wednesday
Execution per new plan
Measurement against Monday's plan → low adherence
Correct
Plan published Monday
Rescheduled on Wednesday
Execution per new plan
Measurement against Wednesday's plan → real adherence
Rule: always measure against the plan version that was in effect at the time of execution. If the plan changed three times, the reference is the third version.
Cause Capture
Measuring adherence is necessary. But measuring without understanding the reason for deviation is just generating numbers nobody knows what to do with.
Each MISS must have a registered cause. In the simulator below, classify each deviation and observe the pattern that emerges.
Root Cause Simulation
Mark each order as HIT (executed as planned) or MISS (deviation). For each MISS, classify the reason.
Learning Loop
Adherence is not control. It is learning. The complete cycle is:
Learning Loop
Define what to produce, when and in what sequence.
Adherence is not control. It is learning. The cycle repeats, and each iteration improves the quality of the next plan.
Each cycle iteration refines planning premises. Lead times become more realistic. Capacities reflect reality. Constraints are parameterized, not improvised.
The goal is not to achieve 100% adherence. It is to create a system that learns from each cycle and improves the quality of the next plan.
Plan → Execute → Measure → Classify → Correct → Plan better. This is the only path to planning that evolves.
Operational Contract
Most factories operate with an imposed plan. Planning generates the schedule and simply publishes it for execution. There is no room for operations to question, flag risks, or validate the feasibility of what was scheduled.
This creates a structural problem. When execution fails, there is always the same justification: "the plan was not executable." And it is frequently correct. A plan that has not undergone any operational validation is merely a hypothesis published as truth.
The alternative model is the operational contract. Before execution, the schedule goes through a review stage by the supervisor. They validate, flag risks, suggest adjustments. Planning incorporates the feedback and publishes the final version. Execution happens on a plan that both sides recognize as feasible.
The difference is profound: when the plan has been validated by operations, adherence stops measuring plan quality and starts measuring execution quality. Accountability shifts sides.
Simulation: Plan with Operational Validation vs Without Validation
Comparison of how execution governance directly impacts adherence.
Illustrative example based on real factory scenarios.
What changes in the process
Without operational validation
- ✕Plan does not consider real constraints
- ✕Risks are not identified
- ✕Operations need to adapt during execution
With operational validation
- ✓Constraints are anticipated
- ✓Risks are flagged before execution
- ✓Plan is adjusted before being published
Expected adherence
30% – 50%
Plan does not reflect actual factory conditions.
Expected adherence
70% – 85%
Plan adjusted based on operational reality.
"Without validation, adherence measures how unrealistic the plan was."
"The problem is not demanding adherence. It is demanding adherence to a plan nobody validated."
Governance and Accountability
For adherence to work as an indicator, it is necessary to clearly define who is accountable for each part:
Planner
Defines feasible schedule, incorporates supervisor feedback, updates premises
Supervisor
Reviews schedule, flags risks, validates or suggests adjustments before execution
Operations
Executes per validated plan, records deviations with root cause
Management
Analyzes adherence trends, defines structural corrective actions
Everyone should be measured by the same indicator. If the planner is not held accountable for plan executability, they will optimize for the model, not for reality. If the operator is not held accountable for adherence, they will optimize for comfort, not for the plan.
Adherence governance is not bureaucracy. It is clarity about who does what, and who is accountable for what.
Practical Insights and Takeaways
BTS consolidation must respect granularity and method
Calculate BTS at the most granular level possible (shift or day). For longer periods, do not use simple averages or consolidate percentages. Always accumulate volume, mix, and sequence and calculate the result on the totals — this ensures volume-weighted average.
No credit for overproduction
Producing above plan does not improve adherence. The indicator is capped at 100% — excess in one item does not compensate for shortage in another.
Volume does not compensate for mix or sequence error
High production with wrong items or out of sequence still represents low adherence. The result is multiplicative — any failure reduces final performance.
Sequence is the most sensitive execution factor
Even with correct volume and mix, small sequence breaks have great impact. Sequencing requires operational stability and executable plan.
Do not mix products for operational analysis
Consolidating BTS across different products or areas can hide real problems. Use aggregations only for management view — improvement happens in the detail.
Include all relevant flow
BTS must consider all finished products and, when applicable, intermediate items with inventory control. Ignoring parts of the flow distorts the analysis.
Adherence also measures plan quality
If the plan is not executable, the problem is not in operations. In these cases, it is necessary to evolve planning — either by adopting an APS for sequencing, or improving the rules and constraints of fine scheduling in the current system.
Use Pareto of causes to drive improvement
Adherence evolution depends on attacking root causes: breakdowns → maintenance; material shortage → planning and supply; scrap → quality; absenteeism → team management; communication → visual management and MES; long setups → SMED and better sequencing.
Adherence requires operations commitment
Before demanding execution, the plan must be validated by those who execute it. The factory manager needs to review and "sign" the schedule — transforming the plan into a real commitment, not just an imposition.
Measure adherence on a live plan, not frozen
BTS makes sense when applied to a constantly updated plan. If the goal is to measure execution against a frozen plan, use BTP — insisting on sequence in this context generates more noise than value.
"If the plan does not consider the probability of execution, it is not a plan. It is merely an optimistic intention."
The problem is not the factory not following the plan. The problem is measuring adherence without considering:
Probability
Each operation has a chance of failing
Horizon
Rigor should decrease with time distance
Context
Not every deviation is an error; some are correct decisions
Validation
Imposed plan does not generate commitment
Adherence is not about blaming the factory. It is about recognizing that planning without execution is theory, and execution without measurement is improvisation.