*The cost of poor quality can begin long before an error appears in performance indicators. And it often begins where nobody is measuring.*
Imagine a commercial negotiation that has been successfully concluded.
The proposal has been approved, the contract formalized, and the information forwarded to the teams responsible for execution and billing.
Everything appears to be working.
Then a discrepancy is found: a contractual condition does not match the customer record, a billing period is incorrect, or required information was not recorded.
The issue needs clarification. Documents are reviewed. People are contacted. Records are corrected. In some cases, work already completed must be done again.
The discrepancy is finally resolved.
But how much did it cost?
And, above all, why did so many steps have to be completed before anyone noticed the error?
1. The cost begins before detection
When incorrect information enters a process, it may be used repeatedly by different departments.
Every activity performed on the basis of that information increases the potential for the error to spread.
An incorrectly recorded commercial term can influence contract drafting, system configuration, operational execution, and invoice issuance.
When the problem is detected only at the end, correcting it may involve several teams.
The cost, therefore, is not limited to the activity in which the discrepancy was discovered.
It may include:
- Time spent investigating and correcting the issue;
- Administrative and operational rework;
- Reprocessing documents and transactions;
- Delays in delivery or billing;
- Additional communication between departments;
- Damage to the customer relationship;
- Potential financial and contractual consequences.
Not all these effects occur in every situation. Ignoring them, however, can produce an incomplete picture of the cost of poor quality.
The point where an error is detected is not necessarily where it originated.
This distinction is essential to continuous improvement.
2. When an indicator reveals the symptom but not the origin
Consider a billing department that tracks the number of documents corrected each month.
The indicator matters. It shows failures detected at that stage.
On its own, however, it cannot answer essential questions:
Where was the discrepancy introduced?
Which information was incorrect?
At what stage could it have been detected?
How many activities were performed before correction?
How much time and which resources were consumed?
Without these answers, an organization risks treating the department that detects the problem as though it caused it.
This can lead to additional checks at the end of the process, increasing inspection effort without reducing the creation of errors.
Effective improvement requires understanding the flow of information from its source.
3. The first tool is a clear definition of the defect
Before building indicators, we must define exactly what counts as a discrepancy.
For example:
*A master-data discrepancy is any inconsistency between mandatory information recorded in a commercial proposal and the corresponding information in the database used to execute or bill the contract.*
This is an illustrative operational definition. Each organization must adapt it to its own process.
A clear definition makes it possible to establish consistent identification and counting criteria.
It also helps distinguish different situations:
- Missing mandatory information;
- Information that is present but incorrect;
- Discrepancies between documents or systems;
- Outdated information;
- Requirements that were misunderstood or interpreted differently by departments.
Without this standardization, two people may examine the same record and reach different conclusions.
In that situation, the indicator loses reliability before it is even put to use.
4. Counting occurrences is not enough: we need to stratify
An organization may record 120 discrepancies in a given period.
That figure alone says little.
Analysis becomes more valuable when occurrences are classified by relevant characteristics, such as:
Type of information: commercial terms, billing period, customer identification, contract value, or mandatory documentation.
Likely stage of origin: initial registration, proposal preparation, contract formalization, or system updates.
Detection point: administrative review, operations, billing, or customer complaint.
Team responsible for correction: the department or role involved in resolving the issue, without turning measurement into a tool for assigning individual blame.
Impact: correction time, number of activities repeated, delays generated, and potential financial effects.
Stratification helps identify patterns that would remain invisible in an aggregate indicator.
A useful tool at this stage is the Pareto chart, which organizes occurrence categories by frequency or impact.
However, the most frequent category is not necessarily the most costly.
A simple discrepancy corrected in a few minutes may occur many times. Another, less frequent issue may disrupt billing, mobilize several teams, and require hours of investigation.
Frequency and economic impact should therefore be examined together.
5. Turning rework into an economic measure
The cost of poor quality can be estimated from the resources consumed in correcting failures.
An initial approach is to measure:
Direct rework cost = number of occurrences × average correction time × hourly cost of the resources involved.
When different roles take part in correction, the calculation should reflect their respective time and cost.
### Illustrative example
Suppose an operation has 120 discrepancies per month.
Each occurrence requires, on average:
- 20 minutes from an administrative employee;
- 15 minutes from a commercial employee;
- 10 minutes from a finance employee.
The total effort is 45 minutes per occurrence.
Over the month, this amounts to:
120 × 45 minutes = 90 hours of work.
If, solely to simplify the example, we use an average cost of BRL 60 per hour, the result is:
90 hours × BRL 60 = BRL 5,400 per month.
Over twelve months, assuming the same conditions, that would represent BRL 64,800 in estimated rework costs.
This amount is not an automatically recoverable saving. It represents the estimated cost of resources consumed by corrections.
Nor does it necessarily include billing delays, penalties, customer impacts, or other indirect effects.
To estimate actual gains, it would be necessary to demonstrate fewer occurrences, measure the resources released, and verify whether expenses were genuinely reduced or productive capacity increased.
Even so, the exercise reveals something important:
An apparently minor administrative error can consume substantial operational capacity when repeated.
6. Preventing errors at the source is different from inspecting at the end
A common reaction to rising discrepancies is to add another review step.
In some circumstances, extra inspection may be needed to contain immediate risks.
But it should not replace investigation of the cause.
If mandatory information is frequently left unrecorded, for example, several questions need answers:
Does the field exist on the original form?
Is completing it mandatory?
Is there a clear standard for recording the information?
Did the people who use the data later help define that standard?
Does the system allow incompatible entries?
Is there validation before the information moves to the next stage?
These questions guide preventive action.
Options include standardizing forms, reviewing input requirements, integrating databases, training teams, and introducing mechanisms that prevent or flag inconsistencies.
In continuous improvement, this approach aligns with quality at the source: creating conditions for information to be recorded correctly and errors to be detected as close as possible to their origin.
Where technically feasible, *Poka-Yoke* controls can help prevent incompatible entries, omissions, or incorrect sequences.
The goal is not to keep increasing the number of checks indefinitely.
It is to reduce the need to correct work that could have been done correctly from the beginning.
7. A smarter set of indicators
To track process improvement, we can combine outcome indicators with preventive indicators.
- Discrepancy rate: Share of records with inconsistencies among all records examined
- Discrepancies by category: Main types of defects
- Detection at the source: Percentage of discrepancies identified before moving further through the process
- Average correction time: Effort required to resolve occurrences
- Estimated rework cost: Resources consumed by corrections
- Recurrence: Reappearance of failures after corrective actions
It is important to define denominators, measurement periods, classification criteria, and data-collection responsibilities.
We must also avoid hasty interpretations.
An initial increase in discrepancies detected at the source may signal improved detection capability rather than process deterioration.
Likewise, a decrease in recorded occurrences does not prove improvement if the organization has reduced its ability to detect problems.
A good indicator system must distinguish defect generation, detection capability, and the consequences of failures.
8. From accurate information to profitability
Information quality may appear to be an exclusively administrative concern.
In practice, it affects productivity and process reliability.
When an organization reduces recurring discrepancies, it can reduce correction effort, free up working capacity, avoid delays, and improve delivery predictability.
These outcomes can contribute to profitability.
But there is an important condition: the benefit must be demonstrated.
Hours of rework avoided do not automatically become cash in the bank. Their economic effect depends on how released capacity is used and which expenses or losses are actually reduced.
Improvement initiatives should therefore connect operational indicators to verifiable economic consequences.
That connection strengthens decision-making and prevents projects from being assessed only by the number of actions implemented.
9. See the error. Understand how it spreads. Choose where to act. Evidence the result.
In Power 4P's approach, this challenge involves different stages of the 6Es.
SEE: identify where discrepancies appear, which information is involved, and how the process is structured.
UNDERSTAND: investigate where inconsistencies originate, how they spread, and the factors behind their recurrence.
CHOOSE: prioritize interventions according to frequency, impact, risk, and feasibility.
EVIDENCE: track indicators to show whether changes reduced defects, rework, and economic impacts.
Implementing actions and evolving standards are also part of the full improvement journey.
The essential point is to understand that a discrepancy should not be analyzed merely as an isolated event.
It may indicate that the process allows inadequate information to advance without sufficient controls.
Conclusion: the most expensive error may be the one nobody notices immediately
When incorrect information reaches the end of a process, the organization may already have invested time, resources, and decisions on an inadequate basis.
Correcting the record is necessary.
But understanding why the error was created, how it spread, and why it was not detected earlier is what turns the occurrence into an opportunity for improvement.
The question, therefore, should not be only:
“How long did it take us to correct the error?”
It must also be:
“How much work could have been avoided if the information had been correct at the source?”
This is where information quality stops being merely an administrative requirement and becomes part of managing productivity, processes, and profitability.
---
Power 4P | Organizational Management and Human Development
*4P: Processes • Profitability*
*6Es: See • Understand • Choose • Execute • Evidence • Evolve*