The most common failure of quality measurement is not measuring wrongly but measuring and not using the result. Findings are recorded, reports are produced and no decision changes. The value of the data lies not in the score of a single job but in the trend it shows over months.
Scope of the data collected
The type of the finding, its severity, its job and its resource are recorded. The fields are defined. The record is structured.
Structured record
Free-text notes cannot be analysed. The categories are kept fixed. The data becomes processable.
The burden of unused data
Data that is collected and not looked at only consumes effort. Measurement becomes meaningless. Use is mandatory.
Trend analysis
A single measurement gives no information. A series is built. Change is monitored.
Distribution of error types
Which type is concentrated becomes visible. The concentrated type is the target for measures. The decision comes from there.
Interpreting the concentrated type
If terminology findings are dense, preparation is lacking. If meaning findings are dense, matching is the problem. The interpretation gives direction.
Analysis by resource
Which type appears with which resource is monitored. Matching decisions rest on this. The data is interpreted with care.
The interpretation trap
Difficult texts naturally produce more findings. The text, not the resource, may be responsible. Comparison is made with care.
Analysis by client
Which finding appears with which client becomes visible. Gaps in preparation emerge. A suggestion is produced.
Analysis by document type
Some document types systematically produce more findings. The process is adapted to that type. The measure becomes targeted.
Turning into a prevention decision
The data shows which assurance step is needed. The decision has a basis. Investment goes to the right place.
Turning into a training plan
Common findings become training subjects. The plan rests on data. The effect is measured.
Turning into a checklist
Recurring findings are added to the checklist as items. Knowledge passes into the tool. Practice becomes lasting.
Turning into tool settings
Automated checks are configured according to frequent findings. Warnings become targeted. Efficiency rises.
Sharing with the client
Trend data is shown in the report. A common view is formed. The decision is made jointly.
Suggestions to the client side
The data also shows gaps that originate with the client. The suggestion is made constructively. The contribution runs both ways.
Storing the data
Records are stored for comparison between periods. Deleted data means a lost trend. An archive is kept.
Stability of the measurement scheme
If the typology and severity change, comparison with the past breaks down. The scheme is kept fixed. Changes are recorded.
Reporting frequency
The analysis is carried out periodically. Analysis on every job creates a burden. A reasonable interval is chosen.
The loop that ties data to decisions
Measurement, analysis, measure and re-measurement form a loop; when this loop does not close, measurement turns into a reporting ritual and nothing improves. Closing the loop is the only meaningful output of measurement. That is why the structure is built.
Limits of the approach
Data shows the past and is no assurance of the future. Conditions may change. The limit is known.