Research question and scope
This protocol measures how often a reworked order traces to a field that was missing at intake. The unit is one corrected order. A missing-field rework means the first pass could not be completed correctly because a required value was absent, blank, or contradictory at intake. The measure describes the intake design, not the diligence of the specialist.
Set the scope to one order desk, one client, and a fixed period. Record the intake form version and the required field list in force. If the form changes during the period, split the sample at the effective date. A different form creates a different denominator.
Keep the measure separate from a general error count. Some rework comes from a mistyped value or a wrong instruction, not a missing field. The purpose is to test whether the intake asks for what the work needs. Report desk errors separately.
Name the decision the result will support before collection starts. The client may use it to add a required field, change a label, or move a check to the source system. A report that only lists corrections leaves the form unchanged. Write the intended action beside the scope.
Record the order types included and the systems that hold the intake form. A phone order and a portal order may collect different fields. Note the channel and the form. The channel often explains why a field is missing.
Population and unit of analysis
Define the population as every order that was corrected after the first pass during the period. Include corrections found by review and corrections found by a customer. The unit is the order, not the correction event. An order corrected twice is one unit when the second correction fixes the same field.
Include orders from every intake channel and mark the channel. Exclude a change the customer requested after the first pass. Report the excluded count and the reason. A customer change is not a quality defect.
Exclude a correction that only reformats a value without changing its meaning. A date format change is not a missing field. Record the rule and an example. The sample should focus on work that had to be redone.
Decide how to treat an order with two missing fields. Count the order once and record both fields. The unit is the order, while the field list is the evidence. Report field frequency separately from the order rate.
Decide whether a cancelled order that was reworked before cancellation counts. Include it and mark the outcome. A cancelled order can still reveal an intake gap. The outcome is a separate column, not a reason to exclude.
Definitions that make the measure reproducible
Define a required field as a value the work instruction says must be present before the order can be completed. Define a missing field as blank, absent, or contradicted by another source. Define a missing-field rework as an order corrected because of that gap. Publish one example of each.
Define a desk error as a value that was present at intake and entered or applied incorrectly. Define a source error as a value that was wrong in the client's system before the desk touched it. Keep the two categories beside the missing-field category. The split decides who can fix the cause.
Define the intake version by the form or template in use on the order date. Record the version and the date. A field added after the order was placed cannot explain the earlier gap. The version keeps the comparison fair.
Define the correction time as the moment the order value changed in the source system. Record the first-pass time and the correction time. The gap shows whether the review caught the error before the customer did. The pair also shows whether the fix was made before or after the customer noticed.
Agree how to classify a field that was optional but should have been required. Keep it as a missing-field finding and mark the design issue. Record the current rule and the proposed rule. The category should point to the form, not to the person.
Sampling and data collection
Draw the sample from the order system's change history for the full period. Record the export date, the filters, and the row count. Do not build the sample from customer complaints. A complaint-based sample misses the corrections caught internally.
Capture the order, the channel, the intake version, the corrected field, the original value, the corrected value, the cause category, and the correction time. Add the customer-facing promise where one changed. Store the export with the report. The field and the correction time are the core evidence.
Check a small overlap sample with a second reviewer. If the reviewers disagree about whether a blank field caused the rework, the definition is too broad. Rewrite the rule and retest before the full sample. Agreement matters more than sample size.
Record the time spent matching the change history to the intake form. A report that requires opening every order will not be repeated. Note the fields the order system does not retain. The collection cost belongs beside the result.
Take a pilot of twenty corrected orders before the full draw. Use the pilot to test the cause list and the export. Record what was missing and how long each order took. A pilot that takes fifteen minutes per order is a warning about the full sample.
Measuring missing-field rework
Report the share of corrected orders with a missing-field cause, a desk-error cause, and a source-error cause. Show the counts behind each share. The three groups describe different failures and should not be collapsed. A single rework rate hides the form gap that the client can fix.
Break the result down by field, channel, and intake version. A missing delivery date may cluster in phone orders, while a missing tax identifier may cluster in a portal form. Compare only groups with the same form and a similar volume. A channel with a different form needs its own comparison.
Report the most common missing field and the share of orders that needed more than one correction. A repeated field is a design issue, while a repeat correction suggests the first fix was incomplete. Record both counts. The field list should be short enough to review in one sitting.
Report the median time from the first pass to the correction. A correction found in review is cheaper than one found after the customer complains. Include the customer-facing message time where a promise changed. The gap shows where the check should move.
Compare the missing-field rate for orders that passed a pre-submit check against orders that did not. A lower rate suggests the check works, while no difference suggests the field is being filled with a placeholder. Record the check and the result. The check is only useful if it blocks the order rather than warns after it.
Limitations and next action
The measure cannot show whether the customer would have supplied the missing value or whether the work could have proceeded with a default. It also cannot show a field that was missing but never caused a correction. Treat the result as a measure of corrected orders. Note the channels and order types that were not included.
Give the result to the operations lead and the client system owner with one change to test. The change might be a required field, a clearer label, or a pre-submit check. Record the effective date and repeat the measure with the same definitions. A controlled change is easier to judge.
Review the boundary cases before adding every field as required. A long form can increase abandonment or placeholder entries. Fix the one field with the most rework first. The goal is a complete order, not a longer form.
Repeat the measure after the change with the same channels and order types. Report the before and after counts together, including the sample size. Record any form or system change in the period. A comparison with a changed intake is not a comparison.
