Consulting & Pricing

Evaluate Workflow Training: Ask the Learner to Perform the Work

The attendance sheet proves someone attended. The draft record shows what they did with the instructions.

Keep both records. Give each a different job. When the question concerns performance, ask the learner to perform a bounded task under stated conditions. Then inspect the result and the help required to produce it.

CDC's training-evaluation guidance distinguishes learning from applying that learning at work. It supports demonstrations and observations as assessment methods. The supplier-intake exercise below adapts that guidance to a fictional business workflow. Its cases, rubric, and review schedule are proposed methods, not a CDC business certification or endorsement.

Decide Which Question Each Record Answers

Attendance answers whether the learner participated in the session. Satisfaction answers how the learner reports experiencing it. Keep feedback about confusing explanations, inaccessible materials, or missing examples. Use it to revise the lesson.

A knowledge check asks the learner to identify or explain a rule. Observed performance asks them to apply the rule while doing a task. Later workplace evidence asks whether they apply it under working conditions.

For this evaluation, record those five types of evidence separately:

Evidence Question for this workflow Limit to record
Attendance Did the coordinator participate in the practice session? Participation alone leaves task performance untested
Satisfaction Which explanation or sample did the coordinator find unclear? Reported experience does not show a completed intake
Knowledge check Can the coordinator identify the incomplete-packet rule? Selecting the rule leaves its application unobserved
Observed performance Can the coordinator prepare and route a synthetic draft? The result applies to the observed cases and conditions
Workplace application Does a later eligible intake show the taught checks? Access, opportunity, task version, and support affect the interpretation

Do not combine these records into an unexplained readiness score. A person may appreciate the session and still need practice. A correct sample may coexist with an inaccessible production screen.

Write the evaluation question first: can the coordinator prepare a reviewable draft and stop at the stated boundary? That question determines the exercise. It does not ask whether the consultant delivered every contractual item or whether the learner holds a professional qualification.

Define the Task and Allowed Support

Use a fictional supplier-intake workflow. The coordinator checks the packet, creates a draft, and routes it for review. A separate approver controls supplier approval. No purchasing or payment changes occur during the exercise.

Prepare an isolated practice space. Use synthetic identifiers, invented organization names, and dummy documents. Give the space no ability to submit to real suppliers, approve live records, or send notifications outside the exercise.

Confirm those boundaries before the learner begins. If the practice environment cannot prevent a real action, use a paper or local mockup instead. Do not turn a training evaluation into an accidental operating change.

State the allowed support. Here, the coordinator may use checklist version three and the help text available beside each field. The trainer may clarify the task wording but must record any hint about the correct decision.

CDC's learning-objectives guidance includes observable behavior and the conditions of performance. Apply that distinction to the assessment. If the objective allows a checklist, do not remove it during observation and judge the result against a memorization requirement.

The task ends with a saved practice draft, a completeness decision, and a selected review destination. Approval remains pending. The observer records the conditions and output before offering coaching.

Build a Normal Case and One Meaningful Exception

Create two packets with the same structure. Make the difference relevant to the decision being taught. Avoid unrelated defects that turn a narrow classification exercise into a search for hidden traps.

The following sample rules are invented for this article. A complete packet needs a requester, service description, required service-scope attachment, and request identifier. Insurance information is outside this fictional exercise. No claim about real supplier requirements follows from these choices.

Input Normal case Exception case
Request identifier SYN-101 SYN-102
Fictional supplier Sample Alder Services Sample Birch Services
Requester Practice coordinator A Practice coordinator B
Requested service Office equipment inspection Office equipment inspection
Required scope attachment Dummy scope document present Scope document absent
Governing rule Complete packets go to intake review Incomplete packets remain on hold with the missing input recorded
Expected final state Draft awaiting intake review Draft on hold; scope attachment requested in the practice note

Give the learner a blank draft for each case. Include fields for identifier, supplier, service, document status, next step, and source-packet reference. Keep the expected outcomes on the observer's copy.

For the exception, stopping is part of the task. The learner should identify the absent scope attachment, preserve the draft, and select the hold state. Do not reward movement toward approval when the exercise requires a stop.

If the learner identifies an ambiguity you overlooked, pause the affected criterion. Ask the process owner to resolve it before scoring that decision. A hidden rule added during observation changes the assessment.

Agree on a Bounded Success Rubric

Write the rubric before the attempt. Check it against the learning objectives and the practice environment. Tell the learner what evidence the observer will collect.

Use three result labels for each criterion: demonstrated independently, demonstrated with a prompt, or not yet demonstrated. Add untested when a condition prevents observation. These labels are choices for this worksheet rather than a universal grading scale.

Criterion Evidence required in this exercise
Correct identification Each draft retains its assigned synthetic request identifier
Required fields Each of the six draft fields contains the correct value or an explicit missing-input note
Normal-case decision SYN-101 goes to intake review with approval pending
Exception decision SYN-102 stays on hold with the absent scope attachment recorded
Boundary respected Neither draft is marked approved; no real action is attempted
Traceable output The observer can connect each draft to its source packet and checklist version

For this fictional exercise, the intended independent result requires every applicable criterion without a decision hint. Apply the normal-case decision to SYN-101 and the exception decision to SYN-102. Apply the other criteria to both drafts. The rubric permits the reference checklist. There is no speed target because this exercise has not established a justified timing requirement.

A correct final state after a hint is useful practice evidence. Record it as prompted. Do not count it as the same observation as independent completion.

The result describes these attempts. It does not certify competence across all suppliers, software versions, exceptions, or future work. Keep that limitation in the review note.

Copy This Observation Sheet

Record the session date, task version, checklist version, observer, and allowed support above the sheet. Use case identifiers rather than private business records.

Case and criterion Independent completion Prompts needed Mistake observed Recovery Support still required
SYN-101: prepare draft Yes in this fictional attempt None None observed Not applicable Workplace application remains untested
SYN-102: classify packet No in this fictional attempt Observer asks the learner to check the attachment rule Learner initially selects intake review Learner identifies the missing scope and selects hold Another exception case with the checklist available
SYN-102: preserve boundary Yes in this fictional attempt None No approval action attempted Not applicable No additional support identified for this criterion
[Your case and criterion] [Yes, no, or untested] [Exact prompt or none] [Observable action] [What changed and who supplied the answer] [Specific next practice or unresolved condition]

These completed rows are fictional. They show how a corrected answer can coexist with a recorded prompt.

Preserve the initial action before recording recovery. If the learner corrects a mistake without help, record self-correction. If the observer points to the missing attachment, record that intervention. Both may produce a correct draft, but their evidence differs.

Ask the learner which instruction they used. Write the explanation as an explanation, not proof that the action occurred. Inspect the draft and selected state separately.

After observation, give specific feedback. Identify the decision that matched the rubric, the decision needing practice, and the support available. Avoid turning one exception error into a statement about motivation or general ability.

Check the Change You Can Support

If your question concerns learning change, use comparable tasks before and after instruction. Keep conditions, allowed aids, and scoring criteria consistent. Use different synthetic identifiers so recalling the first answer does not complete the second task.

CDC explains that a posttest alone cannot establish a change in learning. Without a baseline, report the observed post-session performance and its limits.

For the fictional coordinator, the immediate result is narrow: the normal case meets the rubric independently; the exception requires a decision prompt. The next step is another exception exercise after reviewing the attachment rule.

Do not claim a financial result from those drafts. Do not infer that the session caused every later change in workflow performance. Record what the comparison measures and which conditions differ.

Follow Up After a Real Opportunity to Apply the Task

Choose follow-up timing around an opportunity to perform the workflow. A calendar date alone cannot establish that the learner encountered an eligible case. Agree on the check with the process owner and learner.

CDC's delayed-evaluation guidance recommends assessing application after learners return to work. Its questions include resources and opportunities to use the learning.

Start the follow-up with conditions. Does the coordinator have the required access? Is checklist version three still current? Has an eligible intake arrived? Can the coordinator use the agreed support?

If there has been no eligible case, record “application not yet observed.” If access is blocked, assign the access issue. Neither finding establishes that the learner failed to retain the lesson.

For a fictional eligible case, the reviewer might inspect the draft's document status, review destination, and pending approval boundary. Use authorized evidence in the working system. Keep private supplier attachments outside a broadly shared training report.

Ask what support the coordinator used. A reference checklist is permitted support in this plan. Help from another person belongs in the record too. Keep the difference between independent completion and supported completion visible.

Use operations handover to settle responsibility and ongoing support. Keep this follow-up focused on the observed task. A handover acknowledgment and a learning observation answer different questions.

Revise the Lesson or the Workflow Support

Use each finding to name a next action. If the learner knows the rule but cannot find it, review its placement in the SOP. If two instructions disagree, give the process owner the conflict before adding more practice.

If the exception distinction remains unclear under working conditions, add a contrasting sample and another supported attempt. Then observe a fresh case with the same allowed aids. Keep earlier results rather than replacing them with the successful attempt.

When considering task delegation, use the evidence within its boundary. State which routine and exception cases were observed, what support remains available, and who handles unfamiliar cases. Avoid expanding a narrow result into unrestricted authority.

Close the evaluation with the task version, observed outcomes, unresolved conditions, next action, and review owner. Preserve attendance and learner feedback beside that record. Let the actual work carry the performance claim.

Record the attempt. Improve the support.

Sources

Topic map: Records, access, and continuity

Articles in this cluster, then the 3 clusters that share the most topics with it.

This cluster: Records, access, and continuity

Neighbor: Business reporting and statistics

Neighbor: Spreadsheets, CSV files, and data handoffs

Neighbor: APIs, webhooks, and automation