WF-04 — Annual Programme Monitoring and Periodic Review

This is the PDCA loop for the whole pilot, and its hardest single job is a classification. Every KPI that comes back below target is one of three things: a reporting gap (the activity happened and was not submitted), a data-quality gap (it was submitted and the number is wrong), or a performance gap (the number is right and the performance is low). The three have completely different remedies, and the standing temptation is to book a reporting gain as a performance improvement. MBRU's own score moved from 45.0 to 63.9 in seventeen days in August 2026 without a single student being taught differently. Nothing about that movement was performance, and a monitoring process that records it as improvement will draw exactly the wrong conclusion about what to do next.

1. Purpose

To ensure that each IoL programme is reviewed annually and periodically against its own objectives and against every OBEF KPI it produces, that each shortfall is correctly classified as a reporting, data-quality or performance gap, that a Corrective Action Plan is opened on the MoHESR template where a genuine performance gap exists, and that the actions arising are owned, resourced, dated and closed, so that the programme improves and the improvement is visible rather than asserted.

2. Scope

Scope statement. This process manages programme monitoring from the close of an academic year, through the assembly and interpretation of performance and KPI evidence, the classification of gaps, the agreement of actions and the opening of Corrective Action Plans, to the closure of those actions and their carry forward into the next cycle.

Applies to. PGDipHPE and MScHPE. The annual monitoring cycle, the periodic (multi-year) review on the CAA cycle, and the KPI-level improvement loop for all 22 programme-level KPIs, including those IoL does not itself compute.

Does not apply to. The production of any individual KPI value, which belongs to the workflow that owns it. The OBEF submission itself, which is WF-26. Programme modification, which is WF-01; this workflow proposes changes and WF-01 disposes of them. Individual student appeals and complaints, which are WF-11, though their aggregate trends are an input here.

Applicable requirements. OBEF guide v11.5, all 22 programme-level KPIs; the MoHESR Corrective Action Plan template held at MoE Dashboard/Corrective Action Plan (CAP) for Continuous Improvement.xlsx; CAA requirements for annual programme monitoring and periodic review; MBRU quality assurance policy [IoL to confirm the governing policy and the committee that receives the annual report].

What this workflow is for, stated plainly. Not to produce a report. To produce decisions, and to make next year's numbers reproducible by someone who was not in the room.

3. Trigger, boundary and endpoint

4. SIPOC

5. Accountability

Process owner. IoL Director of Programmes, or the role able to commit IoL resource to an improvement action and to take a change proposal to curriculum governance.

Escalation.

6. Process steps

1. Assemble the evidence set. For each of the 22 programme-level KPIs: the submitted numerator, denominator and result; the Mechanism of collecting the data statement used; the returned dashboard score; the prior two years for the rolling-average KPIs; and the capture-completeness measures from each producing workflow's section 13. [CAPTURE] the evidence set with its extract date. A KPI submitted null is evidence, not an absence of evidence. Record it as a value of null with the reason, because the reason determines the classification at step 3.

2. Compare against target and prior year. Use the prospective targets in 01_Score_Maximisation_Plan.md as the comparison, not a peer average. OBEF anchors are absolute, so the target is knowable in advance and the comparison is meaningful. Record the variance per KPI.

3. Classify every shortfall. This is the step the workflow exists for. Three categories, and each KPI gets exactly one.

   Classification	Test	Correct remedy	Wrong remedy
   Reporting gap	The activity occurred at the required standard, and it was not submitted, or was submitted null, or was submitted without the attributes the KPI requires	Fix the capture point in the producing workflow. The gain appears next cycle and is not an improvement in performance	Opening a CAP to "improve performance" on something that is already happening
   Data-quality gap	The activity occurred and was submitted, and the number is wrong: wrong denominator basis, wrong aggregation unit, a misapplied exclusion, an unreconciled register, or a classification error	Fix the computation and, where material, restate. Record the restatement	Reporting the corrected number as an improvement
   Performance gap	The activity occurred, the number is right, and the value is genuinely below the anchor IoL is aiming at	Open a CAP. This is Category 5 in the charter, and reporting it accurately is the entire response	Reclassifying it as a reporting gap because that is cheaper to fix

   [CAPTURE] the classification per KPI with a written rationale.

4. Validate the classification independently. Quality and IQA reviews the classifications without the Programme Director present at the decision. The person who benefits from a favourable classification does not make it alone. This mirrors the separation the charter requires for the KPI 5.2 eligibility call and the Appendix C grouping decision, and for the same reason. Decision point. Where the two disagree and cannot resolve it, escalate to the pilot sponsor. Do not settle it by choosing the classification that produces the shorter action list.

5. Distinguish the gain from the improvement in the narrative. The annual report must state, for each KPI that moved, how much of the movement is attributable to reporting, to data quality and to performance. Where the split cannot be quantified, say so. A report that attributes a capture gain to teaching quality misleads its own institution, which is a worse outcome than misleading anyone else.

6. Agree actions with owners, resources and dates. One named owner per action. An action with a committee as its owner has no owner. Where an action requires resource that IoL does not hold, say what is needed rather than recording the action as agreed.

7. Open or update the Corrective Action Plan. For every KPI classified as a genuine performance gap, complete a row on the MoHESR CAP template at MoE Dashboard/Corrective Action Plan (CAP) for Continuous Improvement.xlsx. The template is structured per KPI and requires: historical performance by year, benchmarking against other HEIs (mean and 90th percentile), targets after one and two years, improvement actions with resources, the action owner, the expected contribution to the KPI target, progress to date, and progress status. [CAPTURE] the completed CAP row. Note the benchmarking columns and what they mean here. OBEF scoring is absolute rather than peer-relative, so the mean and 90th percentile columns do not set the target; the published anchor does. Complete them where the comparative data is available to IoL and record their absence where it is not. [IoL to confirm what benchmarking data MBRU receives from the dashboard.]

8. Refer design changes to WF-01. Any action requiring a study-plan change, a new programme attribute, a mobility option or a mandatory placement goes to WF-01 and then to curriculum governance. This workflow proposes; governance disposes.

9. Update the pilot's assumptions. The Score Maximisation Plan states plainly that its Scenario 1 values are informed estimates rather than measurements. After the first full cycle, replace them with real numbers and reissue. [CAPTURE] the revised assumption set with the date.

10. Approve and publish. Curriculum governance approves the report. Publish a learner-facing summary alongside it, which is the aggregate half of the "you said, we did" loop that WF-06 runs at course level.

11. Monitor progress at defined points. At least twice between annual reviews, update progress to date and progress status on each CAP row. A CAP updated only at year end is a document, not a plan.

12. Close or carry forward. Every action is closed with evidence or carried forward with a reason. An action carried forward twice is escalated. [CAPTURE] the closure record.

7. OBEF data generated

Capture rule. Two things cannot be reconstructed after the cycle closes. The classification rationale at the time it was made, and the state of the evidence when the decision was taken. A year later, everyone remembers the gap as whichever kind it was cheapest to have been. Write the rationale down while the evidence is in front of you, and keep the evidence set with it.

Reproducibility test. A second analyst reading the annual monitoring report, the gap classification register and the CAP should be able to say, for each KPI, what the value was, why it was below target, who is fixing it and by when, without asking anyone. Where the report says "performance improved" they should be able to see whether the improvement was in the teaching or in the recording. If they cannot, the report has failed at its only real job.

A note on why this section is short. This workflow generates no KPI values. Everything in the table above is metadata about KPI values produced elsewhere. That is proper: a monitoring process that produced its own numbers would be marking its own work.

8. Max-score design

Why this section still matters in a workflow with no KPI. The scarcest resource in the pilot is the attention of the people who can change things. Misclassifying a reporting gap as a performance gap spends that attention on improving something that is already good, and the number does not move because the number was never about the teaching. Misclassifying a performance gap as a reporting gap does the reverse and is worse, because it produces a confident institution with a real problem.

What this workflow must do to reach the target.

1. Classify every below-target KPI, with a written rationale, in the cycle in which it appears.
2. Have the classification independently validated, with the separation of duties actually observed.
3. State, in the annual report, how much of every movement is reporting, data quality and performance.
4. Open a CAP for every genuine performance gap, and only for genuine performance gaps.
5. Update CAP progress at least twice between annual reviews.
6. Replace the Score Maximisation Plan's estimated Scenario 1 values with measured ones after cycle one.

[REDESIGN] actions.

9. Indirect strategy where data cannot be collected

Category 5 applies: performance is genuinely low. This workflow is where Category 5 is operationalised, and the charter is deliberately blunt about it. The sanctioned response is to report the value accurately and open a corrective action plan using the MoHESR template already held in the project folder. That is the entire response. There is no tactic, no reclassification and no presentational route. A programme with a genuine performance gap that reports it accurately and works on it is in a defensible position; the same programme reporting a better number is in an indefensible one and has also lost the information it needed.

The discipline this requires, and why it is harder than it sounds. Category 5 is the only one of the five categories where the correct action produces a worse-looking number in the short run. Categories 1 to 4 all involve doing something that both improves the score and is legitimate. Category 5 involves writing down a low number. The pressure to move a KPI from Category 5 to Category 4 is therefore constant, and the classification at step 3 is where that pressure lands. The independent validation at step 4 exists for exactly this reason, and it is worth more than any other control in this document.

The specific failure this workflow must prevent, stated with its evidence. MBRU's OBEF score moved from 45.0 to 63.9 between 10 and 27 August 2026. That movement was data submission. No teaching changed, no curriculum changed, and no graduate outcome changed in seventeen days. A monitoring process that records that as an 18.9 point improvement in institutional performance will conclude that MBRU is improving rapidly and needs no structural change, which is close to the opposite of the truth. The charter's central proposition is that the gap between what IoL does and what IoL scores is mostly a capture and classification gap. Confirming that proposition requires being able to separate the two, and that separation is made here or nowhere.

The reverse error, which is rarer and worse. A genuine performance gap booked as a reporting gap produces an institution that believes it has a spreadsheet problem when it has an educational one. It generates no CAP, no resource and no action, and the gap persists across the three years of a rolling average, by which point it is expensive to move. Where the evidence does not clearly distinguish the two, classify as a performance gap. That is the conservative direction: it costs an unnecessary CAP row and protects against the more damaging error.

A note on the other four categories, which pass through here. This workflow does not own Categories 1 to 4, but it audits them. Each cycle it should confirm that Category 2 grouping decisions were made before the surveys ran and not after, that Category 3 redistribution classifications remain accurate, that Category 4 capture actions actually closed the gaps they were built for, and that no Category 1 KPI is being treated as beyond influence when its rubric is published and its input is controllable.

Boundary check. This workflow must never:

• classify a genuine performance gap as a reporting or data-quality gap because the remedy is cheaper;
• describe a reporting or data-quality gain as an improvement in performance in any report, internal or external;
• allow the Programme Director to be the sole signatory on a gap classification;
• close an action without evidence that it was done;
• revise a methodological basis after seeing which basis produces a better number, which is the data-quality equivalent of revising an Appendix C grouping after the results are visible;
• complete the CAP benchmarking columns with estimated peer data presented as actual;
• carry an action forward a third time without escalating it;
• treat the CAP as a document produced for the Ministry rather than as IoL's own improvement plan.

10. Service standards

11. Records and evidence

Appendix B readiness. The CAP itself is the most likely artefact to be requested from this workflow, and because it is a Ministry template it should be producible immediately once populated. The record IoL would struggle to produce today is the gap classification register with rationales, because on the assumption that no such register currently exists [IoL to confirm], there is nothing to produce. That is action R1, and it is also the record that would answer the hardest question a reviewer can ask, which is why a number changed.

12. Risks and controls

13. Performance measures

The measure to watch is "reporting or data-quality gains described as performance improvements", and its target is zero. It is the only measure in this pack whose failure is invisible in every other measure. A programme can score well on everything else in this table while systematically misunderstanding why its own numbers moved.

14. Change control