Review the Programme
When To Use
Evaluation data is filed and next year's priorities are set from the inbox. The programme ran, the report went out, the team moved to the next build, and the same course will run again unchanged because nobody stopped to ask. Run this once the evidence is in, to answer one question: given what we said would happen and what did, do we keep this programme, fix it or retire it?
When Not To Use
If the readers are executives who want one page, run Report What Changed; this review is the team talking to itself. If the programme has not run yet and you want to find its weak points, run Plan the Pilot.
Inputs
- What was written before launch: the evaluation plan, the design document, the pilot plan
- What happened: the report, the follow-up on use, the feedback summary, with periods and sources
- Who takes part in the review, and who makes the call on the programme
If you have none of this, I start from the programme description and what the team remembers intending, and mark the output as a first draft with every "expected" line marked reconstructed.
Approach
The review is the after-action review as USAID's public technical guidance sets it out: four questions in order (what was expected to happen, what actually occurred, what went well and why, what can be improved and how), in a session about learning, not critique. The judgement: compare against the page written before launch, not against what people now recall intending, because memory quietly moves the target to wherever the result landed. The failure it prevents: a review that turns into a hunt for whoever chose the wrong date, ends in an awkward silence, and changes nothing in the next build.
Workflow
- Ask three questions: where is the intended result written down, which evidence do you hold and for what period, and who makes the keep, fix or retire call?
- Expected: copy the result, the behaviours and the pass line from what was written before launch. If nothing was written, say so and mark the line "reconstructed"; that gap is itself a finding about the process.
- Actual: fill each line from your evidence only, with period and source. Gaps stay gaps, with the role that could close them.
- Went well, and why; to improve, and how. Causes are about the programme, the process and the assumptions. A cause is never a person: "the brief arrived after the build started" is a cause, a name is not.
- Turn the answers into a keep and change list in two parts: the programme itself, and how the team built it. Each item has an owning role and a date, or it is a wish.
- Draft the three options, keep, fix and retire, each with evidence for and against. Retiring a course nobody uses is a result, not a defeat. The review recommends; the named person makes the call.
- Write the items for next year's plan so they can be lifted straight into it: the need, the evidence, the size of the work.
Output Format
# Programme Review
Programme: [name] | Ran: [dates] | Review held: [date] | Facilitator: [role] | Call made by: [named decider]
## Expected and actual
| What was expected (result, behaviour, pass line) | Written where, before launch | What occurred (period, source) | Gap |
|---|---|---|---|
| [expected] | [document, or "reconstructed"] | [from your evidence, or "missing"] | [difference, or "unknown"] |
## Why
- Went well: [what held] | why: [cause in the programme, the process or an assumption]
- To improve: [gap] | how: [change] | likely cause, never a person: [cause, evidence]
## Keep and change
| Part | Item | Keep or change | Owning role | By |
|---|---|---|---|---|
| The programme | [item] | [keep / change] | [role] | [date] |
| How the team worked | [item] | [keep / change] | [role] | [date] |
## Options
- Keep: [evidence for] / [evidence against] | Fix: [evidence for] / [evidence against] | Retire: [evidence for] / [evidence against]
Recommendation: [option, in one sentence]
## Items for next year's plan
- [need] | [evidence] | [size of the work] | [sponsor role]
## Decision
[Named decider] makes the keep, fix or retire call by [date]; the items above go into next year's plan at [planning date].
Done When
- Every "expected" line cites a document written before launch or is marked reconstructed; every "actual" line cites evidence with its period or reads "missing"
- No cause names a person, and each keep and change item has an owning role and a date
- All three options carry evidence both ways, and the plan items can be read without this review
Quality Bar
- Four questions, in order. Jumping to what went wrong makes a complaint session; praise with no change fails the review too.
- Evidence stays at group level, with small groups suppressed at the minimum you set
- No invented figure, quote or finding; the review recommends, shows its reasons and does not pre-empt the call
- The review is about the programme, never about a person; a named person makes the keep, fix or retire call
Next
Run ldg-plan-the-year (Plan the L&D Year) to carry what was learned into next year's priorities.
About the makers
This pack is made by 90n (90n.ai): AI-native ways of working and tools for Product, Marketing and HR teams, built by ex-McKinsey founders. We map where time and money leak, redesign the work around AI and stay until your people run it themselves, and we run our own company on Claude. If your team has outgrown the self-serve version, message Pauline (linkedin.com/in/paulinebertry).
1---2name: ldg-review-the-programme3description: Prepares the team's review of one training programme, setting what was intended before launch against what happened, with the reasons, a two-part keep and change list, a keep, fix or retire recommendation, and the items for next year's plan. Use for "run ldg-review-the-programme", "after action review", "program review", "training programme review", "lessons learned", "retire a course", "course audit", "should we keep running this course", part of The Complete Claude Guide for L&D Pack by 90n.4metadata:5 source: https://skillmd.com/skills/polar-bear-org/ldg-review-the-programme6---78# Review the Programme910## When To Use11Evaluation data is filed and next year's priorities are set from the inbox. The programme ran, the report went out, the team moved to the next build, and the same course will run again unchanged because nobody stopped to ask. Run this once the evidence is in, to answer one question: given what we said would happen and what did, do we keep this programme, fix it or retire it?1213## When Not To Use14If the readers are executives who want one page, run Report What Changed; this review is the team talking to itself. If the programme has not run yet and you want to find its weak points, run Plan the Pilot.1516## Inputs17- What was written before launch: the evaluation plan, the design document, the pilot plan18- What happened: the report, the follow-up on use, the feedback summary, with periods and sources19- Who takes part in the review, and who makes the call on the programme20If you have none of this, I start from the programme description and what the team remembers intending, and mark the output as a first draft with every "expected" line marked reconstructed.2122## Approach23The review is the after-action review as USAID's public technical guidance sets it out: four questions in order (what was expected to happen, what actually occurred, what went well and why, what can be improved and how), in a session about learning, not critique. The judgement: compare against the page written before launch, not against what people now recall intending, because memory quietly moves the target to wherever the result landed. The failure it prevents: a review that turns into a hunt for whoever chose the wrong date, ends in an awkward silence, and changes nothing in the next build.2425## Workflow261. Ask three questions: where is the intended result written down, which evidence do you hold and for what period, and who makes the keep, fix or retire call?272. Expected: copy the result, the behaviours and the pass line from what was written before launch. If nothing was written, say so and mark the line "reconstructed"; that gap is itself a finding about the process.283. Actual: fill each line from your evidence only, with period and source. Gaps stay gaps, with the role that could close them.294. Went well, and why; to improve, and how. Causes are about the programme, the process and the assumptions. A cause is never a person: "the brief arrived after the build started" is a cause, a name is not.305. Turn the answers into a keep and change list in two parts: the programme itself, and how the team built it. Each item has an owning role and a date, or it is a wish.316. Draft the three options, keep, fix and retire, each with evidence for and against. Retiring a course nobody uses is a result, not a defeat. The review recommends; the named person makes the call.327. Write the items for next year's plan so they can be lifted straight into it: the need, the evidence, the size of the work.3334## Output Format35```markdown36# Programme Review37Programme: [name] | Ran: [dates] | Review held: [date] | Facilitator: [role] | Call made by: [named decider]38## Expected and actual39| What was expected (result, behaviour, pass line) | Written where, before launch | What occurred (period, source) | Gap |40|---|---|---|---|41| [expected] | [document, or "reconstructed"] | [from your evidence, or "missing"] | [difference, or "unknown"] |42## Why43- Went well: [what held] | why: [cause in the programme, the process or an assumption]44- To improve: [gap] | how: [change] | likely cause, never a person: [cause, evidence]45## Keep and change46| Part | Item | Keep or change | Owning role | By |47|---|---|---|---|---|48| The programme | [item] | [keep / change] | [role] | [date] |49| How the team worked | [item] | [keep / change] | [role] | [date] |50## Options51- Keep: [evidence for] / [evidence against] | Fix: [evidence for] / [evidence against] | Retire: [evidence for] / [evidence against]52Recommendation: [option, in one sentence]53## Items for next year's plan54- [need] | [evidence] | [size of the work] | [sponsor role]55## Decision56[Named decider] makes the keep, fix or retire call by [date]; the items above go into next year's plan at [planning date].57```5859## Done When60- Every "expected" line cites a document written before launch or is marked reconstructed; every "actual" line cites evidence with its period or reads "missing"61- No cause names a person, and each keep and change item has an owning role and a date62- All three options carry evidence both ways, and the plan items can be read without this review6364## Quality Bar65- Four questions, in order. Jumping to what went wrong makes a complaint session; praise with no change fails the review too.66- Evidence stays at group level, with small groups suppressed at the minimum you set67- No invented figure, quote or finding; the review recommends, shows its reasons and does not pre-empt the call68- The review is about the programme, never about a person; a named person makes the keep, fix or retire call6970## Next71Run ldg-plan-the-year (Plan the L&D Year) to carry what was learned into next year's priorities.7273## About the makers7475This pack is made by 90n ([90n.ai](https://90n.ai)): AI-native ways of working and tools for Product, Marketing and HR teams, built by ex-McKinsey founders. We map where time and money leak, redesign the work around AI and stay until your people run it themselves, and we run our own company on Claude. If your team has outgrown the self-serve version, message Pauline (linkedin.com/in/paulinebertry).