Document a Program Failure
Document and structure learning from a program that did not meet its targets.
Draft: Document a Program Failure
Fill In Your Details (2 Blanks)
Your Prompt (With Your Details)
This is the complete prompt, ready to copy and paste.
Evidence and applicability: use only supplied or verifiable evidence for historical facts, values, quotations, completed activities and approvals. Preserve conflicting sources with their provenance; do not choose a preferred value without adjudication. Mark missing evidence as unknown or not evidenced, not as proof an activity never occurred. Keep proposed plans and illustrative examples separate from findings. Never invent a baseline, quote, pilot, validation or sign-off to complete a template. Match methods to the purpose and population; a suitable single-method design is valid. Distinguish descriptive change from causal attribution. Check obligations in the applicable funding agreement rather than assuming a universal donor rule.
You are a senior MEAL specialist documenting learning from a program that did not meet its targets.
The goal is honest, useful documentation of underperformance, written with discipline and without blame.
Requirements:
- State the underperformance. Quantify it: what was the target, what was achieved, over what period, and for which sub-groups. Avoid softening language.
- Root-cause analysis. Identify 3-5 plausible drivers across design, implementation, context, and assumptions. For each, provide supporting evidence and counter-evidence.
- What was tried. Document the adjustments that were attempted during implementation and whether they helped, hurt, or had no measurable effect.
- Counterfactual reflection. Briefly consider what likely would have happened with a different design choice, and what is unknowable.
- Lessons. Translate the analysis into 4-6 lessons. Each must be specific and decision-oriented. Avoid generic statements such as 'we needed more time.'
- Tone. Factual, accountable, blame-free. Distinguish between individual conduct (not in scope) and structural drivers (in scope).
Output Format:
Produce:
- A short framing paragraph.
- A performance table (target, actual, gap, period, sub-group note).
- A root-cause section with evidence per driver.
- A 'what was tried and what happened' table.
- A lessons list, with each lesson tagged to the decision it should inform.
Check: Score It with the Lessons Learned Quality Rubric
The Review Prompt
Paste the draft at the bottom of this prompt and run it. It scores each criterion from 1 to 5 and says why.
Use source evidence; retain unknowns/conflicts. N/A or missing evidence means no total. Critical ethics/validity failures block use at any score. Scores never authorize fielding or compliance.
You are an expert evaluator and learning specialist. Score the lessons learned section I will provide using the rubric below. Treat the lessons learned as a standalone product: a useful lessons learned section should help readers in similar programs make better decisions, not just summarize what this team did.
SCORING RUBRIC - Lessons Learned Quality Rate each dimension 1-5:
DIMENSION 1: Evidence Basis
- Score 5: Every lesson cites the specific finding, data point, or event it draws from. The reader can trace each lesson back to evidence in the report. Where multiple findings support a lesson, all are cited. Lessons that contradict surface-level data acknowledge what was reconciled and how.
- Score 4: At least 80 percent of lessons cite their source finding or evidence. The remainder are plausible inferences from the body of the report but the link is implied rather than stated.
- Score 3: Roughly half of lessons are traceable to findings. The rest read as the author's general impressions. Some lessons appear consistent with the report but no direct link is shown.
- Score 2: Most lessons appear to reflect author opinion without documented evidence. Citations to findings are rare or absent. The section reads as commentary detached from the analysis.
- Score 1: Lessons are untraceable to anything in the report. No evidence basis is shown.
DIMENSION 2: Specificity
- Score 5: Every lesson is specific to a context, mechanism, or condition. Each lesson names the population, sector, intervention type, or operating condition it applies to. No lesson could be cut and pasted into an unrelated program report and still make sense.
- Score 4: At least 80 percent of lessons are specific in this sense. A small number drift toward general statements but the bulk are anchored.
- Score 3: Roughly half of lessons are specific. The rest read as generic ("partnerships matter," "context is important," "data quality is critical") without naming what made them apply here.
- Score 2: Most lessons are generic platitudes that could be lifted into any program report. Specificity is rare.
- Score 1: All lessons are platitudes with no specificity at all.
DIMENSION 3: Generalizability
- Score 5: Every lesson explicitly states its scope of applicability. The reader knows whether a lesson applies only to this program, to similar programs (and what makes them similar), or to the sector broadly. Boundary conditions are named where they matter.
- Score 4: At least 80 percent of lessons state their scope. A small number leave scope implicit but the surrounding context makes it inferable.
- Score 3: Some lessons name their scope; others do not. The reader has to guess whether a lesson is a one-off observation or a broader principle.
- Score 2: Scope is rarely addressed. Lessons are presented without any signal about whether they generalize.
- Score 1: No lesson states its scope. The reader cannot tell which lessons might apply elsewhere.
DIMENSION 4: Counter-Lessons
- Score 5: Lessons about what did not work, what proved harder than expected, what produced unintended or negative effects, and what the team would do differently are surfaced with equal weight to successes. The section reads as a balanced learning product, not a marketing artifact.
- Score 4: Counter-lessons are present and meaningful, though they may receive somewhat less space than success lessons. Difficulties and missteps are not hidden.
- Score 3: One or two counter-lessons appear but are framed softly or buried. The dominant tone is still success-oriented.
- Score 2: Counter-lessons are nearly absent. Difficulties, if mentioned, are reframed as opportunities or external factors rather than lessons.
- Score 1: No counter-lessons. The section reads as a success story.
DIMENSION 5: Action-Oriented Framing
- Score 5: Every lesson is framed so that a reader can derive an action or decision from it. Lessons name what to do differently, what to keep doing, what to stop, or what to test. Verbs and decisions are visible.
- Score 4: At least 80 percent of lessons are action-oriented. A small number remain observational but the bulk are decision-ready.
- Score 3: Roughly half of lessons are action-oriented. The remainder are passive observations that leave the reader unsure what to do with the information.
- Score 2: Most lessons are passive observations or descriptive statements. Few suggest action.
- Score 1: Lessons are descriptive only. No action framing is present.
OUTPUT FORMAT: Return your assessment as a table followed by a summary:
| Dimension | Score (1-5) | Evidence | Priority Revision |
|---|---|---|---|
| Evidence Basis | |||
| Specificity | |||
| Generalizability | |||
| Counter-Lessons | |||
| Action-Oriented Framing |
Total: X/25 Band: Strong (22-25) / Adequate (17-21) / Needs Revision (11-16) / Substantial Revision (5-10) Single Most Important Revision: [One specific sentence]
For any dimension scored 1 or 2, add a brief explanation and a concrete revision example using one of the lessons from the document.
LESSONS LEARNED TO SCORE: [Paste your lessons learned section here]
Fix: Revise What Scored Low
The Revise Prompt
Paste the draft and its scores at the bottom of this prompt. It revises the parts that scored below 4.
Use source evidence; retain unknowns/conflicts. N/A or missing evidence means no total. Critical ethics/validity failures block use at any score. Scores never authorize fielding or compliance.
You are an expert evaluator and learning specialist. Review the lessons learned section I will provide and fix every problem that falls below the Good (4) threshold on the Lessons Learned Quality rubric.
WHAT TO LOOK FOR AND FIX:
Evidence problems: Lessons not tied to findings or data. Lessons that appear to reflect author opinion without documentation. Add a specific finding or data point citation to each lesson, or flag as "single-author observation" where no evidence exists.
Specificity problems: Generic platitudes ("partnerships matter," "context is important," "data quality is critical"). Rewrite each platitude to name the context, population, mechanism, or condition that made the lesson apply here.
Generalizability problems: Scope of applicability missing. Add a one-sentence statement for each lesson clarifying whether it applies only to this program, to similar programs (and what makes them similar), or to the sector broadly.
Counter-lesson problems: Section reads as success story. Add at least two counter-lessons covering what did not work, what was harder than expected, or what the team would do differently. Give them equal weight in framing.
Action-framing problems: Passive observations. Rewrite each passive lesson with an action verb (do, stop, keep, test, change, defer) and the decision it implies.
OUTPUT FORMAT: For each problem found:
- Dimension: [which dimension]
- Lesson affected: [quote or paraphrase]
- Problem: [one sentence describing what is wrong]
- Revised version: [the corrected lesson]
Then provide the complete revised lessons learned section incorporating all corrections.
LESSONS LEARNED TO REVISE: [Paste your lessons learned section here]