Monitor-Evaluate-Pivot Loop
Track implementation, evaluate outcomes, and revise faulty assumptions
- Difficulty
- Moderate
- Time to result
- ~ongoing to results
- Steps
- 7
- Confidence
- 94%
The Monitor-Evaluate-Pivot Loop separates two complementary forms of learning. Monitoring follows the program while it is being implemented, revealing whether activities, resources, and intermediate milestones are proceeding as expected. Evaluation examines the outcomes produced after implementation and asks whether the intended change actually occurred. Both streams are compared with the assumptions made at the program's starting point. If delivery is weak, managers can repair execution; if delivery is strong but outcomes are poor, they may need to revise the intervention or its underlying diagnosis. The loop makes measurement operational rather than ceremonial: evidence protects beneficiaries, informs resource allocation, and creates explicit opportunities to pivot before additional capital is committed to an ineffective approach.
Origin
Extracted from Masters in Business, where Saadia Madsbjerg explains the Rockefeller Foundation's distinction between monitoring during implementation and evaluation of outcomes.
Core principles
- 01Implementation data and outcome evidence answer different questions
- 02Measurement exists to protect intended beneficiaries
- 03Initial assumptions must remain open to revision
- 04Programs should adapt while evidence can still improve them
How to run it
- 1
State the intended outcome
Define the change expected in beneficiaries' lives and the time horizon in which it should appear.
Pro tip Use an outcome that can be observed rather than a broad aspiration.
Watch out Activities such as grants issued are not beneficiary outcomes.
- 2
Record the starting model
Document the causal assumptions linking program activities to the intended outcome.
Pro tip Identify the assumptions carrying the greatest uncertainty.
Watch out Unwritten assumptions cannot be tested or revised transparently.
- 3
Design dual measures
Select indicators for implementation progress and separate indicators for final or intermediate outcomes.
Pro tip Keep the indicator set small enough to influence decisions.
Watch out Monitoring delivery alone can make an ineffective program appear successful.
- 4
Monitor implementation
Collect evidence while the program operates and address deviations in timing, reach, quality, or resources.
Pro tip Set review points before launch instead of waiting for a crisis.
Watch out Do not alter the program for every short-term fluctuation.
- 5
Evaluate outcomes
Assess whether the intended effects occurred and whether the program plausibly contributed to them.
Pro tip Compare results with a baseline or credible counterfactual when possible.
Watch out Positive anecdotes cannot substitute for representative outcome evidence.
- 6
Diagnose and pivot
Determine whether shortcomings arise from execution, the intervention design, or the original understanding of the problem, then adjust accordingly.
Pro tip Match the size of the pivot to the strength of the evidence.
Watch out Defending sunk costs can keep disproven assumptions alive.
- 7
Repeat the loop
Continue monitoring and evaluation after changes to determine whether the revised program performs better.
Pro tip Maintain a decision log connecting evidence to each change.
Watch out A pivot without follow-up measurement produces no reliable learning.
In the wild
A program monitors whether farmers use a new market-linking platform and finds high enrollment but delayed crop collection. Evaluation then shows that spoilage has not fallen. Managers discover that transport capacity, rather than farmer participation, is the binding constraint and redirect resources toward coordinated pickup and local processing.
→ The program changes its causal mechanism instead of mistaking enrollment for impact.
A nonprofit monitors timely payment delivery and recipient reach, then evaluates whether household financial stability improves. If payments arrive correctly but the target outcome remains unchanged, it investigates benefit size, timing, complementary services, and the original diagnosis.
→ Implementation success is distinguished from actual beneficiary impact.
Common mistakes
Confusing monitoring with evaluation
Knowing that a program was delivered as planned does not establish that it produced the intended outcome.
Measuring without decision points
Data collection creates administrative burden rather than learning when nobody is responsible for acting on the results.
Refusing to revisit the starting point
Weak results may expose a faulty diagnosis, not merely poor execution.
Is it for you?
Best for
Multi-stage programs where managers can observe delivery, evaluate effects, and adjust the intervention over time.
Not ideal for
One-off activities with no meaningful outcome measurement or opportunity to adapt.
From the transcript
“monitoring, which is doing it as the program is implemented, and then evaluation, which is the actual outcomes of the program once it has been…”
“we want to make sure that the impact we want to have on the life of pinvulnerable people is actually achieved”
“whether the starting point that we had for when we started the program was the right one or whether we need to pivot along the…”
From the episode
Interview With Saadia Madsbjerg: Masters in Business (Audio)
Saadia Madsbjerg