At a Glance
Process vs Outcome at a Glance
| Factor | Process indicators | Outcome indicators |
|---|---|---|
| What they measure | Implementation: delivery, fidelity, dose, coverage, quality of activities | Change: shifts in behavior, knowledge, capability, condition, status |
| What question they answer | Is the program running as designed? | Is the program producing the intended change? |
| Typical data source | Activity records, roster, supervision checklists, attendance logs | Surveys, assessments, biometric tests, observation of applied skills |
| Who is accountable | Program staff (implementation team) | Participants and the program jointly (outcomes require participant action) |
| Collection frequency | Set by delivery and decision needs | Set by what you measure, the expected change, and decision needs |
| Cost per collection cycle | Lower (administrative data) | Higher (requires survey or assessment) |
| Interpretation | Straightforward (did we deliver?) | Requires attribution thinking, meaning you must ask whether the program caused the change |
| Where they sit in results chain | Activity and output levels | Outcome level (may inform impact) |
Process and outcome indicators are not interchangeable. They serve different purposes, require different measurement methods, and answer different questions. A MEL plan that treats them as equivalent, or substitutes one for the other in reporting, produces data that does not tell the program's full story.
For the conceptual grounding, see indicator and indicator selection.
Where Do Process and Outcome Indicators Sit in the Results Chain?
The standard results chain is: inputs → activities → outputs → outcomes → impact/goal. Process and outcome indicators sit at specific points along this chain.
| Results chain level | Indicator type | Example |
|---|---|---|
| Inputs | Input indicators | Budget disbursed, staff hired |
| Activities | Process indicators | Training sessions delivered with full curriculum; client intake completed per protocol |
| Outputs | Output indicators | Number of kits distributed; number of households reached |
| Outcomes | Outcome indicators | Percentage of trained participants applying new skill 3 months later |
| Impact / Goal | Impact/goal indicators | Reduction in under-five mortality; improvement in livelihood security |
Process indicators overlap with output indicators. The distinction: outputs count (800 kits distributed); process indicators describe how well (training sessions delivered with 94% curriculum fidelity and 87% participant satisfaction). For many programs, the two are reported together in one indicator: "Percent of community health workers who pass the competency check after training delivered with at least 80% curriculum fidelity."
See output vs outcome vs impact for the full level taxonomy.
When to Use Each
When Do Process Indicators Matter Most?
Use process indicators when the program is early-stage, when implementation quality drives whether outcomes can emerge, or when adaptive management cycles depend on rapid feedback.
Early-stage programs. Before outcomes can meaningfully be measured, the program must be running. Process indicators tell you whether implementation is actually happening at the planned dose, fidelity, and reach. In the first 6-12 months of a multi-year program, process indicators are your primary monitoring data; outcome indicators may not have produced signal yet.
Implementation-fidelity-sensitive programs. Some programs depend on precise delivery: a trained evidence-based intervention for health, a specific pedagogical method in education, a standardized protocol in protection casework. For these, process indicators measuring fidelity (did the delivery match the protocol?) are outcome-critical. An outcome failure without process data leaves you unable to diagnose whether the intervention failed or just was not delivered.
Adaptive management. Quarterly or monthly adaptive cycles need indicators that change quickly enough to inform decisions. Process indicators update as activities happen; outcome indicators typically lag by weeks or months. Programs using real-time or quarterly learning cycles rely heavily on process indicators as the feedback stream.
Accountability for implementation. Program staff are directly accountable for process indicators (they control delivery). Outcome indicators depend partly on participant action and external context, so accountability is shared. Programs with strong implementation accountability cultures lean heavily on process indicators for day-to-day management.
When Do Outcome Indicators Matter Most?
Use outcome indicators when the program has been running long enough for change to emerge, when donor accountability is at the results level, or when evaluation questions focus on effectiveness.
Mid-term and endline phases. Outcome change needs comparable starting and follow-up measurements; endpoint attainment and some causal designs can be assessed with other data structures. Mid-term reviews and endline evaluations are outcome-indicator-heavy. If your endline report has more process than outcome data, the evaluation cannot answer the fundamental question of whether the program worked.
Reporting needs vary. Funding agreements and program decisions may call for process, output, coverage, and outcome measures. Check the actual requirements and choose measures at the level needed; neither process nor outcome measures belong exclusively to internal or external reporting.
Impact contribution. Outcome indicators are the evidence bridge between what the program did (process/outputs) and what changed at population scale (impact). Without outcome indicators, you cannot argue that the program contributed to the goal-level change.
Program learning about effectiveness. The question "what intervention design produces the best outcomes?" requires outcome data. Process data alone cannot answer this; two programs can have identical process indicators and very different outcome results.
How Do You Use Process and Outcome Indicators Together?
The strongest MEL plans use process and outcome indicators in pairs at each layer of the results chain. The pairing structure:
- For each outcome, specify the process indicators that feed it
- For each process indicator, specify the outcome the process is designed to produce
- When outcomes fail to emerge, process data tells you whether it was an implementation failure or an intervention-design failure
Example pair: Outcome indicator "% of trained birth attendants applying WHO-aligned hygiene protocol during delivery" (measured at 6 months post-training via observation). Process indicators feeding it: "% of birth attendants completing full 5-day training curriculum," "% of training sessions delivered with ≥80% curriculum fidelity," "% of trained attendants receiving at least one supervision visit within 3 months."
When the outcome is 45% and you were targeting 75%, the process indicators let you diagnose: if fidelity was 95% and supervision was 90%, the intervention design was weaker than expected; if fidelity was 60% and supervision was 30%, the implementation was weak and the intervention design may still be sound. Without the process indicators, you have a failed outcome and no explanation.
For the broader framework of how to size your indicator portfolio, see mistake too many indicators.
Designing Each Type
Designing Process Indicators
Good process indicators measure implementation with enough specificity that a spot-check could verify them. Three design rules.
Name the specific practice, not just the activity. "Trainings delivered" is weak; "Trainings delivered with full 5-day curriculum and pre/post test administered" is measurable and verifiable. The indicator should fail when implementation shortcuts are taken.
Tie to an implementation protocol. Process indicators presume that a defined protocol exists. If the training has no written curriculum, "curriculum fidelity" cannot be measured. Build the indicator around the protocol, not the protocol around the indicator.
Measure how well, not only how many. A program can train 100 community health workers (CHWs, the frontline staff who screen and counsel families), which is an activity count, not an indicator by itself. If only 60 of them pass the competency assessment, the process indicator shows the training needs work. The count alone would have hidden that. Three examples from a child nutrition program:
- Percent of education sessions that use the standard curriculum
- Percent of CHWs who pass the competency assessment
- Average days from screening to referral for children with severe acute malnutrition (SAM, the most serious form of undernutrition)
Collectable within the activity itself. Process data should be captured as the activity happens, not reconstructed later. Training completion from the attendance roster, fidelity from a supervisor observation checklist, adherence from session logs. Reconstruction from memory produces unreliable data.
Designing Outcome Indicators
Good outcome indicators measure the specified change. A defined intention can be an outcome, but it does not establish subsequent behavior. Three design rules.
Match measurement to the intended result. Knowledge, attitudes, confidence, intentions, skills, behavior, and status can all be outcomes when the theory of change specifies them. An intention measure does not establish behavior. Use demonstration for demonstrated skill, and an appropriate self-report instrument for perceived confidence or intention.
Allow enough time for change to emerge. Timing follows what you measure. Immediate learning can be a valid short-term outcome; retained knowledge, applied practice, and durable health or livelihood change need appropriate follow-up. Specify the expected emergence and recall period rather than a universal waiting time.
Tie to the theory of change explicitly. Each outcome indicator should correspond to a specific claim in the theory of change. If the TOC says "trained workers will apply WHO hygiene protocol in home births," the outcome indicator should measure exactly that application, not a proxy that sounds similar.
See SMART indicators deep-dive for the quality gate every outcome indicator should pass.
Sector Examples
Sector Examples
The scenarios below are illustrative. Their figures, costs, and review outcomes are teaching assumptions, not reported studies or evidence of donor approval.
Health: Safe Delivery Program in East Africa
A program trained 420 community health workers in WHO-aligned safe delivery practices. Process indicators: training completion (% completing all 5 days), fidelity (supervisor observation of training sessions), supervision coverage (% trained workers receiving ≥1 supervision visit in 6 months post-training). Outcome indicators: % of trained workers applying WHO hygiene protocol in supervised delivery observation 6 months post-training, % of women delivered by trained worker reporting no complications. Mid-term review found training completion at 91% but protocol application at 38%. Process-outcome pairing revealed that supervision coverage was only 52%, and workers without supervision visits had 18% application rates vs 61% for those with ≥2 visits. Program redesigned supervision schedule for the second cohort.
Education: Girls' Education Program in South Asia
A girls' education program delivered weekly mentoring sessions to 2,400 adolescent girls across 60 schools. Process indicators: session attendance rate, mentor fidelity to session guide (monthly spot-checks), completion of 36-session curriculum. Outcome indicators: % of participating girls still enrolled at end of school year, % of mothers reporting intention to support daughter continuing to secondary school, literacy assessment scores at endline. The process data showed strong attendance (83%) and reasonable fidelity (76%). Endline outcomes showed enrollment retention at 91% (vs 68% control), an observed difference requiring a documented comparison design, baseline/context assessment, and uncertainty before causal interpretation. The balance between process and outcome indicators let the evaluation both confirm effectiveness and diagnose which schools had weakest implementation.
WASH: Community Water Program in West Africa
A rural water program installed 80 water points and trained water committees in operation and maintenance. Process indicators: % of water points installed to engineering standard, % of committees completing 3-day O&M training, % of committees receiving at least one post-training technical visit. Outcome indicators: % of water points functional at 12 months post-installation, average household distance to improved water source, household water storage practices. Endline found 85% functionality (target 90%) and revealed that committees without post-training technical visits had only 56% functionality vs 94% for those with ≥2 visits. The process-outcome linkage drove a redesign of the ongoing mentoring approach for subsequent program phases.
Food Security: Livelihoods Program in Southern Africa
A livelihoods program delivered savings-group formation and agricultural training to 3,200 households across 28 communities. Process indicators: % of savings groups meeting monthly, average meeting attendance, % of trained farmers applying at least 3 of 5 trained practices within one planting season. Outcome indicators: household food security score (HFIAS) at baseline vs endline, household diet diversity, seasonal food gap reduction. The high process compliance (savings group monthly meeting rate 88%) paired with moderate outcome change (HFIAS improvement in 62% of households) raised questions about dose, implementation, contextual shocks, measurement, and the causal pathway for households without improvement. Process compliance alone cannot diagnose the cause. Program commissioned a targeted qualitative study to understand non-responders.
Mistakes & Checklist
Common Mistakes
Mistake 1: Reporting process indicators as if they were outcomes. "500 women trained" is not an outcome. "500 women reporting increased confidence" can be a self-reported confidence outcome if change and measurement are defined; confidence is distinct from intention and from observed behavior. Report each construct accurately.
Mistake 2: Selecting only process indicators for ease. Process indicators are cheaper and easier to collect than outcome indicators. This can bias indicator selection toward what is easy to measure, producing an MEL plan that runs the program well but cannot show whether the program works.
Mistake 3: Selecting only outcome indicators and losing implementation signal. When outcomes fail, you need process indicators to diagnose why. A plan with rich outcome measurement but no process data produces post-mortems with no actionable diagnosis.
Mistake 4: Confusing output indicators with outcome indicators. Outputs count deliverables; outcomes measure change. "Number of handwashing stations built" is an output; "Percentage of households practicing handwashing at critical times" is an outcome. The two are not interchangeable.
Mistake 5: Measuring outcomes too soon. A program that measures "skill application" one week after training is measuring short-term retention, not durable behavior change. Use timing justified by the intended outcome; immediate learning and later behavior answer different questions.
Mistake 6: Not pairing process and outcome indicators. When outcome data comes in disappointing, the absence of paired process data means you cannot diagnose whether it was implementation weakness or intervention weakness. Design them as pairs, not as separate lists.
Mistake 7: Using stated intention as an outcome proxy. "Participants report they plan to use new skill" captures recall and social desirability, not behavior change. Measure intentions if they are the intended outcome, and measure actual behavior separately when that is the question.
Mistake 8: Pretending process data shows effectiveness. A 95% training completion rate does not mean the training worked. Donor reports that highlight process metrics as if they were evidence of effectiveness create credibility problems at evaluation time.
Process vs Outcome Selection Checklist
Run through this for each indicator as you build or review the MEL plan.
For each process indicator:
- Measures implementation quality, not just count of deliverables
- Tied to a documented protocol or delivery standard
- Collectable during or immediately after the activity
- Paired with at least one outcome indicator it is expected to feed
For each outcome indicator:
- Measures the defined outcome without substituting intention for observed behavior
- Has data appropriate to the claim: comparable rounds for change, an endpoint for attainment, and identification evidence for attribution
- Collection timing and recall period fit what you measure, from immediate learning to durable change
- Tied to a specific theory of change claim
- Paired with at least one process indicator that feeds it
For the indicator portfolio:
- Ratio matches program stage (early: more process; mid/endline: more outcome)
- No substitution of process indicators for outcome indicators in donor reporting
- The MEL plan states how process and outcome results are read together
- Mid-term review protocol includes explicit process-outcome diagnostic logic
For the broader indicator design framework, see SMART indicators deep-dive and custom vs standard indicators. For how these indicators fit in a MEL plan, see how to write a MEL plan. For current document workspaces, see the Playbooks catalog.