Lesson 5 of 6
Measure Change: Define One Outcome You Can Defend
To measure change, define one outcome precisely: who is eligible, what counts, when you observe it, the numerator and denominator, the source, the owner and the known limitation. Report the result with its coverage, keep missing answers separate from “no”, and don’t claim the program caused the change unless you have a baseline or comparison that supports it.
You will make: One outcome measure with its population, timing and limitation.
Builds on: The context from lesson 3 and the review cycle from lesson 4.
THE METHOD AT A GLANCE
- Separate delivery, outcome and cause.
- Write the measure: population, observation, window, numerator, denominator, source, owner, limitation.
- Report the result with its coverage.
- Add a baseline if you want to claim improvement.
- Use open-ended answers to explain, not to count.
What is the difference between delivery, outcome and cause?
In short: Delivery is what you did. An outcome is what changed for people. Cause is proof that you made it change, and it needs the most evidence.
DELIVERY
40 people completed training
What your team controls. Easy to count.
OUTCOME
15 of 25 respondents used the skill
A change you observe in people, at a set time.
CAUSE
The training led to that use
Needs a baseline, a comparison or strong supporting evidence.
The same split applies elsewhere: a closed ticket is delivery, a customer saying the issue is resolved is an outcome. An award is delivery; what the grantee achieves with it is an outcome.
How do you write one outcome measure?
In short: Fill in eight fields. If two people would compute a different number from your definition, it isn’t finished.
OUTCOME MEASURE · TRAINING EXAMPLE (ILLUSTRATIVE)
How do you report the number honestly?
In short: Always show three numbers together: the result, the response coverage and the unknowns.
All 40 completers at the 30-day follow-up (illustrative).
SAY
“15 of 25 respondents (60%) used the skill within 30 days. 25 of 40 completers responded (62.5%); outcomes are unknown for 15.”
DON’T SAY
“60% of learners use the skill.” That treats non-respondents as if they answered, and implies lasting use.
Keep “no reply” as its own value, never as “no”. And a single checkpoint shows use at day 30, not continued use or improvement.
Do you need a baseline?
In short: Only if you want to claim improvement. Measure the same thing at intake and at follow-up for the same people, linked by their ID.
Maria rated her confidence 2 out of 5 at intake and 4 out of 5 at exit. That’s change for one person. For a group, compare only people with both waves, and say how many that is. To claim the program caused the change, you also need a comparison, such as people who haven’t started yet, or strong supporting evidence like coach observations.
How should you use open-ended answers?
In short: Use them to explain the numbers and to find what to investigate, not to estimate how common something is.
If 12 of 18 interviewees mention scheduling problems, that’s a strong lead for your next review. It doesn’t tell you what share of all 40 completers faced the same problem. Keep the original words, and put your theme in a separate field.
Where does Sopact Sense fit?
In short: In Sopact Sense, you define a measure once. It’s calculated as responses arrive, with coverage and unknowns shown next to the result, and baselines linked through each person’s ID. The AI Assistant explains the result using open-ended answers and cites the responses behind every figure. Your team approves the definition and the claim.
Add this to your plan
Open your working evidence plan ↗
Write one outcome measure for your workflow using the eight fields above. Then write the sentence you would report, including coverage and unknowns.
Check your reasoning
A strong measure keeps the 40 eligible, the 25 respondents and the 30-day timing together; treats non-response as unknown; and makes no causal claim without a baseline or comparison. If a colleague would calculate a different number from your definition, tighten it.
Frequently asked questions
What is the difference between an output and an outcome?
An output (or delivery) is what your team did, such as sessions run or people trained. An outcome is a change observed in people or organizations afterwards, such as using a new skill or resolving an issue. Report both, and never relabel an output as an outcome.
What response rate is good enough?
There’s no single threshold. Report coverage every time, and check who is missing. If the people who didn’t respond differ from those who did, for example by site or attendance, the result may not represent everyone.
Do we need a control group?
Not to report an outcome. You need a baseline to claim improvement, and a comparison group or strong supporting evidence to claim the program caused it. Say clearly which of these your evidence supports.
How should we handle missing answers?
Record them as “no reply” or “unknown”, never as “no”. Report how many there are next to the result, and try to follow up with a sample of non-respondents if the decision depends on it.
Can AI calculate the measure for us?
Yes, if the definition is written down and the records are linked. Check the first calculation by hand: recompute the numerator and denominator, and confirm the time window and population match your definition.
When you need more depth: frameworks and valuation
To explain why your work should lead to change, or to plan delivery against outcomes, use the Theory of Change and logic model methods in the Measurement & Reporting course.
Build a Theory of Change →