Capability template for analytics and reporting
Program outcome map
Before an evaluation, the program and evaluation leads get one map linking each activity to the outcome it is meant to serve, with proposed measures, data sources, assumptions and the questions the evidence cannot yet answer. It claims no impact.
Listed under
Add it to a workflow in the Fully Powered Planner, beside the teams and roles that will run it. Book a fit call
The outcome
Everyone preparing for the evaluation looks at the same page. Each activity sits beside the outcome it is meant to serve, with the measures you already collect, the measures worth adding, the data each would come from and the assumption that has to hold for the activity to lead to the outcome.
The map is honest about what a count can show. Sessions delivered and people reached say the program ran; they do not say anyone changed. The page keeps those counts apart from measures of a condition, labels what each link actually has behind it, and turns every gap into a numbered question: what the figures cannot tell you about a starting point, about other things happening at the same time, about who was not heard and about effects nobody planned for.
The leads decide. They approve or change each proposed measure, confirm or drop each link the assistant had to propose, and choose which further evidence to collect. The assistant draws the map and claims no impact, decides no one's eligibility for help and works only from de-identified or supplied program information.
Who starts the work, and when
The program lead or the evaluation lead starts it when an evaluation, a funder review or a program redesign is coming and the team wants to agree what is being tested before anyone collects more data. Run it again when an outcome changes, a new measure is added or fresh data arrives.
What they supply
- The activities: what the program does, how much of it, how many people it reaches and over what period, with any outcome the plan already names for it.
- The intended outcomes: each in the plan's own words, for whom, by when, and where the plan states it.
- The measures you already have: what each counts, where the data comes from, how often it is collected and its latest figure.
- Any starting figures, where a baseline was taken.
- Participant feedback that participants have approved for use, without names or contact details.
- Your settings: the program name, the evaluation and its date, the leads, the data period and anything else happening at the same time that might move the outcomes.
What comes out
One page in a fixed order: the totals; the map, a block for each link from an activity to an outcome; activities with no outcome and outcomes with no activity; the counts that are not outcome evidence; the proposed measures; the numbered questions the evidence cannot answer; the decisions for the leads; and a check that every activity, outcome, measure and feedback item is accounted for.
An invented example
Illustrative example with invented data. Kestrel Lane Reading Project is an invented after-school reading program. Its staff, partner schools, figures and feedback are invented, and so are all names. First come the inputs the program lead supplied, in full, then the map the capability returned.
The inputs
ACTIVITIES (export, run Tuesday 10 November 2026)
Activity ID | Activity | Dose | Participants reached | Period | Stated outcomes | Stated assumptions | Notes
A1 | Weekly small-group reading | 60 minutes weekly, 28 weeks planned | 62 children | 1 Sep 2025 to 30 Jun 2026 | O1, O2 | Children attend regularly | Groups of 4 to 5 children with a volunteer tutor
A2 | Volunteer tutor training | 3 sessions before each term | 18 volunteers | 1 Sep 2025 to 30 Jun 2026 | | | Covers how to run a session
A3 | Family reading evenings | 4 evenings | 41 families | 1 Sep 2025 to 30 Jun 2026 | O3 | | Families try reading activities together at the school
A4 | Book giveaway | Once a term, 3 times | 62 children | 1 Sep 2025 to 30 Jun 2026 | | | Books go home to keep, for families to read together
INTENDED OUTCOMES
Outcome ID | Intended outcome | For whom | Timeframe | Stated in
O1 | Children read more fluently | Children in the program | 30 Jun 2026 | Program plan, section 2
O2 | Children feel more confident reading aloud | Children in the program | 30 Jun 2026 | Program plan, section 2
O3 | Families read together at home more often | Families of children in the program | 31 Dec 2026 | Program plan, section 2
O4 | Children attend school more regularly | Children in the program | 31 Dec 2026 | Funding application
MEASURES
Measure ID | Measure | Activity ID | Outcome ID | Data source | How often | Latest value | Latest date | Number covered
M1 | Reading fluency, median words read correctly per minute | A1 | O1 | Tutor check sheets | Termly | 52 | 15 Jun 2026 | 55 children
M2 | Sessions delivered | A1 | | Session log | Weekly | 27 of 28 planned | 30 Jun 2026 |
M3 | Children attending 20 or more sessions | A1 | | Session log | Termly | 38 | 30 Jun 2026 | 62 children
M4 | "I feel confident reading aloud", four-point answer | A1 | O2 | Child feedback form | Yearly | 44 chose agree or strongly agree | 18 Jun 2026 | 51 children
M5 | Families attending at least one family evening | A3 | O3 | Sign-in sheets | Per event | 41 | 30 Jun 2026 |
BASELINES
Measure ID | Baseline value | Date | Population
M1 | 41 | 22 Sep 2025 | 60 children assessed
FEEDBACK (de-identified; Approved for use is the participant's or parent's own approval)
Item ID | Activity ID | Summary | Approved for use
F1 | A3 | Several parents said the evenings gave them ideas for reading at home | Yes
F2 | A1 | Children said they liked choosing their own book | Yes
F3 | A1 | A comment about a tutor | No
F4 | A3 | Some parents could not attend evenings because of work hours | Yes
SETTINGS
Program: Kestrel Lane Reading Project. Evaluation: external evaluation, fieldwork from 1 Feb 2027.
Leads: program lead and evaluation lead. Data period: 1 Sep 2025 to 30 Jun 2026.
Other influences: I1, a daily 15-minute reading period at both partner schools from Jan 2026 (may touch O1, O2). I2, a free family reading hour at the town library from Mar 2026 (may touch O3).
The map
PROGRAM OUTCOME MAP: Kestrel Lane Reading Project
For: Program lead and evaluation lead, before the external evaluation fieldwork from 1 Feb 2027. Data period 1 Sep 2025 to 30 Jun 2026.
Prepared: 10 Nov 2026, from 4 activities, 4 outcomes, 5 measures, 4 feedback items, 1 baseline row, settings supplied.
Totals: Activities 4. Outcomes 4. Links 4 (3 stated in the plan, 1 proposed). Measures supplied 5. Measures proposed 3. Questions 13.
THE MAP
Link 1: A1 Weekly small-group reading → O1 Children read more fluently (stated in the plan)
Reach: 62 children; 60 minutes weekly, 28 weeks planned; 1 Sep 2025 to 30 Jun 2026. Intended by: 30 Jun 2026.
Condition measures: M1 Reading fluency, median words read correctly per minute: baseline 41 on 22 Sep 2025 (60 children); latest 52 on 15 Jun 2026 (55 children); difference 11, arithmetic on the supplied figures only. Source: tutor check sheets, termly.
Counts on this link: M2, M3.
Comments: F2 (A1).
Measures proposed: none.
Evidence status: Condition measured at two points.
Assumptions: Children attend regularly (from the plan). Regular practice with a tutor in a small group carries over to how fluently a child reads on their own (assistant's reading, leads to confirm).
Questions: 1, 2, 3.
Link 2: A1 Weekly small-group reading → O2 Children feel more confident reading aloud (stated in the plan)
Reach: 62 children; 60 minutes weekly, 28 weeks planned; 1 Sep 2025 to 30 Jun 2026. Intended by: 30 Jun 2026.
Condition measures: M4 "I feel confident reading aloud", four-point answer: 44 of 51 children chose agree or strongly agree on 18 Jun 2026. No baseline supplied. Source: child feedback form, yearly.
Counts on this link: M2, M3.
Comments: F2 (A1).
Measures proposed: P1.
Evidence status: Condition measured once.
Assumptions: Children attend regularly (from the plan). A child's answer on the form reflects how confident they are reading aloud in front of others (assistant's reading, leads to confirm).
Questions: 4, 5, 6 (3 asked at Link 1).
Link 3: A3 Family reading evenings → O3 Families read together at home more often (stated in the plan)
Reach: 41 families; 4 evenings; 1 Sep 2025 to 30 Jun 2026. Intended by: 31 Dec 2026.
Condition measures: none.
Counts on this link: M5.
Comments: F1, F4 (A3).
Measures proposed: P2.
Evidence status: Comments only.
Assumptions: Families who attend an evening can fit shared reading into their week at home (assistant's reading, leads to confirm).
Questions: 7, 8, 9, 10, 11.
Link 4: A4 Book giveaway → O3 Families read together at home more often (proposed, leads to confirm)
Reach: 62 children; once a term, 3 times; 1 Sep 2025 to 30 Jun 2026. Intended by: 31 Dec 2026.
Condition measures: none.
Counts on this link: none.
Comments: none approved for A4.
Measures proposed: P2, P3.
Evidence status: Nothing supplied.
Assumptions: The books are ones a family can read together at home (assistant's reading, leads to confirm).
Questions: 12, 13 (10 asked at Link 3).
ACTIVITIES WITH NO OUTCOME
A2 Volunteer tutor training, 18 volunteers. The plan states no outcome for it, and no outcome names the same group of people or the same place and the same area of change. See decision 3.
OUTCOMES WITH NO ACTIVITY
O4 Children attend school more regularly (funding application). No activity states it, and none names the same group and area of change. See decision 4.
COUNTS THAT ARE NOT OUTCOME EVIDENCE
M2 Sessions delivered, 27 of 28 planned (activity count). It shows the sessions ran, not that any child changed.
M3 Children attending 20 or more sessions, 38 of 62 (participation count). It shows who took part, not what changed for them.
M5 Families attending at least one evening, 41 (participation count). It shows who came, not what they did at home afterwards.
UNPLACED MEASURES
None.
PROPOSED MEASURES
P1, for O2 (Link 2). Children choosing agree or strongly agree on "I feel confident reading aloud", asked of every child at the start of the year and again in June on the same four-point form; count of children who answered both times. Answered by children, collected by whoever the leads name. Data source: the child feedback form (M4), with a new start-of-year round. Proposed; the leads approve.
P2, for O3 (Links 3 and 4). Number of days in a week a family reads together, asked on a short parent form when families join and again at the end of the year; families counted by answer. Answered by parents. Data source: new collection (a parent form). Proposed; the leads approve.
P3, for O3 and A4 (Link 4). Families who say they read at least one giveaway book together in the past month, asked on the same parent form. Answered by parents. Data source: new collection (the same parent form). Proposed; the leads approve.
No target or expected value is proposed for any measure.
QUESTIONS THE EVIDENCE CANNOT ANSWER
Q1 (same people; Link 1). M1 covers 60 of the 62 children reached at baseline and 55 in June. Are the 55 among the 60, and why were 5 not assessed in June? Who can answer: program lead. Evidence that would help: any record that lets the two rounds of check results be compared for the same children, without names.
Q2 (other influences; Link 1). No comparison group was supplied, and I1, a daily reading period at both partner schools from Jan 2026, may touch O1. How much of the difference of 11 could the reading period account for? Who can answer: evaluation lead. Evidence that would help: the termly check results from before and after Jan 2026.
Q3 (unintended effects; A1). Did the small-group sessions have effects the plan did not intend, helpful or not, on the children, their classes or the tutors? Who can answer: evaluation lead. Evidence that would help: open questions put to children, tutors and teachers.
Q4 (starting point; Link 2). M4 has no baseline. How confident were children before the sessions began? Who can answer: program lead. Evidence that would help: any earlier answers to the same question; P1 would give a starting point only for the next group.
Q5 (who was heard; Link 2). M4 covers 51 of the 62 children reached, and one approved comment (F2) is on file. What did the 11 children who did not answer experience? Who can answer: program lead. Evidence that would help: the session log's attendance, if it records who was absent.
Q6 (other influences; Link 2). No comparison group was supplied, and I1 may touch O2. How much of the answer on M4 could the reading period account for? Who can answer: evaluation lead. Evidence that would help: the same form asked of similar children outside the program.
Q7 (starting point; Link 3). No measure shows how often families read together before the evenings. What was it? Who can answer: program lead. Evidence that would help: any earlier parent answers; otherwise P2 from the next intake.
Q8 (who was heard; Link 3). Two approved comments (F1, F4) sit beside 41 attending families, and F4 says some families could not come because of work hours. What do the families who did not attend, and the attenders who did not comment, say? Who can answer: program lead. Evidence that would help: a short call or form for non-attenders, with approval to use the answers.
Q9 (other influences; Link 3). No comparison group was supplied, and I2, a free family reading hour at the town library from Mar 2026, may touch O3. How much of any home reading could it account for? Who can answer: evaluation lead. Evidence that would help: parent answers on whether they used the library hour.
Q10 (timing; O3, Links 3 and 4). O3 is intended by 31 Dec 2026 and the data period ends 30 Jun 2026. What has happened since June, and what will be known by the fieldwork? Who can answer: program lead. Evidence that would help: parent answers gathered since June.
Q11 (unintended effects; A3). Did the family evenings have effects the plan did not intend, helpful or not? Who can answer: evaluation lead. Evidence that would help: open questions put to parents and school staff.
Q12 (what would show it; Link 4). Nothing supplied relates to the book giveaway, and the link is proposed. If it stands, what would show whether the books reach shared reading at home? Who can answer: evaluation lead. Evidence that would help: P3.
Q13 (unintended effects; A4). Did the giveaway have effects the plan did not intend, helpful or not? Who can answer: evaluation lead. Evidence that would help: open questions put to parents and children.
DECISIONS FOR THE PROGRAM LEAD AND EVALUATION LEAD: 5
1. Link 4 (A4 → O3) is proposed from the wording of the notes. Confirm it, change it or drop it.
2. Proposed measures P1, P2 and P3. Approve, change or drop each.
3. A2 Volunteer tutor training has no outcome. State the outcome it serves, or record that it serves none in this evaluation.
4. O4 Children attend school more regularly has no activity. Name the activity meant to lead to it, or leave it off this map.
5. Questions 1 to 13. Choose which further evidence to collect, and who collects it.
Nothing is recommended.
CHECK
Activities 4 = linked 3 (A1, A3, A4) + no outcome 1 (A2). Outcomes 4 = linked 3 (O1, O2, O3) + no activity 1 (O4). Links 4 = stated 3 + proposed 1.
Measures 5 = condition measures 2 (M1, M4) + counts 3 (M2, M3, M5) + unplaced 0. Placed: M1 on Link 1; M2 and M3 on Links 1 and 2; M4 on Link 2; M5 on Link 3.
Feedback items 4 = approved 3 (F1, F2, F4) + not approved 1 (F3, not used).
Questions 13 = Link 1 three + Link 2 three + Link 3 five + Link 4 two.
Every figure traces to the files supplied. The map claims no impact, decides no one's eligibility for help, changes no record and sends nothing.
Where it fits
- Analytics and reporting, which owns the measures and the reporting behind an evaluation.
- Project and delivery, because the activities are the program's delivery.
- Nonprofits, where funders ask what changed as well as what was done.
- Education, for tutoring, outreach and family programs run by schools and foundations.
- Government, for public programs reviewed against their stated aims.
- Established businesses, for organizations with a program in place and an evaluation due.
- Growth & scaling, when a program is about to expand and the case for it needs to be clear.
It is not analysis plan, which plans the questions to put to a particular dataset, and it is not operations KPI scorecard, which reports a fixed set of measures month by month. Assumption map ranks the untested beliefs behind a product plan; this one lays out the activity-to-outcome links of a program.
What has to change for your company
Decide who the two leads are and who answers each kind of question. Use your own names for activities, outcomes and measures, and keep their identifiers stable from one run to the next. List the other influences you want asked about, such as a policy change or a partner's new service, each with the outcomes it might touch. Agree what counts as approved feedback and who confirms the approval. Keep the file names and column headings your team already uses, or map them once.
What has been checked
Fully Powered AI wrote this capability template. It was editorially reviewed on 10 November 2026. It was tried once on its invented example in Claude on 10 November 2026, and the record is published with it. Edits made to the template and the example after that run, including a sharper rule for proposing a link, were checked by hand and not run again from the start; the record lists them. It has not yet been tested in a client environment.
Testing in your environment means running it on a past program whose evaluation you already know, with cases written in advance: an activity with no stated outcome, an outcome with no activity, a measure that only counts sessions, feedback not approved for use. The evaluation lead checks every figure against the files supplied before anyone relies on the map.
The full capability template
Everything above is the outline. The complete capability template is the working version: the roles, the inputs, the procedure, the exceptions and stop conditions, the output format, the reviewer checklist, the acceptance tests, the instructions for an AI assistant and what to customize for your company. It comes with a download and the link to its folder in Fully Powered’s public repository.
Free with an account. Sign in by email: no password and no card.
Next step
Put this capability to work for your team: on your plan, or built with you in a sprint.
Build this in a sprint
In a sprint, Fully Powered adapts this capability template to your rules, proves it on your own work and puts it in the hands of the roles that own it.
Two ways to begin: