How our tools reach their conclusions.
These tools use your answers and transparent arithmetic to prompt a discussion. Published frameworks inform the questions; they do not validate our scoring or establish findings about your institution.
What counts as evidence here
Published foundations
Named literature and policy frameworks inform the questions we ask. Their presence is not an endorsement of Onetsa or validation of our tools.
Onetsa editorial choices
Diagnostic weights, radar scenarios and teaming outlines are discussion aids. They are not benchmarks, verified institutional findings or eligibility decisions.
Engagement evidence
Institution specific conclusions need checked data, contextual review and documented limitations. Validation against completed Onetsa engagements remains future work.
Reference titles, attribution and conceptual scope were checked against publisher and official organisation records on 9 November 2026. This is an editorial source review, not independent expert validation, a complete literature review or field calibration. Living standards and applicable national rules need a fresh check before use.
1. What the diagnostic is, and is not
It is a structured reflection on your own practice. Onetsa has assigned the answer weights as editorial judgements, not as validated measurements of institutional performance.
It is not a benchmark, a maturity rating, or a statistical sample. It makes no comparison with other institutions, and the percentages shown are relative to your own highest score.
2. The scoring, in full
Each answer carries whole number weights from 0 to 3 on four dimensions. Answers describing a healthy state carry zero.
Sd = Σ wi,d for each dimension d
The bottleneck is the dimension with the highest total. Bars show each score as a share of that highest score.
Tie rule, upstream first. If two dimensions tie, we name the earlier link in the chain:
Data & digital foundations → MEL & timing → Decision authority → Institutional capability
This is an ordering convention for selecting a starting point, not a mathematical law or a proven causal relationship. Tied areas should be considered together.
3. The calendar latency test
Where timing matters, we use calendar days, not scores:
Blind window = day report is ready − day decision is made
Example: validated data is ready 45 days after quarter end; the budget allocation meeting sits on day 15. The report is ready 30 days after the decision date. This does not establish that other evidence was unavailable. If the result is zero or negative, evidence arrives in time. You supply both dates, we only do the subtraction.
4. Every question, and why we ask it
Q1 · Evidence at the point of decision
When leaders make a major decision, where does the evidence usually come from?
Separates absence of evidence (data), lateness of evidence (MEL), and evidence that exists but is ignored (systems).
- Scattered spreadsheets and reports we struggle to reconcileData & digital foundations +3
- Monitoring reports, but they arrive after the decision is madeMEL & timing +3
- Good data, but it rarely changes what we actually decideDecision authority +3
- Mostly the experience and judgement of senior staffInstitutional capability +2, Data & digital foundations +1
Q2 · Digital Public Infrastructure (DPI)
Your systems can now exchange data through shared digital rails (ID, payments, data exchange). What happens next?
Technical interoperability is not decision interoperability. Tests whether data flows are matched by mandate to act on them.
- We are not yet connected to national digital railsData & digital foundations +3
- Data flows, but nobody has the mandate or sign off to act across agenciesDecision authority +3
- Data flows, but quality and definitions differ between agenciesData & digital foundations +2, MEL & timing +1
- Data flows and there are clear rules on who acts on it0
Q3 · Digital Public Goods (DPGs)
Think of your open source platforms (e.g. DHIS2, OpenMRS, RapidPro). Who keeps them running after donor support ends?
Open licensing does not make implementation or recurrent stewardship free. This question prompts discussion of skills, funding and continuity; it does not assess DPG recognition or local implementation assurance.
- Mostly external partners or a vendorInstitutional capability +3
- A few internal people, if they leave, it stallsInstitutional capability +2, Decision authority +1
- Our own team, but there is no recurrent budget line for itDecision authority +2, Institutional capability +1
- A funded internal unit with a budget line and succession plan0
Q4 · AI readiness
How is your institution approaching AI (dashboards, prediction, chat assistants)?
AI amplifies the quality and governance of its inputs. Tests data integrity and accountability before automation.
- Pilots are planned, but we are unsure our underlying data can be trustedData & digital foundations +3
- Pilots exist, but no one has decided who is accountable for AI informed decisionsDecision authority +3
- Interest is high, but staff lack the skills to judge or use AI outputsInstitutional capability +3
- We are not pursuing AI yet, and that is a deliberate choice0
Q5 · Monitoring, evaluation & learning
How would you describe your monitoring, evaluation and learning?
Distinguishes compliance oriented MEL from learning that feeds management and budget cycles.
- Mainly compliance reporting for fundersMEL & timing +3
- We track indicators, but rarely adapt because of themMEL & timing +2, Decision authority +1
- We have a framework, but staff lack time or skills to use itInstitutional capability +3
- It works and regularly informs decisions0
Q6 · Timing of evidence
Compare when your reports are ready with when budget or planning decisions are made.
Self reported calendar latency: if validated data arrives after the statutory decision window, the decision is structurally blind.
- Reports usually arrive after the decision window has closedMEL & timing +3
- Reports are on time, but the meeting does not use themDecision authority +3
- We do not know, the two calendars have never been comparedMEL & timing +1, Decision authority +2
- Reports are timed to arrive before key decisions0
Q7 · Persistence of problems
Why do your programmes or reforms most often stall?
Recurring problems point to structure (feedback loops, incentives) rather than one off failures.
- We cannot see clearly what is happening on the groundData & digital foundations +3
- Fixes address symptoms; the same problems keep returningDecision authority +3
- We cannot tell which interventions are actually workingMEL & timing +3
- Skills and momentum leave when consultants or key staff leaveInstitutional capability +3
5. 2026 guardrails: DPI, DPGs and AI
DPI
Connected rails are necessary, not sufficient. We ask who holds the mandate to act on shared data, and whether sign off rules allow it.
DPGs
Open licensing does not make implementation, hosting, integration or stewardship free. We ask how recurrent costs and responsibilities will be met; DPG recognition does not assure a local implementation.
AI
AI amplifies its inputs. We check data integrity and name who is accountable for an AI informed decision before any pilot scales.
6. Foundations
- Leverage Points: Places to Intervene in a System
Donella Meadows · 1999
Scope: A conceptual account of intervention points. Onetsa's five category teaching ladder is a simplification, not a validated effectiveness ranking.
- Thinking in Systems: A Primer
Donella Meadows · Chelsea Green · 2008
Scope: Systems concepts including feedback and delays. Does not establish the reporting cycles or causes in Onetsa's sector examples.
- Building State Capability: Evidence, Analysis, Action
Matt Andrews, Lant Pritchett and Michael Woolcock · Oxford University Press · 2017
Scope: Institutional capability, isomorphic mimicry and Problem Driven Iterative Adaptation. Conceptual grounding, not evidence about a named Onetsa client.
- Utilization Focused Evaluation
Michael Quinn Patton · SAGE · Fourth edition, June 2008
Scope: Evaluation designed around intended use by intended users. Does not validate Onetsa's diagnostic weights or the two week exercise.
- Developmental Evaluation: Applying Complexity Concepts to Enhance Innovation and Use
Michael Quinn Patton · Guilford Press · Print publication 2010; ebook 2011
Scope: Evaluation for adaptation, complexity and innovation. The publication year depends on the edition or format cited.
- Universal DPI Safeguards Framework
Multistakeholder framework under UNDP and UN technology office stewardship · 2024 launch; living framework
Scope: Guidance for safe and inclusive digital public infrastructure. The launch was stewarded by UNDP and the Office of the Secretary General's Envoy on Technology, not jointly authored by ITU or GovStack. Verify the applicable version for each engagement.
- Digital Public Goods Standard
Digital Public Goods Alliance · Living standard
Scope: Recognition criteria for a core digital solution, not assurance of a local implementation or a guarantee of cost free deployment. Check the version used before applying it.
- The Open Source Definition
Open Source Initiative · Definition and licensing criteria
Scope: Free redistribution does not mean implementation, hosting, support, integration or ongoing stewardship are free.
- AU Data Policy Framework
African Union · 28 July 2022
Scope: Continental policy direction for data governance. Does not establish a country's applicable law or implementation status.
- Continental Artificial Intelligence Strategy: Harnessing AI for Africa's Development and Prosperity
African Union · July 2024
Scope: Continental strategy, not a national legal opinion or a complete model assurance checklist.
- Everybody's business: strengthening health systems to improve health outcomes
World Health Organization · WHO framework for action, 2007
Scope: A general health systems framework, not evidence of a specific HMIS, stock allocation calendar or country condition.
- PEFA 2016: Framework for assessing public financial management
PEFA Secretariat · 2016 framework, second edition
Scope: Public financial management assessment framework. Onetsa's budget timing scenario is not a PEFA assessment or finding.
7. How we keep it current
- Review before external use: verify the questions and sources against the country, sector and decision in question. Record substantive changes with a new version number.
- Policy anchoring: reviews should check applicable continental and national frameworks (AU data and AI strategies, national DPI and data protection laws) rather than older assumptions about missing technology.
- Field calibration: validation against completed engagements remains future work. The weights are not yet empirically calibrated.
- Stated limits: until that field evidence exists, results are indicative. We say so on every result.
