Services
Data systems and M&E design
We help you trust your own data. We assess the data you already collect, design what you should measure, and train your teams to use it.
Questions this answers
- Can we rely on the data behind our plans?
- What should we measure, and how often?
- Do our teams use the data they collect?
The kinds of data systems and M&E work we run
Most engagements combine two or three of these.
Monitoring and evaluation frameworks
The overall structure that says what a programme will track, how, and how often, agreed before implementation starts.
Results framework and indicator design
What each indicator actually measures, how it is calculated, and how often it can realistically be collected.
Routine data quality assessments
Checking the data your own systems already produce, such as health and school data systems, against the records behind it.
Data use studies
Whether the data being collected is actually used to make decisions, or filed and forgotten.
Gender and inclusion data audits
Whether your data can be split by sex, age, disability and location at all, and where it cannot.
Training in research methods and data use
Building your own team’s ability to collect, check and use data, not only handing over a report.
How we assess and design data systems
A data system fails for one of three kinds of reasons. We check all three, in every assessment.
The technical layer.
The forms, tools and systems themselves — are they too complex, error-prone, or simply not built for what you need to know?
The behavioural layer.
Do the people entering and using data have the skills, and the understanding of why it matters, to do it well?
The organisational layer.
Does the wider organisation support good data with training, feedback and the time to check it, or does it only ask for reports?
Fix-first, not just diagnose.
Every assessment ends with what to fix first, not only a list of what is wrong.
The three things that determine whether a data system works
Adapted from PRISM (Performance of Routine Information System Management), set out in 2009 by Aqil, Lippeveld and Hozumi and developed into a set of tools by MEASURE Evaluation, a USAID-funded project, to assess routine health information systems.
Technical determinants
The data collection forms, the systems, the processes — whether the tools themselves make good data easy or hard to produce.
Behavioural determinants
The knowledge, skills, values and motivation of the people who actually collect and use the data.
Organisational determinants
The culture, structure, resourcing and roles that decide whether good data collection is supported or merely demanded.
What a data systems assessment can find
Published studyOne ministry of health was concerned that district and facility staff rarely used their own routine data to spot gaps, plan or track progress. A PRISM assessment found very high data errors, caused by data collection forms that were too complex, inaccurate transfer from patient records, and calculation errors. The response was not more pressure to use the data. It was simpler forms, refresher training in collecting and processing data, and regular meetings to share performance results.
From our own work
Real work of ours that used this kind of research, not an invented example.
Bauchi State Ministry of Budget and Planning
Real exampleBuilding M&E capacity for Bauchi State’s social protection policy
44 Bauchi State officials. Implementation of the Bauchi State Social Protection Policy; training held in Jos, Plateau State, in 2023.
Reading a fix-first list
Illustrative| Problem found | Layer | Fix |
|---|---|---|
| Facility form asks for the same figure in two places, entered differently each time | Technical | Redesign the form to ask for it once |
| Data clerks were never told why the monthly report matters | Behavioural | A short briefing on how the data is used, not just how to fill the form |
| No one reviews the data before it is submitted upward | Organisational | A named supervisor sign-off step, before submission |
Three different kinds of problem need three different kinds of fix. A list that only says “improve data quality” could not tell you which of these three to do first.
Questions worth asking a data systems and M&E provider
We built five questions from the PRISM framework’s three determinants, and we answer all five in every proposal.
Does the assessment cover the forms and systems, or only the numbers they produce?
Both. A number can look wrong for reasons that have nothing to do with the number itself.
Does it look at whether people understand why the data matters, not just whether they collect it?
Yes. Motivation and understanding are checked, not assumed.
Does it look at whether the organisation supports good data, or only asks for it?
Yes, including whether anyone reviews data before it is used.
Will the recommendations say what to fix first?
Every assessment ends with a fix-first list, not only a diagnosis.
Can our data be split by sex, age, disability and location?
Checked explicitly, and stated plainly where it cannot yet.
What you receive
An assessment with a fix-first list, a framework your team can run, and training.
The assessment
Findings across the technical, behavioural and organisational layers, not just a data quality score.
The fix-first list
What to fix first, and why, in an order your team can act on.
The framework and training
A framework built for your team to run, and training so they can.
Questions people ask about data systems and M&E specifically
What’s the difference between an M&E framework and a results framework?
An M&E framework covers the whole system: what will be tracked, how, and how often. A results framework, or indicator design, defines each individual measure inside it — what it counts, and how it is calculated.
Can you assess a system we already use, like DHIS2 or a school data platform?
Yes. We assess the system as it is actually used in practice, not only how it is designed to work.
Why don’t our teams use the data they already collect?
Often because the data was designed to be reported upward, not to answer a question the team itself has. A data use study checks which is true before recommending a fix.
How is training different from a one-off workshop?
Training is built around your actual data and your actual team’s gaps, found in the assessment, not a generic curriculum delivered once and left behind.
Not sure your own data can be trusted?
Tell us what the data is meant to tell you, and what has made you doubt it. We will say what an assessment would look for, and what it would take.