Avoid the most common M&E system failures with this pre-build checklist from 10 years of implementations.
In ten years of building M&E systems for NGOs, we have seen the same failures repeat themselves: a system nobody uses, a dashboard that does not match the logframe, a data collection tool abandoned after month one. Most of these failures were preventable — they stem from starting the technical build before answering a set of fundamental questions.
Programme Design Questions
1. Do you have a finalised logframe or results framework? If not, stop. Build the logframe first. An M&E system is an instrument for measuring logframe indicators. Without a finalized logframe, you will build a system that measures the wrong things.
2. Are your indicators SMART? 'Improved health outcomes' is not an indicator. 'Percentage of ANC1 visits completed within the first trimester, measured monthly at health facility level' is an indicator. Each measurable indicator needs a clear numerator, denominator and disaggregation.
3. Who owns indicator calculation? Will calculations be automatic (built into the system) or manual (someone computes them in Excel and enters the result)? Manual calculation invites error. Agreement on this before build saves months of argument after.
Data Collection Questions
4. Who are your data collectors? Field officers with smartphones? Clinic nurses at a desktop computer? Community health workers with feature phones? The answer determines your data collection technology.
5. What is the internet connectivity like in your programme areas? Test it before you design. 'Low connectivity' in Nairobi means slower 4G. 'Low connectivity' in northern Kenya may mean GPRS or no data at all.
6. What data protection requirements apply? Kenya's Data Protection Act 2019 requires a Data Protection Impact Assessment for processing sensitive categories of data including health records, GPS location, and national ID numbers.
7. Who validates data quality at the point of collection? Field officers make errors. Forms need validation rules that catch impossible values, logical contradictions and missing required fields.
Reporting Questions
8. What exactly does your donor want to see? Get a copy of your most recent donor report. Every table and figure in that report must map to a field or calculation in your system. If it cannot, your system will not produce the report without manual intervention.
9. How often do you need to report at each level? Field officer → supervisor (weekly?), supervisor → programme manager (monthly?), programme manager → donor (quarterly?). Each level may need a different view.
10. Do you need real-time visibility or scheduled reports? Real-time requires infrastructure (streaming pipelines, live dashboards). Scheduled reports (daily ETL, email delivery) are cheaper and simpler. For most NGOs, daily refresh is sufficient.
Technical Questions
11. Where will data be stored? Cloud (which cloud? which region? what are the data residency implications?) or on-premises? On-premises reduces connectivity dependency but requires local IT capacity for maintenance.
12. What are your data retention requirements? Most donors require programme data to be retained for 5–10 years after project close. Object storage (S3, MinIO) is the right answer for long-term retention at acceptable cost.
13. Who maintains the system after go-live? This question kills more M&E systems than any technical failure. If the consultant who built it leaves and the NGO has no internal capacity, you will have an abandoned system within 18 months.
Change Management Questions
14. How will you onboard field officers? Plan for a 2-day training, not a 2-hour overview. Include hands-on practice with real forms, troubleshooting common errors, and a clear escalation path when something goes wrong in the field.
15. What is the plan when data looks wrong? Every M&E system eventually produces a number that does not look right. Who investigates? What is the process for identifying whether it is a data quality issue, a calculation error, or a programme reality?