What to do when there is no baseline: four honest options
Oct 12 2026 · Donor reporting
No baseline is not a dead end. Four honest options, ranked, plus how to report them without inventing a number.
When there is no baseline, you have four honest options, in order of preference: reconstruct one from existing records such as registers, MIS exports or prior reports; run a rapid baseline now and label it with the month you collected it; measure a comparison group alongside your treatment group; or report change from the first measurement onward and say plainly that no baseline exists. Never invent a figure or back-date a survey. Most donors accept 'no baseline; first measurement collected at project start' far more readily than a number with no source. State which option you used, and why, in the reporting notes.
What do you do when there is no baseline?
A baseline is the measurement of an indicator before your intervention starts. When it is missing, you are not stuck, but you do have to choose an approach and be transparent about that choice.
The options below are ranked from strongest evidence to weakest. Take the highest one you can actually deliver with the time and budget available.
- Reconstruct a baseline from existing records: administrative data, registers, management information systems, earlier reports.
- Run a rapid baseline now and date it clearly.
- Measure a comparison group at the same time as your first round of data collection.
- Report change from the first measurement onward and state that no baseline exists.
Which option is best? A quick comparison
| Option | What it involves | Evidence strength | Main risk |
|---|---|---|---|
| Reconstructed baseline | Pull pre-project figures from registers, MIS exports or prior reports | Strong when records are complete and cover the same population | Records may not match your indicator definition |
| Rapid baseline now | Collect data at project start, before activities reach participants | Moderate to strong | Activities may already have begun |
| Comparison group | Measure a similar unexposed group alongside your participants | Moderate | The groups may differ in ways you cannot control |
| First measurement only | Use the first round as the reference point | Weak | You cannot show change from the true starting point |
Can you reconstruct a baseline from existing records?
Often yes. Before you accept that no baseline exists, check what is already there.
- Attendance or service registers from the previous period.
- Routine health, education or agriculture management information systems.
- Prior project reports and endline evaluations covering the same area.
- National or district survey data, if the geography and population match.
- Your own logframe tracking sheet from an earlier phase.
The test is comparability: same population, same indicator definition, same unit of measurement, comparable period. If any of those differ, say so when you report the figure. A tracker such as the logframe template in Google Sheets is a practical place to record both the original figures and their source.
How do you run a rapid baseline now?
- Freeze the indicator definition and unit of measurement first.
- Collect data before activities reach participants, even if the project has formally started.
- Record the exact collection window and describe the sample.
- Keep the raw data so someone else can re-check it.
- Label it 'rapid baseline' or 'first measurement', never plain 'baseline', in every report.
A small, well-documented rapid baseline beats a large, undated one. If you only have a few days, use a short mobile data collection form and keep the sample description precise.
Is a comparison group a substitute for a baseline?
Partly. A comparison group tells you what happens without your intervention, measured at the same time as your treatment group. It supports statements about differences between groups, not change over time within one group.
If you use one, describe how the group was selected and what makes it similar to your participants. Do not present between-group differences as though they were before-and-after change.
What do you do if you genuinely cannot measure anything?
Then say so, and treat the first measurement as your reference point. Write 'to collect' in the baseline cell, enter the actual month when you collect the data, and report change from there.
This is a legitimate position. What is not legitimate is a blank cell filled later with an estimate presented as measured data.
How do you write it up in a donor report?
State the option, the date, the source and the limitation, in that order, in a short note beside the indicator table.
- No baseline available; figures reconstructed from district service registers covering the previous year.
- No baseline; first measurement collected at project start using a rapid participant survey.
- No baseline; change reported from the first measurement onward.
- Baseline left as 'to collect' because the activity had already begun when the indicator was agreed.
'No baseline; first measurement collected at project start.' A short honest note costs you very little. A fabricated figure can cost you the contract.
Checklist: what to do when there is no baseline
- Confirm the indicator definition, unit and disaggregation.
- Search existing records before assuming nothing exists.
- Check comparability: population, definition, period.
- Choose the strongest option you can realistically deliver.
- Date every measurement and keep the raw data.
- Label it clearly as reconstructed, rapid or first measurement.
- Write the limitation in the report in one or two sentences.
- Agree the approach with your M&E lead or donor before reporting.
What should never appear in a baseline cell?
- A number with no stated source.
- A collection date you did not actually use.
- Targets, planned figures or expected values.
- An estimate labelled as measured data.
If you are under pressure to fill the box, escalate it. The conversation is uncomfortable for an afternoon; the alternative can affect future funding. If you need a written plan that holds up to scrutiny, the evaluation plan builder and the indicator builder both walk you through the wording.
Written for M&E professionals. Not a guarantee of accuracy — always verify every claim against its source, and let a named human sign off. Get the free handbook.