Why is a submitted figure not yet a result?
Field staff send in numbers all the time: people reached, items given, sessions held. Most are right. Some are wrong because of a typing slip, a different definition, a double entry, or a site that counted the wrong period.
If every submitted number goes straight into a report, errors become results. An approval step puts a pause between “someone sent this” and “we stand behind this”.
What does an approval step do?
It does three jobs.
- Catches errors. A second person looks at the number with fresh eyes.
- Creates a decision record. You can show who accepted the figure, when, and on what basis.
- Separates roles. The person who collects the data is not the person who accepts it, which protects both of them.
How should the process work?
- Collect. A field officer or partner submits the value, with a note and the period it covers.
- Review. The reviewer looks at the figure next to the current recorded value and the source evidence, such as an attendance sheet.
- Decide. The reviewer approves or rejects. Either way, they add a note.
- Accept. An approved figure becomes part of the result. A rejected one does not.
- Report. Reports use approved figures only.
Our article on field data quality checks lists the checks a reviewer can use at step 2.
What should the reviewer look at?
- Is the figure complete for the period and site?
- Does it follow the indicator definition?
- Is it plausible next to previous periods?
- Does the evidence support it?
- Could any people be counted more than once? See reach totals and double counting.
A reviewer who cannot answer these questions should ask, not guess.
What should you record about each decision?
Keep a simple record for each decision:
| Item | Why |
|---|---|
| The submitted figure | The original claim |
| Submitter and date | Who said it and when |
| Reviewer | Who decided |
| Decision and date | Approved or rejected |
| Note | The reason, always for a rejection |
When an auditor or donor asks how a figure was accepted, you can show the record in a minute.
What if the reviewer is also the submitter?
Do not allow it. Self-approval removes the point of review. In a small team with few staff, ask a colleague from another project, or the project manager, to act as reviewer. Write the arrangement down.
Also decide what happens when a reviewer is away: name a deputy, for a set period, and record that the deputy acted.
A short example shows why the step matters. A site officer submits “900 kits distributed” for the quarter. The reviewer, comparing with the previous figures and the distribution lists, sees that the site received only 600 kits from the warehouse. They reject the figure with a note: “Warehouse shows 600 sent. Please check the distribution lists.”
The officer checks and finds that a second round of 300 was counted twice. They submit 600. The reviewer approves it.
Without the step, the report would have shown 900 kits, more than the project owned. A donor who later checked stock would have questioned every other number.
Approval does not need to be heavy. A reviewer who knows the project and spends a few minutes per figure can catch most such errors. The key points are that the reviewer is independent, the decision is recorded, and nothing counts until it is approved.
Lastly, set simple service targets for the review itself. For example, a reviewer looks at each submission within three working days, and a rejection always includes a reason the submitter can act on. Track how many figures are rejected and why. A high rate of rejection for the same cause, such as a unit mistake, shows a training need or an unclear indicator definition. Use that information to improve the form, the reference sheet and the guidance for field teams, so fewer errors reach the reviewer at all.
Where EMPOBase fits
In EMPOBase, a collection link lets a field team or partner submit a value for a project indicator and period. An authorised project reviewer reads it beside the current actual and approves or rejects it, with a note. Only approved submissions update the indicator actual, and pending or rejected ones are not counted. The review decision records who acted, when and any note. Reviewers can approve or reject only, so a rejected figure has to be submitted again. Field reports are a separate narrative record and do not write indicator actuals. See the field-to-results workflow.