By Sopact · Updated September 12, 2026

Track investees against an impact agreement by comparing reviewed results with the target, definition and date agreed for each measure. Calculate the difference only when the observations are comparable. Keep performance status separate from evidence status, explain the reason for a material variance and assign the next action. Preserve the original agreement when targets or definitions change.

This lesson is for portfolio and impact teams reviewing investee or grantee progress. Bring the agreement, dictionary and evidence review brief from the related lesson. You will build a variance table and a short action record. This is an operational monitoring method, not a judgment about an investment’s financial value.

1. Identify the agreed measure, target and observation date.
2. Confirm that the reviewed result uses the same definition and scope.
3. Calculate the appropriate difference.
4. Apply a documented review rule, keeping uncertainty visible.
5. Investigate the explanation and agree an action.
6. Retain the decision and review it at the next checkpoint.

What belongs in the impact agreement?

A target is usable only when its meaning is clear. Record the population, unit, baseline if relevant, observation window, direction of improvement and evidence expected. A statement such as “improve employment” is an aim; it is not enough to calculate progress consistently.

Principle 6 of the Operating Principles for Impact Management connects monitoring against expected impact with an appropriate response. The worksheet in this lesson is a practical implementation exercise, not an official standard or compliance determination.

Check comparability before calculating variance

Compare like with like: the same unit, population, definition and period. If any of those changed, record the difference before presenting a performance label. A figure can be accurate in its own scope and still be unsuitable for the proposed comparison.

• Is the target quarterly, annual or cumulative?
• Does the result describe the target population or only respondents?
• Is the measure a count, rate, percentage-point change or milestone?
• Are higher values better, lower values better or a range acceptable?
• Is the observation due, or does it require longer follow-up?
• Has the source been reviewed, and what uncertainty remains?

Do not compare a quarter’s result with an annual target as though the annual deadline had passed. Dividing an annual target evenly into four quarters is appropriate only if that pacing assumption was agreed and fits the work. Seasonal activity and delayed outcomes often require a different expected path.

Use the right calculation for the measure

For a count with a nonzero target, show both the absolute difference and, when useful, the relative difference. For rates, distinguish percentage points from percentage change. For a milestone, compare completion criteria and due date rather than inventing a numeric score.

A zero target makes the relative-difference formula undefined. Use the absolute difference and the agreed review rule. Missing observations cannot be substituted for zero. Where estimates or sampling uncertainty matter, show those limitations instead of assigning more precision than the evidence supports.

Practice: separate performance from evidence

Self-contained fictional example using Partner A’s employment update. The partner confirms 35 starts against the agreed target of 40. Five future offers are excluded. This supports a five-start shortfall, or 12.5% below target, for that specific measure and quarter.

The same package reports that 14 of 20 follow-up respondents were employed at the three-month check. It labels this “70% sustained employment,” although interval evidence has not been supplied. The agreement expects 80%, but the eligible cohort and response coverage have not yet been confirmed. Do not label this a verified ten-point performance shortfall. First establish whether the observation matches the agreement, including whether the target concerns employment at a checkpoint or continuous employment. If it does, calculate the gap while retaining the coverage limitation.

This two-column assessment prevents a common mistake: treating a missing or unclear result as poor performance, or treating a complete submission as proof that performance is on track.

Define status rules before assigning colors

Use labels such as on track, needs review and off track only with an explicit rule for that measure and checkpoint. Keep “not assessable” available when evidence is missing or incomparable. One universal percentage threshold is rarely appropriate for every measure in a portfolio.

A workable rule specifies the expected path, tolerance if any, materiality, evidence requirement and response. For a quarterly activity count, a shortfall might trigger an explanation. For a critical safeguard or a time-sensitive commitment, the review may need a different process entirely. Agree these rules with the responsible team rather than letting a dashboard color make the decision.

Do not average several status labels into an overall green without stating how the assessment works. Strong performance on one measure does not erase an unresolved concern on another. Show the affected commitments and let the reviewer understand the pattern.

Investigate the gap before choosing an action

The variance tells you where to look; it does not establish why the result differs. Read the partner’s explanation alongside relevant evidence, then distinguish a delivery issue, timing difference, changed context, unrealistic target or measurement problem.

• Observation: what differed from the agreement?
• Explanation: what does the partner believe caused it?
• Supporting evidence: what makes that explanation credible?
• Alternatives: what else could explain the difference?
• Action: what will be changed, clarified or supported?
• Follow-up: who owns it and when will the result be reviewed?

In the fictional case, the five future offers may help explain timing, but they are not guaranteed future starts. An action could be to check those actual start dates next period and investigate whether additional support is needed. Keep the present quarter’s shortfall visible.

What if the target or definition needs to change?

Record the original agreement, proposed revision, reason, approval and effective date. A justified change may improve future monitoring, but it should not silently rewrite prior performance.

If a definition changes, assess whether earlier values can be restated from source evidence. If they cannot, mark the break in comparability. If a target changes prospectively, distinguish the result against the original commitment from progress against the revised plan. Do not make a historical shortfall disappear by replacing the denominator.

Keep the review record with the relevant reporting period. A future colleague should be able to see what was agreed at the time, what evidence was available and why the decision was made.

Build a useful portfolio view

At portfolio level, show where action is needed without discarding the underlying differences. Group comparable measures where appropriate, but preserve partner-specific commitments and unresolved evidence. A count of red or green organizations is not a substitute for understanding which outcomes or obligations are affected.

Include measure, actual, target, checkpoint, performance status, evidence status, owner and next review date. Link each row to its source and decision record. Use the impact reporting guide and report examples when turning that view into a narrative for readers outside the review team.

What should you test in Sopact?

Bring an agreed target, a reviewed submission and one definition change. Check whether the actual result remains linked to its source and period, whether the comparison uses the right definition, and whether reviewers can inspect the reason behind a status.

Use configured analysis to prepare candidate variances and questions. Confirm calculation rules, permissions and review ownership during setup. People should approve consequential interpretations and target changes. The useful result is a comparison that can be explained and revisited, not just a colored dashboard.

Frequently asked questions

What is impact variance?

It is the difference between an observed result and an agreed expectation for a defined measure and time. Its meaning depends on comparability, the direction of improvement and the evidence supporting the result.

Can all measures use the same threshold?

Usually they need different rules. Counts, rates, milestones and critical commitments have different timing and consequences. Document the appropriate rule for each measure.

Should missing data be marked off track?

Keep evidence status separate. A missing result may prevent assessment, but it does not by itself establish the underlying performance.

Is a ten-percentage-point gap the same as a ten-percent gap?

No. A change from 80% to 70% is ten percentage points. Relative to 80%, it is a 12.5% decrease. Label the calculation clearly.

Can a target be revised?

Yes, through an agreed review process. Preserve the original, reason, approval and effective date so the revision does not silently erase prior commitments.

What happens after a variance review?

Assign a specific action, owner and review date. Retain uncertainty and source references, then check whether the action addressed the issue at the next checkpoint.

Bring the result back to your module

Related practice: design portfolio monitoring alerts. Use the response rules and ownership decisions from this lesson to decide which changes deserve attention.

Add this part to your plan

On workbook page 4, map a revised C-12 return and a changed outcome definition. Show what the earlier board report still points to.

Self-check: Can a new colleague reconstruct an earlier report from the retained record?

Questions you may have

Should the latest upload replace all earlier evidence?

No. Preserve its relationship to earlier versions and identify which version is approved for each reporting use.

Can one company belong to several funds?

Yes. Model the distinct investment relationships and define attribution and aggregation rules.

Apply the method to your work

Use the five-part plan to assess the sources, analysis and permissions your team needs. The solution page shows where a connected platform can support that workflow.