# Evidence and honest impact reporting

Choose the claim your evidence can support before choosing the report headline. Keep reach, delivery, use, return, and evidenced change separate. Follow up without leading the answer, retain missing information as unknown, and publish findings beside their method, limitations, and next decision. A report can be complete without establishing causation or verified impact.

By Daniel Ziedins (https://www.danielziedins.com).
Updated: 2026-10-21
Canonical: https://www.agent-six33.com/journal/topics/evidence-and-impact

A public report should help a reader inspect the work, not ask them to trust the energy of the presentation. This collection connects measurement vocabulary, follow-up conversations, and reporting decisions. It is useful after a small delivery, before a public update, or when a team needs to distinguish an encouraging story from a supportable outcome claim. SIX33's own current count remains separate from every hypothetical example in these guides.

Begin with the difference between distribution and change. A page being published is delivery; somebody using it is a different observation; a later account about a changed action is different again. Follow-up helps you understand what happened next, but the response rate, source, timing, assistance, and other influences still matter. Do not infer an outcome for people whose experience you could not establish. A moving quotation cannot fill a missing denominator or prove that a result was typical.

When the evidence is organised, use the reporting guide to put the finding beside the promise, method, limitations, and decision. The ministry guide applies the same care to participation, stories, and spiritual claims without treating a person's life as a marketing unit. Choose measures appropriate to the actual work and use specialist evaluation or professional review when the stakes require it. The scorecard and proof pack can organise a record and flag missing context; they cannot independently verify whether the underlying accounts are true.

## Choose the guide for your decision

### 1. Vanity metrics vs impact metrics: how to tell them apart.

When a large number is doing too much work: distinguish reach, delivery, use, return, and evidenced change before combining any claims.
Read: 6 min
https://www.agent-six33.com/journal/vanity-metrics-vs-impact-metrics

### 2. How to follow up after a pilot: questions, script, and evidence.

When the delivery is complete but the next chapter is unknown: ask neutral questions and report responses and missing information separately.
Read: 7 min
https://www.agent-six33.com/journal/how-to-follow-up-after-a-pilot

### 3. How to write an impact report people will actually trust.

When a finding is ready to share: put the method, limitations, and next decision beside the result, then review the draft.
Read: 9 min
https://www.agent-six33.com/journal/how-to-write-an-impact-report-people-trust

### 4. How to measure ministry impact without inflating the numbers.

When reporting community or ministry work: handle participation, stories, and claims about change with care rather than inflation.
Read: 6 min
https://www.agent-six33.com/journal/how-to-measure-ministry-impact

## Put it to work

Organise a report in the proof pack: https://www.agent-six33.com/proof-pack#build-proof-pack
Use the local-first builder to retain the finding, method, limitations, and decision. A completeness check is not independent verification.

## Questions

### What should I measure before writing an impact report?

Start with the intended promise and the people and time window involved. Separate distribution, delivery, use, return, and evidenced change, and record the method and limits of each finding. Choose measures suited to the actual decision rather than collecting large numbers first and inventing a claim afterward.

### Can follow-up responses prove that a project caused a change?

Not by themselves. Accounts can help describe what happened afterward, but other influences, who responded, timing, and missing information affect interpretation. A causal claim requires an appropriate evaluation design, not simply a positive quote or a before-and-after story.

### Does passing the proof-pack review verify impact?

No. The review checks completeness and certain data-quality issues in the record provided. It cannot independently establish that a finding is true, representative, or caused by the project. Publish the method, limitations, and review boundary beside the claim.