Original workflow / Source you can inspect
Less noise. More signal.
FIELDNOTES / A fictional 30-day social review. Find the posts worth learning from, keep the calculations visible and make the next test concrete.
These are rendered results from this source, not a recording of a live app. Chapter buttons seek the movie; press Play to continue.
Channel performance
| Channel | Posts | Impressions | Engagement % | CTR % | Spend cents |
|---|---|---|---|---|---|
| 3 | 63000 | 4.33% | 1.02% | 37000 | |
| 2 | 22000 | 3.64% | 1.89% | 7500 |
Inspect the actual post rows.
The next experiment is a proposal.
| Hypothesis | Action | Source IDs | Guardrail / Owner |
|---|---|---|---|
| The process opener may earn more useful responses. | Test two different process openers | sep-video-1, sep-video-2 | Keep topic, format and 7-day observation window matched. / Content lead |
| A reusable checklist may improve intent signals. | Publish a controlled checklist draft | sep-carousel-1 | Do not infer purchases from click counts. / Community lead |
Calculation boundaries
Engagement counts likes, comments and shares. Clicks are a separate action count; all rates divide by impressions. Zero-denominator rates remain unknown. Impressions are not unique reach. Spend uses supplied integer cents, without currency conversion.
Edit the source. Make it yours.
Edit the full input, download it next to any supplied source files, then rebuild into a new folder. Existing MP4 and PowerPoint downloads keep their generated content until rebuilt.