The product / Context changes everything
From a single ad to the bigger picture.
Keep the context an ad library leaves behind. See what launches, what stays in the mix, and what has a recorded ending.

CreativeFirst seen
ContinuesRecorded end
CompleteGaps visible
What changedWhat is unknown
Claude CodeCodexMCP
01 / Build a record
Save the before.
Understand the after.
A creative’s history belongs next to the creative. Dated observations make it possible to return to the evidence, not just a remembered impression.
- 01First observed
The creative enters the public record.
- 04Additional variant
A related variation appears while the original continues.
- 07Context preserved
Compare the dated observations. Business results remain unknown.
Fictional advertiser. Example records, not live data.
02 / Compare the changes
What changed?
What stayed with it?
Compare the opening, offer and visual format. The presence of a variant is observable; its commercial impact is not.
- What changed
- The opening message and colour treatment.
- What stayed
- The product, format and central visual.
- What is unknown
- Spend, conversions and which creative performed better.
Authored examples for a fictional advertiser. Not performance evidence.
03 / Inform the brief
Better questions.
A more useful starting point.
Bring the evidence.
Leave room for your own idea.
- Which opening or offer keeps appearing?
- Where are related variants exploring a different angle?
- What could we test — without assuming their results?
A question for your team, not a recommendation to copy an advertiser.
Take a closer look
Three ways to read the record.
Follow the activity, investigate the creative, and check what the evidence supports.
Competitor monitoring
Follow public ad activity across dated observations.
Explore the details02 / The productCreative research
Turn observed creative patterns into better research questions.
Explore the details03 / The productEvidence & coverage
See what a record supports before drawing a conclusion.
Explore the detailsKnow what you can rely on
Every signal has a boundary.
See how complete observations and qualified interpretations fit together.
Read the method