CapCut received a new creative angle with its own measured reason to believe.
Panoverse linked the recommendation to engagement performance by creative angle, so the content team could inspect the metric cited in the proposal.
4.3%
Engagement rate on newly proposed angles
8 in 10
Angle proposals approved on first review
11
New viable angles identified per quarter
“The new-angle proposal cited our own measured ER — first rec that came with its evidence.”
Company
CapCutCreative software
Region
Global
Team
Content marketing
Use case
Creative intelligence · Better briefs
10use cases running on PanoverseClient statement
01USE CASE
Ship angles that carry their own measured reason to believe
The problem
New angles were easy to propose and hard to prioritize. Without a visible link to measured audience response, a recommendation came down to taste and persuasion.
The result
The content team judges the idea and its evidence in one pass, which makes the brief easier to challenge, improve and approve.
Engagement rate on newly proposed angles
1.8%Before Panoverse
4.3%Today
Goal
Give a creative proposal a measured reason to be executed, so it can be challenged on evidence rather than settled on taste.
Deployed
Engagement performance organized by creative angle across the account's own published content, at the account's own granularity.
How it works
Angle is treated as a property of the content, not of the campaign it ran in, so the same angle can be compared across creators, formats and weeks. A campaign-level average cannot do this: it reports the mix, and the mix is what the proposal is trying to change.
The proposal names the angle, the measured response it was derived from, and the window that response was measured over — and the cited metric stays openable from inside the brief. A reviewer who doubts the number reads it instead of asking for it.
Result: held until the client approves it for publication.
02USE CASE
Rank angles by response, not by who argued hardest
The problem
In a review with no shared measure, the ranking produced is a ranking of confidence, and confidence is distributed independently of whether an angle works.
The result
Angles carry their measured response and its window, so the ranking is a lookup that any reviewer can reproduce.
Angle proposals approved on first review
3 in 10Before Panoverse
8 in 10Today
03USE CASE
Find the angles working for the category and missing from you
The problem
An account's own history can only rank the angles it has already tried, and the most valuable one is by definition not in that set.
The result
Category content is read on the same angle taxonomy as the account's own, so an absence shows up as an absence rather than as nothing at all.
New viable angles identified per quarter
2Before Panoverse
11Today
04USE CASE
Scale the angle or the creator — but know which
The problem
A single strong post supports both explanations equally well, and scaling the wrong one is the more expensive mistake.
The result
The angle is compared across creators and the creator across angles, so the two are separable before the budget follows either.
Scaled bets placed on the wrong variable
4 in 10Before Panoverse
1 in 10Today
05USE CASE
Know which half of the video to rewrite, the same day
The problem
A single engagement figure blends the decision to start watching with the decision to keep watching, and those two failures need opposite fixes.
The result
The two are reported as separate series against the same creative, so the brief says which half to rewrite.
Time to diagnose which half of a creative failed
2 weeksBefore Panoverse
same dayToday
06USE CASE
Stop funding an angle a month after it stopped working
The problem
Decay is gradual and each week's drop is individually explainable, which is how an angle keeps its budget for a month after it stopped working.
The result
Decay is graded against the angle's own curve rather than week over week, so the trend is raised as a trend.
Days of spend on an angle already past its turn
24 daysBefore Panoverse
6 daysToday
07USE CASE
Set the bar before the creative ships, not after
The problem
A metric chosen after a creative has run will nearly always find something that improved, because something in the account always did.
The result
The metric, the window and the expected direction are written before the creative ships and cannot be edited after.
Creatives scaled that then underperformed the control
35%Before Panoverse
12%Today
08USE CASE
Re-check any number in the brief without leaving it
The problem
A number quoted in a brief and held in another system is a number nobody re-checks, and the brief inherits whatever error it carried.
The result
The figure links to the content rows it was computed from, so a reviewer settles the question by reading rather than by asking.
Review rounds per brief before approval
3.4Before Panoverse
1.6Today
09USE CASE
Stop re-pitching an angle that already failed twice
The problem
Angles that failed leave no trace, so the same idea is re-proposed every few months with the same confidence as the first time.
The result
Failed angles are kept against the test they were graded on, so the re-proposal starts from the record instead of from scratch.
Repeat proposals of an angle already known to fail
5 per yearBefore Panoverse
0Today
10USE CASE
Attribute the result to the angle or the format, correctly
The problem
An angle proven in one format is routinely commissioned in another, and the result is read as a verdict on the angle.
The result
Angle and format are carried as separate fields, so a brief states both and a result can be attributed to the right one.
Results attributed to the correct variable
46%Before Panoverse
91%Today
02WHAT RUNS TODAY
The system that is deployed
Panoverse organized engagement performance by creative angle, identified a relevant movement and wrote the proposal around the observed ER rather than a generic trend claim.
03STARTING POINT
The state before any of it
New angles were easy to propose and hard to prioritize. Without a visible link to measured audience response, a recommendation still depended on taste and persuasion.
04HOW IT WAS MEASURED
What the number rests on
Evidence kept
Measured engagement by creative angle
What fired
An angle relevant to the next brief
What changed
Write the proposal around observed ER
What was measured
The cited metric remains inspectable
Caliber
Client statement
The result shown is drawn from a client statement or from a backtest against the client's own history. Results vary with data, execution and market conditions.
05HOW THE TEAM WORKS NOW
The operating model
The content team could assess the idea and its evidence in the same pass, making the brief easier to challenge, improve and approve.
CapCut