Which creative actually closed revenue. And what nobody can measure.
Edge ranks your creative by the deals it closed rather than the clicks it earned. Where the ad platform mixes your assets and makes a finer answer impossible, it says so in plain language, tells you exactly what it can still prove, and shows you how to restructure so the finer answer becomes available.
The claim
Your revenue backs one creative over another, and Edge will name it.
Not impressions, not clicks, not a platform-reported conversion. Edge ranks your creative on the deals it closed, and every entry in the ranking carries the sample it rests on. Where a family has not produced enough outcomes to rank, it is withheld rather than listed at the bottom.
- Best closing videoScale-readySolid
Rank 1 · 34 attributed outcomes · 90d
- Video 2HoldSolid
Rank 2 · 21 attributed outcomes · 90d
- Video 3Needs more dataMaturing
Rank 3 · 12 attributed outcomes · 90d
Order, not amounts. A ranking without its sample size is a claim rather than a verdict, so the evidence line is not optional here. Families below the sample floor are withheld from the list entirely.
The limitation
No platform can tell you your single best headline under dynamic mixing. Including this one.
When your campaigns run dynamic creative, the ad platform mixes your headlines, bodies and images across ads and reports performance only at ad grain. The single-element answer does not exist in the data anyone receives. Edge diagnoses this per campaign, states it in those words, and then tells you what it can still prove.
These campaigns use dynamic creative. The platform mixes your assets across ads and reports at ad grain, so no tool can isolate which single headline closed revenue.
- What we can prove: which creative family your revenue backs, and on what sample.
- What we can prove: the complete element set inside your best closing ad.
- What we can prove: your best videos by closed revenue.
- What nobody can prove here: which one headline did the work.
The diagnosis is a computed read over your ingested campaign structure, per campaign cohort, not a banner triggered by a guess.
Why this is the pitch
Anyone selling you element-level rankings under dynamic mixing is selling you a number they invented.
We checked. The data required to rank a single headline under dynamic creative is not in what the platform returns, so a ranking presented at that grain was modelled and then presented as measured. Edge refuses that trade. It shows the achievable grain, names the ceiling, and treats the ceiling as a staircase rather than a wall.
The staircase
Restructure for clarity, and the measurement upgrades itself.
Edge teaches the campaign structure that makes element-level truth measurable, recommending you isolate the family your own data already prefers. When your restructure produces separable variance, the grain upgrades automatically and says so, citing the ads that made it possible. Revert, and it degrades back just as visibly.
Acting on it
Pause, activate, adjust budget, duplicate a winner, without leaving the verdict behind.
Every action executes a specific verdict and records which one, so the trail reads as a decision with its evidence rather than an untraceable change. Budget moves are bounded, nothing destructive is on the menu, and the aftermath is observed and taught back to the brain that recommended it.
Generation
Variations written from your own closed-revenue evidence, and tracked from birth.
Signet drafts new creative from the brief your data writes: the winning family's members, the element sets inside your best closers, your diagnosed structure. Every variation is registered before it ships, so when it runs, its performance attaches to the recommendation that produced it. Sigil suggests, and nothing goes live until you adopt it.
Second act
The sales call, read against criteria you wrote.
Edge also reads your recorded calls against a rubric you author yourself. Publishing mints a version, and every call records the version that read it, so an answer from March is still explainable in December. Route different call types to different rubrics, and test a change against real calls before it goes live.
Creative decisions with the evidence attached.
Every module is part of one platform. You get all of it, and the tiers differ by scale rather than by which capabilities are switched on.