Which creative actually closed revenue. And what nobody can measure.
Forge 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 closed revenue divides across creative families, and Forge will name them.
Not impressions, not clicks, not a platform-reported conversion. Where the conservation gate passes, Forge assigns each eligible closed-revenue contribution to one creative family. Every visible index carries the closed-deal sample it rests on.
- Leading creative familySolid
Index 100 · 12 closed · 90d
- Creative family 2Solid
Index 85 · 7 closed · 90d
- Creative family 3Solid
Index 65 · 13 closed · 90d
- Creative family 4Solid
Index 28 · 4 closed · 90d
- Creative family 5Solid
Index 4 · 3 closed · 90d
18 families. 100% coverage. Values are indexed to the leading family, not amounts. 13 more below the closed-deal floor.
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. Forge 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. Forge 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.
Forge 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.
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.