Creative analytics
Creative analytics answers the question every marketer actually cares about: which content makes money. Not which ad got clicks — which hook, which angle, which pairing of the two put revenue in the bank.
Because every creative you make in Omnymous is built from known parts — a hook, an angle, a customer avatar, an offer — the platform can take performance apart by any of those parts on its own. You never touch a UTM or invent a naming convention; the lineage travels with the creative from the Library all the way to the order it drove.
Open Creatives under ANALYTICS. The switcher at the top right has four views: Overview, Combinations, Pages and Insights. The date range and attribution model pickers above them apply to everything.
Overview
The Overview is your creative command center: what you produced, how the portfolio is performing, and who’s actually winning.
Creative Output
The band at the top reads in two halves, on purpose:
- Produced this period — how many items you Generated and how many creatives Launched (entered the market for the first time), with the launch rate between them. Below sit your AI cost (period), Cost / launched and the Generated mix, plus the Production cadence chart — generated and launched, day by day.
- In market this period — every creative With Spend (whenever it launched), how many are Winners, and the hit rate between them.
The dial on the right shows your hit rate against the 5–10% industry expectation band — if you’re inside it, you’re producing at a healthy standard. Next to it, Winning creatives lines up the winners themselves, each with its return and spend.
Decide what “winning” means
The Winner ≥ … control in the band’s corner is yours to set. Click it and define your own bar — for example ROAS ≥ 2.0x — and every winner count, the hit rate, the dial and the leaderboard marks recalculate to your definition. Insights use the same bar, so a Creative Winner finding means winning by your rules.
Leave it unset and Omnymous uses a statistical default: 1.3× your portfolio’s pooled ROAS. Either way the bar is judged on a shrunk ROAS — small samples get pulled toward your average, so a creative can’t fluke its way to winner on a lucky $100.
The portfolio charts
- Portfolio Map — every creative as a dot: spend across, ROAS up, with your average and the winner bar drawn in. Winners are the violet dots above the line; a dot far right and low is where budget is leaking.
- Spend Concentration — how dependent you are on your top creatives (“Top 3 of 18 creatives carry 24% of spend”). A curve that shoots up immediately means one or two ads carry the account.
- Where the Spend Went — your budget split By format and By hook type, with the return each share earned.
Creative Leaderboard
Every launched creative, ranked by attributed return — no naming conventions required.
Each row carries its thumbnail and, under the name, a Winner mark when it clears your bar plus its hook · angle — so you can see at a glance that your top performer is, say, a question hook on the Science-backed results angle. Share shows each creative’s slice of spend; the attributed columns run through Orders, Revenue, ROAS, POAS and CPA.
Combinations — the DNA lab
Single parts win, but pairings often win bigger. The Combinations view is one connected lab for the whole question: which variables work, which combinations work, and which combinations you’ve never even tried.
Pick the lens first — POAS, ROAS, CVR or Spend — and all three zones repaint.
The Variables
One column per dimension — Customer Avatar, Angle, Hook, Offer, Product, Hook Type, Headline and more — each value ranked by how much spend it carried, painted by your chosen metric. The header counts tell you how much of your library the market has actually seen: 4 of 4 avatars tested, but maybe 6 of 18 hooks.
Click any value to focus it: every other column recomputes for that value alone. Focus The Busy Parent and the Hook column becomes “which hooks work for The Busy Parent” — that’s the question generic ad reporting can’t answer.
The Matrix
Choose any two parts — Hook × Avatar, Angle × Offer, Headline × Landing Page — and every pairing becomes a cell, colored by performance and marked “1 tested of 72 possible pairs” so the size of the unexplored space is always in view.
- Colored cells are tested pairings — violet above the breakeven or benchmark, rose below. Hatched cells ran on under $50 of spend: treat them as hints, not verdicts.
- Dashed cells never ran together. Click one and Omnymous tells you whether each half has proven itself separately — an untested pairing of two individually-strong values is your cheapest next experiment.
- Click any cell and the panel shows its spend, orders, revenue, CM2 profit, the CVR confidence range and the creatives carrying the pairing — then hands off with Generate with this pairing, which opens the Studio with both parts preloaded.
- Show untested adds library values that have never launched at all; All tested values expands past the top rows and columns.
Winning Recipes
The table at the bottom ranks every 2–3-part combination by CM2 — the contribution profit it produced, not just revenue. Use the preset pairings or compose your own three-part recipe, and raise the minimum-spend filter ($0 / $50 / $100 / $250) when a “winner” looks too good to be true — sparse pairings get lucky on a few dollars all the time.
Pages
The Pages view covers where your traffic lands.
- The Funnel walks every visit from Sessions → View Content → Add to Cart → Checkout → Purchases, with the conversion between each stage and your session-to-purchase rate, session revenue and revenue per session below.
- Survival by Channel shows the same journey per traffic source — if Meta traffic reaches the cart at twice the rate of TikTok traffic, it’s visible in one glance.
- Page Templates ranks which of the built-in page templates converts paid traffic best, with each template’s Share of spend and its CVR Range confidence bar.
- The Landing-Page Funnel breaks the journey down per entry page, so a page that loses everyone at the cart is obvious at a glance.
Reading the confidence columns
Alongside exact numbers, ranked tables add two columns that tell you how much to trust a result:
- CVR Range — a confidence range for the true rate, drawn as a small bar. A wide bar means there isn’t enough spend behind it yet to be sure; a tight bar means the number is solid.
- P>Avg — the chance this value truly beats your account average. High is good; near zero means it’s a laggard.
Insights
The Insights view is Omnymous watching your data for you. Eight detectors run daily — creative winners, winning and losing concepts, creative fatigue, wasted spend, budget shifts, page tests and tracking health — and each one is statistically gated, so it only fires when there’s enough evidence to trust.
- Every finding carries a one-click action back into the work — Open in Launcher, Generate more on this concept, Compare campaigns — so an insight is never a dead end.
- Why this fired under each card opens the exact evidence: the spend, the returns and the thresholds it cleared. No black boxes.
- Work the feed with the New / Acknowledged / Actioned / Dismissed filters: Acknowledge what you’ve seen, dismiss what you disagree with, and anything you action is tracked as handled.
- How insights work (top right) explains all eight detectors in plain language.
How much of your spend is explained
Wherever creative analytics ranks by lineage, it shows a quiet honesty note: Explains N% of ad spend. That’s the share of your ad spend these rows can actually account for. Ads you imported from a platform (or any creative without lineage) can’t be decomposed, so they’re left out — and the note keeps that in view, turning amber when coverage is low, so you never scale a “winner” measured on a sliver of your budget.
Launching your creatives through Omnymous is what keeps that number high — every part is tracked automatically, so more of your spend becomes explainable.
Troubleshooting
The leaderboard (or the DNA lab) is empty. These views only rank creatives whose lineage Omnymous tracked — which happens when you build and launch through the platform. Launch a batch from the Launcher and the tables fill themselves as spend comes in. Ads imported straight from a platform have no lineage to decompose.
Every range is wide and the numbers look unsure. There isn’t enough spend behind them yet. A CVR Range bar narrows as a hook, angle or page accumulates budget — give the winners more spend, or widen the date range, and the estimates firm up.
No insights are showing. Detectors run daily and are deliberately strict — a quiet feed means nothing has cleared the evidence gates yet, which is normal in the first days after connecting. Check back once your campaigns have a week of spend behind them.