Metrics
Every number on the Dashboard and the Analytics screens, with what it counts and the formula behind it.
The screens carry a short form of each formula: rest the pointer on a figure to read it, or read it under the figure on a phone. This page gives the whole definition, once. Each heading is the name the screen uses, so you can search for it.
Three things hold for every figure. Money is in your store's currency, and ad spend in another currency is converted at each day's rate. Days are your store's days, in its time zone. And "vs prev." compares the period with the same number of days before it. See How to read the numbers.
Sales
Gross Sales
The sum of price × quantity on every line of the period's countable orders. It is before discounts, and it leaves out tax and shipping.
Discounts
The order-level discounts given on the period's countable orders.
Refunds
The goods refunded in the period, counted on the day of the refund. A refund in October of an order from September lowers October. The tax and shipping refunded with the goods are not in it.
Net Revenue
Gross sales − discounts − refunds. What your orders brought in for their goods. It is the basis of every ROAS on these screens, of AOV and of the cost of a new customer.
Orders
The period's countable orders. An order counts when it is paid, partly paid, partly refunded or refunded. Cancelled orders, and orders that are pending, only authorized or voided, are left out.
AOV
Net revenue ÷ orders. The average order value. On a table of ads, campaigns or creatives it is attributed revenue ÷ attributed orders.
Costs and profit
Each line takes the one above it and subtracts a cost. You enter most of those costs yourself: see Enter your costs.
COGS
The cost of the goods sold in the period's orders. For each line, Omnymous uses the first cost it finds: your COGS override for the product, then the cost per item set in Shopify, then your fallback percentage of the line's revenue. With none of them, the line costs nothing.
CM1 (Gross Margin)
Net revenue − COGS.
Shipping Revenue
The shipping your buyers paid, less the shipping you refunded. It counts toward CM2 and the lines below it, never toward net revenue or ROAS.
Shipping
What fulfilment and carriers cost you, from your shipping cost rules. Omnymous does not read your carriers' bills, so without a rule this line is zero.
Payment Fees
The payment processing fees on the period's orders, read from Shopify. The tile on Store's Profit view calls them Gateway Fees. Do not enter them as a cost rule: they would count twice.
Variable Costs
The costs your variable rules add per order, per line, per refund, by weight, by quantity, or as a share of revenue or of COGS. Handling, packaging and warehouse fees belong here.
CM2 (Contribution)
CM1 + shipping revenue − shipping − payment fees − variable costs. What each order leaves before advertising. Credited to ads and atoms, it is the basis of POAS.
Ad Spend
What your ad accounts spent, on every connected platform. Spend in another currency is converted at that day's exchange rate. A day with no rate yet is left out, and the tables that miss it say so.
CM3
CM2 − ad spend.
Fixed Costs
Your flat monthly cost rules, spread evenly over the days of the period. A $3,000 monthly rule adds about $100 to each day.
AI Spend
What your AI providers charged for work on your keys, converted to your store's currency. It is estimated from the providers' prices, day by day.
Net Profit
CM3 − fixed costs − AI spend. It is only as accurate as your costs.
Net Margin
Net profit ÷ net revenue.
Marketing efficiency
Blended ROAS (MER)
Net revenue ÷ ad spend, across every platform. It credits no ad in particular: it says what the whole store earned for each unit of ad spend. The Dashboard calls it Blended ROAS; Store calls it MER.
aMER
New-customer revenue ÷ ad spend. The same, counting only the revenue of first orders.
nCAC
Ad spend ÷ new customers. What a new customer cost you in ads.
Customers
New Customers
Customers whose first order was in the period.
Returning Customers
Customers who ordered in the period and had ordered before. Each counts once, however many orders they placed. Orders without a customer account cannot be recognised as returning, so they are never counted.
New-Cust. Revenue
The net revenue of the period's first orders.
Retention
The share of a cohort's customers who ordered in a given month after their first. A cohort is the customers whose first order fell in one calendar month. Month 0 is always 100%.
Cum. LTV
The revenue a cohort's customers brought in, per customer, from their first order up to a given month.
LTV 90d and LTV 360d
The net revenue per customer within 90 or 360 days of the first order. Only customers whose first order is at least that old are counted, so a recent month never pulls the figure down.
Delivery
What the ad platforms report about showing your ads. The attribution model does not change these.
Spend
What the platform charged for the row's ads in the period.
Impressions
How many times the row's ads were shown. Tables write it Impr.
Clicks
How many clicks the platform counted on the row's ads.
CTR
Clicks ÷ impressions.
CPC
Spend ÷ clicks.
CPM
Spend ÷ impressions × 1,000. The cost of a thousand impressions.
What the platforms claim
Reported Revenue
The conversion value the ad platform claims for its ads. Only Meta reports it; Google and TikTok report delivery only, so their rows show a dash. Tables write it Reported Rev.
Reported ROAS
Reported revenue ÷ the spend of the platforms that report revenue.
Claim Gap
Reported ROAS ÷ attributed ROAS. How much more the platform claims than Omnymous credits it. Above 1× is normal: a platform counts every sale it believes it caused, while Omnymous credits each order once, across every touch in the buyer's journey.
What Omnymous credits
These depend on the attribution model chosen at the top of the screen. Each order's net revenue is split across the touches that led to it, and each touch passes its share to its campaign, ad set, ad and creative.
Orders (attributed)
The orders credited to the row, counted by their share of the credit. An order split between two ads counts as half an order on each, so a row can show 0.5.
Revenue (attributed)
The net revenue credited to the row.
ROAS
Attributed revenue ÷ spend.
POAS
Attributed contribution (CM2) ÷ spend. Profit on ad spend: 1.0× is breakeven, after the cost of goods, shipping, fees and the ads themselves.
CPA
Spend ÷ attributed orders.
NC-ROAS
Attributed revenue from new customers ÷ spend.
Attributed Revenue and Attributed ROAS
On Campaigns' Performance view: the revenue credited to your Meta campaigns, and that revenue ÷ the ad spend of the channels selected. Google and TikTok campaigns are not counted in the revenue, though their spend is in the division.
Credited revenue
On Journeys: the credit the attribution model gives to the sessions drawn, in revenue.
Creative output
The creative figures on the Dashboard and in Creatives. Winners are judged on last-click credit whatever model is chosen, so changing the model never changes who wins.
Generated
The items Studio completed in the period, of every type: images, videos, hooks, angles, copy and pages.
Launched
The creatives whose first ad went live in the period. A creative launched in an earlier period and added to a new ad does not count again.
Launch rate
Creatives launched ÷ the images and videos generated in the period, at most 100%. Creatives Overview prints it between the two as "of ad creatives launch". A creative made in an earlier period and launched in this one counts too.
With Spend
The creatives whose ads spent in the period, whenever they were launched.
Pooled ROAS
The last-click revenue of every creative with spend ÷ their spend. The average return of your portfolio, which the winner bar is measured from.
Shrunk ROAS
(Revenue + k × pooled ROAS) ÷ (spend + k). A creative's ROAS pulled toward the pooled ROAS, so that a few lucky orders on little spend cannot make a winner. k is the spend of five orders at the pool's average cost per order, and never less than 200 in your currency. The more a creative spends, the less it is pulled.
Winners
The creatives with at least 100 of spend, in your store's currency, whose shrunk ROAS reaches the winner bar. The bar is your own target ROAS when you set one, otherwise 1.3 × the pooled ROAS. Creatives Overview calls them Efficiency winners.
Win Rate
Winners ÷ creatives with spend. Creatives Overview calls it the hit rate, and prints the 5–10% that is usual in the industry beside it.
Adoption winners
The creatives that spent at least ten times the median creative's spend in the period. What your account chose to scale, whatever it returned.
AI cost
What your AI providers charged for the period's generations, in US dollars. Providers bill in US dollars, so it is not converted.
Cost / launched
AI cost ÷ creatives launched. The Dashboard calls it AI cost / launched.
Share
On the Creative Leaderboard: the creative's spend ÷ the spend of every creative in the period.
Effects and estimates
The Lab, Experiments, Combinations and Pages estimate what the data supports, not only what happened. Each estimate is pulled toward the average of its kind until there is evidence to say otherwise, and comes with a 90% interval. See Attribution Lab.
Effect vs family
The atom's rate at its stage ÷ the average atom of its dimension, adjusted for everything else its ads carried. 1.00× is average. 1.20× means 20% more clicks per impression, add-to-carts per session or orders per click than the average hook, angle or offer, all else equal. The bar around it is its 90% interval.
P>family
The probability that the atom truly beats the average of its dimension.
Incremental
The revenue the atom earned compared with the average atom of its dimension. Negative when it cost you money. Every atom's incremental revenue, plus what an all-average ad would have earned, adds up to the total credit.
Ratio
The stronger atom's effect ÷ the weaker one's, on the Lab's Contrasts, with its 90% interval and P(beats), the probability that the stronger one truly is.
Expected rate and Value / 1k
For a pair of atoms that never ran together: the rate the model expects at their stage, and the contribution (CM2) it expects per 1,000 exposures. An exposure is an impression at the attention stage, a visit at the intent stage and a click at the conversion stage. On the Lab's Gaps the rate is named after the stage, such as Expected CTR.
CVR Range and P>Avg
The estimated orders per click of a row, with its 90% interval, and the probability that the row truly beats the average of the rows beside it.
Identified share
The share of a stage's evidence carried by atoms whose effect can be told apart, as in "78% identified at attention" on the Lab.
Sessions and the funnel
Sessions
Visits to your store recorded by the Omnymous pixel. A session ends after 30 minutes without activity.
Funnel stages
The sessions that reached each step: a product page, an add to cart, a started checkout, a purchase. Pages names them View Content, Add to Cart, Checkout and Purchases; Journeys names them Saw a product, Added to cart, Started checkout and Ordered. A session counts once at each step, however many products it viewed. The percentage beside a step is the share of the step before it that went on.
Session → purchase
Purchases ÷ sessions. On the Landing-Page Funnel the same figure, per entry page, is the CVR column.
Session revenue
The value of the purchases the pixel recorded in the period's sessions.
Revenue / session
Session revenue ÷ sessions.
Buyers and Buy rate
On Journeys' Paths: the sessions that ended in an order, and buyers ÷ sessions.
Measurement health
How far the other figures can be trusted. See When your numbers look wrong.
Pixel match rate
The share of the last 14 days' orders that the pixel tied to a visit. It turns amber below 70%.
Ad match rate
The share of the last 14 days' visits from your ads whose link named the ad. Only a visit that names its ad can be credited to it. It turns amber below 70%. See Tag your ads.
Conversions (14d)
The orders of the last 14 days that went through attribution, whether or not a visit could be found for them.
Pixel Events (24h) and Sessions (24h)
What the pixel sent in the last 24 hours: every event, and the sessions they make up.
Coverage
The spend of the ads a table can place ÷ all ad spend in the period, printed as "Explains 100% of ad spend" when every ad can be placed. A table of creatives or atoms can only place an ad whose creative is in your Library. It turns amber below 60%: the rows then rest on a small part of your spend.