Metrics explained

Every metric in the results table and what it tells you.

Results table columns

Column	Definition
Sessions	Number of PAGE_VIEW events for this variant
Orders	Attributed orders (from webhooks via cart attributes)
CVR	Orders ÷ Sessions — the conversion rate
ATC Rate	Add-to-cart events ÷ Sessions
Checkout Rate	Initiate-checkout events ÷ Sessions
Revenue	Total order revenue attributed to this variant
Rev / visitor	Revenue ÷ Sessions — revenue per visitor
AOV	Revenue ÷ Orders — average order value
Lift	Percentage improvement in CVR vs control

Which metric to optimise for?

For most Shopify stores, CVR (conversion rate) is the primary metric. It directly measures whether the variant drives more purchases.

Revenue per visitor is better than CVR when you’re testing price changes or upsell flows — a variant could have lower CVR but higher AOV, making it more profitable overall.

ATC Rate is a useful leading indicator — it moves faster than CVR so you can spot direction early. Don’t conclude from ATC alone.

Lift

Lift is calculated as:

lift = (variantCVR - controlCVR) / controlCVR

A lift of +12.5% means the variant converts 12.5% more visitors than control, not 12.5 percentage points more.

Sample size guidance

As a rough guide, you need approximately:

CVR	Minimum sessions per variant for 80% power
1%	~10,000
2%	~5,000
5%	~2,000
10%	~1,000

These are estimates for detecting a 10% relative lift. Smaller expected lifts require more sessions.

Novelty effect

A novelty effect warning appears when a variant’s CVR was significantly higher in the first 48 hours than in subsequent days (≥40% drop-off). This usually means returning customers are inflating early results because the change feels novel. Wait for the effect to stabilise before concluding.