Four measured counts, everything else derived
Every offer records four things directly: how many times it was shown, how many times a product was added to cart from it, how many times it converted, and how much revenue it produced. Conversion rate, add-to-cart rate and cart-to-purchase rate are all ratios of those four, so the funnel is internally consistent — you can see whether a weak offer is failing to get attention or failing to close.
Because the counts are per offer, per placement, you can compare the same product offered on the product page against the cart drawer against post-purchase, and see which moment actually works for it.
Revenue is net of discounts
If an offer sells a $40 product at 25% off, it reports $30 — not $40, and not the value of the whole order. That is the only honest way to compare a discounted offer against an undiscounted one, and it keeps the headline number from flattering itself by counting revenue the offer didn't create.
Attach rate and upsold AOV
Attach rate is the share of your orders that contain at least one upsell — a store-wide measure of reach rather than a per-offer conversion rate.
Upsold AOV compares the average order value of orders that included an upsell against the average of orders that didn't. That second cohort is the important part: the obvious way to calculate this is to compare upsold orders against your overall average, but your overall average already contains the upsold orders, which understates the real lift. UpsellPlus compares against a clean non-upsold cohort instead, and only reports the figure once both cohorts have enough orders to be meaningful.
Revenue by placement
Revenue is attributed across the surfaces it came from — checkout, post-purchase, cart, product page and storefront pages — so you can see where your incremental revenue is actually being generated rather than inferring it from conversion rates. For a store running offers at several stages, this is what tells you which stage deserves more attention.
Bots are excluded
Requests from bots and crawlers are identified and served nothing, so they never count as impressions. Without that, crawler traffic silently inflates impression counts and deflates every conversion rate derived from them — which makes offers look worse than they are and makes two offers on differently-crawled pages incomparable.
Rewards report separately
Reward programs have their own analytics, distinct from offer analytics, because a reward measures progress toward a threshold rather than acceptance of a product recommendation. If you run both, read both — offer-level revenue does not include reward-driven revenue.
