Analytics, A/B tests & customers

How offers and orders are counted

Every product a shopper takes from an offer before checkout is marked on its order line, and a thank-you offer's second order carries a mark too, so the upsell revenue in Luv Upsell is revenue you can find in your own Shopify orders. A post-purchase offer counts once Shopify confirms the charge. A view counts every time a shopper sees an offer, whether or not they go on to buy, so every rate is out of everyone who saw the offer.

What each number means

• Upsell revenue: what shoppers paid for the products they took from your offers, after discounts and before tax and shipping, counted the same way at every place.
• Views and adds: how often shoppers saw an offer, and how often they took it.
• Take rate: adds ÷ views.
• Revenue per view: what each view earned, for comparing offers fairly.
• Orders with an upsell: how many orders carried at least one product from an offer, and their share of your orders.
• AOV with and without an upsell: the average order value of orders that included an upsell, next to those that didn't. Shoppers who take an upsell may have spent more anyway, so read the gap as a guide rather than proof that the offer caused all of it.
• Post-purchase: how many shoppers saw a post-purchase offer, how many took it, and what it added.
• Declines: how often shoppers said no thanks.

The numbers are shown day by day, by offer, by place and by funnel. Pick the last 7, 30 or 90 days or the last year, and export any table to CSV. Refunds aren't taken off; your Shopify reports have those.

A/B tests

In an offer, turn on Test a version B of this offer and give version B its own discount, words or products (on the thank-you page, words or products); anything left as Same as A stays the same. Choose how many shoppers see version B, from 5% to 95%. Before checkout, each shopper keeps the version they first got in the same browser; post-purchase and thank-you offers pick a version per order, and in-checkout offers one per checkout.

In Insights, By offer shows both versions side by side, with what the numbers say so far in plain words:

• Collecting: nothing is called until each version has at least 100 views, and it says how many more are needed.
• Ahead, not proven yet: one version has the higher take rate, but the gap could still be chance.
• Wins: the gap in take rate is statistically significant (p below 0.05), and the p value is shown.
• No difference: both versions have had enough views and their take rates are all but the same.

You decide when a test ends: keep the version you want, and switch the test off.

Customers

Insights › Customers lists the shoppers who bought an offer, with their lifetime value and order count from Shopify, when they became a customer, and what they accepted and where. Tap a name to open their Shopify profile, or search by name or email. Names and emails come from Shopify each time you open the page; Luv Upsell doesn't keep them. Erase data on a customer's card removes everything Luv Upsell holds about them.

Orders

Insights › Orders lists the orders that carried an upsell, newest first, with what was added and where. Tap an order to open it in Shopify.

Bought together

Insights › Bought together shows the products your customers buy in the same order, learned from your last 60 days of orders: each pair, how often (the share of orders with the first product that also have the second) and lift (how many times more often that happens than chance). Best partner offers pick from these. The table refreshes itself every day, you can refresh it yourself every 30 minutes, and Create funnel turns any pair into a draft funnel.

Questions? hi@customerluv.app — a founder answers every message.