How to measure how well your store turns visitors into orders, from five counts and your revenue. The five counts are sessions, product views, carts, checkouts and orders. From them come the five numbers: conversion rate, average order value, revenue per visit, add-to-cart rate and checkout completion rate, each worked out as one sum. You learn where to find each count in Shopify Analytics and in a GA4 funnel exploration, and how revenue is built from sessions × conversion rate × average order value. You then record a baseline from one date range, with every sum written out, so every later change can be compared against it. The result is cro-baseline.md. Terms explained include session, conversion rate (CVR), average order value (AOV), revenue per visit and baseline.
How to find the exact step where your store loses the most buyers. The lesson lays out the funnel as five counts and four steps, each with its own step rate and the number of visitors who left. You draw the same funnel twice, for phone and desktop, because the two often leak at different steps, and you check which landing pages start sessions that rarely order. Each leak is then matched to one of five reasons visitors leave: Fit, Clarity, Trust, Cost or Effort, with the evidence for it from your support questions, reviews, policies or analytics. The result is leak-map.md, with every step's counts, rate, reason and evidence. Terms explained include funnel, step rate, drop-off, leak, device split and landing page.
How to work out which leak is worth fixing first. For each step, you raise its rate to a target you write down, such as 10% higher or your desktop rate, hold every other step the same, and multiply the extra orders by your average order value. You learn how to choose a target rate, work through a full table for every step on phone and desktop with every sum written out, and see why the values are never added together. The result is an estimated value at the target rate, not a forecast: it ranks the steps so you know where to look first. You then sort the leaks by value, check each has evidence and enough visits, and choose the first leak to fix. The value column is added to leak-map.md. Terms explained include target rate, relative rise, percentage point, benchmark and estimated value.
The six prompts from the SOP & AI Prompts pack, run in order in one Claude conversation on your own analytics export. They build the baseline, map the funnel and step rates by device, collect the evidence and assign a reason to each leak, calculate the estimated value per step, sort the leaks and choose the first one to fix, and check every sum in both files. You leave with cro-baseline.md and leak-map.md, and the one leak to fix first.
The SOP shows its example figures in £ and its prompts round money to the penny. Your own will be in your currency.