In short
Learning goal: by the end of this lesson you can read your store's funnel, see at which step most shoppers leave, and choose the one step to work on first.
"Why are people not buying from my store?" is really four questions. Do people arrive? Do they find a product they want? Do they put it in the cart? Do they finish checkout? Each step loses some shoppers, and every store loses most of them somewhere. The useful question is where your store loses more than it needs to, and whether that's something you can fix this month.
The funnel in four steps
A shopping funnel follows visits from arrival to order. Shopify's own analytics counts sessions, sessions with cart additions, sessions that reached checkout and sessions that completed checkout. Most funnel reports, ours included, use the same idea.
| Step | What it counts | Rate to watch |
|---|---|---|
| Sessions | Visits to your store | None: this is the base |
| Added to cart | Visits where something went into the cart | Added to cart per 100 sessions |
| Checkouts started | Visits that reached checkout | Checkouts per 100 sessions, and per cart |
| Orders | Visits that ended in an order | Orders per 100 sessions (your conversion rate) |
What to measure at this step
This lesson is about diagnosis, so most of what you look at is a diagnostic: it explains where the problem is, but it doesn't decide whether a change worked.
| Metric | Definition | Role |
|---|---|---|
| Orders per visitor (conversion rate) | Orders ÷ visitors | A good primary goal for most changes |
| Revenue per visitor | Revenue ÷ all visitors, buyers or not | A good primary goal when a change affects order size |
| Added to cart per 100 sessions | Cart additions ÷ sessions × 100 | Diagnostic: shows whether product pages persuade |
| Checkout completion | Orders ÷ checkouts started | Diagnostic: shows whether checkout loses people |
| Device split | Each rate for phones and for computers | Diagnostic: shows whether one device has a problem |
The difference matters later in the course. A change that gets more people to add to cart but not more people to buy didn't help. Lesson 3 comes back to this.
How to read it
Look at the biggest drop you can do something about
Every funnel narrows sharply at the first step: most visitors look and leave. That's normal browsing, and Baymard's research on why shoppers abandon purchases excludes it for the same reason (42% of US online shoppers said they were "just browsing", page updated 22 September 2025). So don't fix the biggest drop by default. Look for the drop that's bigger than it should be for a reason you can name.
Compare devices before anything else
Split the funnel by phone and computer. If computers turn carts into orders at twice the rate of phones, the problem is probably something on the phone: a slow page, a popup over the button, a checkout field that's hard to fill in. Device differences are the most common clue a small store can act on.
Compare with yourself, not with averages
You'll find "average Shopify conversion rates" online. They mix stores selling $5 stickers with stores selling $900 sofas. Your own funnel last month is a better benchmark than any of them.
Common problems, by step
Few carts per 100 sessions. The product page isn't answering the shopper's questions: price, delivery date, returns, what's in the box, what it looks like in use. Or traffic is arriving on the wrong page, such as an ad that promises one thing and lands on another. Lessons 2 and 3 cover these.
Carts that don't become checkouts. A surprise in the cart, usually shipping. Baymard lists "extra costs too high (shipping, tax, fees)" as the most common reason for abandoning a purchase, at 40% of shoppers who were not just browsing (same page). Lesson 4 covers the cart.
Checkouts that don't become orders. Payment methods missing, a forced account, a confusing delivery choice, or a broken step. Lesson 5 covers what you can control on a standard Shopify plan.
Test ideas, and why they'd work
At small-store traffic most of these are fixes, not tests. Lesson 7 shows how to tell the difference.
| Idea | Mechanism |
|---|---|
| Show the shipping cost or threshold on the product page | Removes the surprise that makes carts stall |
| Delivery date under Add to cart | Answers the gift buyer's main question before they leave to check |
| Move a popup off phone screens | Uncovers the button the shopper was about to press |
| Show express payment buttons in checkout | Saves typing on a phone, where typing is slowest |
A store's funnel for one week: 2,000 sessions, 140 added to cart (7.0 per 100), 60 checkouts started (3.0 per 100) and 36 orders (1.8 per 100). Checkout completion is 36 ÷ 60 = 60%.
The owner wonders whether to test a checkout change or a product page change.
- A checkout change judged on checkout completion: to detect a 15% improvement (60% to 69%) at 95% confidence and 80% power needs 443 checkouts per version. At 60 checkouts a week, that's 15 weeks.
- A product page change judged on orders per session: to detect a 15% improvement (1.8% to about 2.07%) needs 40,860 sessions per version. At 2,000 sessions a week, that's 41 weeks.
Computed with sampleSizePerVariant from Benchmyrk's statistics engine. The store is illustrative.
Two lessons from the example. First, a step with a high rate (checkout completion) needs far fewer people per version than a step with a low rate (orders per session), but there are also far fewer people at that step. Second, at 2,000 sessions a week both tests take months. The sample-size calculator does the same sum for your own store. That's why the first job is finding problems you can simply fix, which is what the rest of this course is about.
A 10-minute exercise on your own store
- Open your analytics for the last 30 days (Shopify's, or Benchmyrk's if you use it). Write down four numbers: sessions, sessions with an add to cart, sessions that reached checkout, orders.
- Work out three rates: carts per 100 sessions, checkouts per 100 sessions, and orders ÷ checkouts.
- If your analytics can filter by device, do the same for phones only and computers only.
- Circle the step where phones fall furthest behind computers. If none does, circle the step where you know of a reason shoppers might stop.
- Place a test order on your phone through that step. Note anything slow, unclear or broken. That's your first fix.
You don't need any tool beyond Shopify for this.
Do it in Benchmyrk
The Black Friday series for small stores starts with the same idea, fix first and test second: see Black Friday for a small Shopify store.
Check yourself
- Your carts per 100 sessions are the same on phones and computers, but phones complete checkout at half the rate. Where do you look first? Answer: checkout on a phone. Place a test order on your phone and look for missing express payment buttons, hard-to-fill fields or a slow step.
- A change raised added-to-cart per 100 sessions but orders per session didn't move. Did it work? Answer: not yet, as far as you can tell. Add to cart is a diagnostic. The change only helped if more visitors bought or spent more.
- Why is your own funnel last month a better benchmark than an industry average? Answer: averages mix stores with very different products, prices and traffic. Your own numbers share your products, prices and customers.
Sources
- Shopify developer docs, ShopifyQL sessions schema (sessions with cart additions, reached checkout, completed checkout), read 12 October 2026
- Baymard Institute, Cart abandonment rate statistics (updated 22 September 2025), read 12 October 2026
- Shopify Help Center, Placing a test order, read 12 October 2026