The trap is small changes. A small store doesn't get enough visits to show whether a small change worked. That new headline you tried last Tuesday may have done nothing at all, and you'd never know.
"Use industry best practices that were proven, but stop wasting resources tweaking things you can't evaluate!" Ron Kohavi, who led Microsoft's online testing team
He wrote that for startups, but it fits a shop with a few thousand visits a month just as well. So that's what you'll find here: proven fixes, in the order that gets you money soonest, and a simple way to see which ones paid off.
What your conversion rate means
Your conversion rate is the share of visits that end in an order. Divide your orders by your visits, then multiply by 100. Say you had 40 orders from 3,000 visits last month. That's 1.3%, or about 13 orders for every 1,000 visits.
That's also about normal. Littledata tracked 438 Shopify stores, and the store in the middle got 13 orders per 1,000 visits. The top fifth got 27 or more. The bottom fifth got 5 or fewer. Every one of those stores made about one sale a day or more, so a newer shop should expect less.
What you sell changes the picture. Skincare stores in that group got about 26 orders per 1,000 visits. Clothing and furniture stores got about 10. So 10 is fine for a sofa shop and low for a face cream shop. Before you judge your own number, look up the average ecommerce conversion rate for your kind of product.
After that, stop comparing. The number to beat is your own from last month.
One warning for Shopify owners. Shopify changed how it counts visits in September 2026. If your shopify conversion rate jumped or fell around then, your customers may not have changed at all. The counting did.
What one fix is worth
Work this out before you give up a weekend. Say you get 3,000 visits a month and your average order is $60.
- At 13 orders per 1,000 visits, you get 39 orders.
- At 16 orders per 1,000 visits, you get 48 orders.
- That's 9 more orders, or $540 more in sales every month, with nothing extra spent on ads.
Now try your own store. Take your monthly visits, divide by 1,000, multiply by 3, then multiply by your average order. That's what 3 more orders per 1,000 visits would bring in each month. If the answer is $40, go get more visitors first. If it's $4,000, clear your afternoon.
Is your store big enough for this?
Your traffic decides what's worth doing. Find your row:
| Visits a month | What to do |
|---|---|
| Under 1,000 | Mostly get more of the right visitors. Do the test purchase and watch five people buy. Your numbers are too small to tell you much yet. |
| 1,000 to 10,000 | Everything in this guide. Fix what's clearly wrong and ask shoppers what stopped them. Skip A/B tests. |
| 10,000 to 60,000 | Everything in this guide. You can also test big changes, like a whole new product page. |
| Over 60,000 | Everything in this guide, plus regular testing of one change at a time. |
An A/B test shows two versions of a page to two groups of visitors at the same time, then counts who buys. It's the most reliable way to learn what works. It also takes a lot of visits.
Say a new page would truly bring you 20% more orders. At 13 orders per 1,000 visits, this free calculator says you'd need about 30,000 visits on each version. That's 60,000 visits in all. At 5,000 visits a month, one test takes a year.
Kohavi gives a second reason to stick with proven fixes. Even at companies that test properly, about 8 in 10 ideas don't improve the number they were meant to improve. Your clever idea probably has the same odds. A broken checkout is a safer bet, and you don't need a test to fix one.
Find the step where shoppers give up
Shoppers usually give up at one of three points: the product page, the cart or the checkout. Your reports show you which.
- On Shopify: Go to Analytics, then Reports. Click the Category filter, choose Behavior and open Conversion rate breakdown.
- On other platforms: Google Analytics can show this if it's set up to track your store. Go to Reports, expand Monetization and click Purchase journey.
You want the number of visits, the number that added to cart, the number that started checkout and the number that bought. Then find your row. The typical figures come from the same 438 stores.
| What you see | What it usually means | Start with |
|---|---|---|
| Under 3 in 100 visits add to cart (typical is about 7) | The page doesn't convince them, or they were never real shoppers | Fixes 6 and 7, and check who's visiting |
| Plenty add to cart, few start checkout | The total put them off | Fix 2 |
| Under 26 in 100 who start checkout finish it (typical is about 47) | Something in the checkout stops them | Fixes 1, 3, 4 and 5 |
| All of this looks fine, but sales are low | You need more visitors | Stop here and work on traffic |
To check who's visiting, look at where your visits come from. Someone who searched for your product was already shopping. Someone who tapped a video ad was watching videos a moment ago. And visitors from a country you don't ship to can't buy at all, however good your page is.
Fix it in this order
Start with the step your report points to. If your numbers are too small to read, start closest to payment, since those shoppers have already decided to buy.
1. Buy from your own store
Do it on your phone, since about two out of three visits to a typical store come from one. Use a real card, buy your cheapest product and refund yourself afterward. Some fees may not come back, so think of it as a very small consulting bill.
Look for anything sold out by mistake, a country you forgot to ship to, a discount box that errors, or a page that jumps around under your thumb. On Shopify, also go to Orders, then Abandoned checkouts. A failed payment usually doesn't create an order, so it lands there. You may find people who tried to pay you and couldn't.
2. Show the full cost early
Baymard, a research firm, asked US shoppers why they'd left a cart. Set aside the ones who were just browsing. Of the rest, 4 in 10 said extra costs were too high: shipping, tax and fees. Nothing else came close. Slow delivery was next, at 2 in 10.
So stop saving the bad news for the end:
- Put the shipping cost and delivery date on the product page.
- Or build shipping into your prices and say so.
- If you offer free shipping over a set amount, set it a little above your average order. Check that you still make money at that amount.
A line like this does the job: Free shipping over $75. Order today and it arrives by Friday.
3. Let people buy without an account
Nearly 1 in 5 shoppers in that same survey left because the site wanted them to create an account. A first-time buyer doesn't want a relationship yet. They want the candle.
- On Shopify: Go to Settings, then Checkout. Under Customer contact method, turn off "Require customers to sign in to their account before checkout." While that's on, Shopify also hides fast options like Apple Pay from the cart.
- On WooCommerce: Go to WooCommerce, then Settings, then Accounts & Privacy, and turn on guest checkout.
If you want accounts, offer one after the order is placed.
4. Cut the boxes
The average US checkout shows about 15 boxes to fill in. Baymard says 7 or 8 can be enough. Count yours, on your phone, with your thumb.
On Shopify, go to Settings, then Checkout, then Customer information. Set Company name to "Don't include" unless you sell to businesses. Set Address line 2 to "Optional." Set the phone number to "Optional" too, unless your delivery company needs it.
Tuck the coupon box behind a small link as well. Baymard has watched shoppers spot that empty box, feel they're overpaying and go off hunting for a code.
5. Offer a fast way to pay
Typing a card number on a phone is slow and annoying. Turn on Apple Pay, Google Pay, Shop Pay or PayPal, whichever your platform supports. Pick the few your buyers really use, because a long row of logos is hard to read on a small screen.
6. Make the top of the product page do its job
Leave the button color alone. You'll be tempted anyway. It takes ten minutes and feels like work, and it's the store owner's version of cleaning the garage when the taxes are due.
Open your best product page on your phone instead. Show a friend the first screen for five seconds, then take the phone away. Can they say what it is, who it's for and what it costs? If not, fix that first.
Then fix the rest of the page:
- Photos. In Baymard's testing, more than half of shoppers went straight to the photos. Show the product from every side, in use, with something next to it for size.
- Words. "Organic cotton, 200 GSM" says what it is. "Still soft after 50 washes" says why they'd want it, as long as it's true.
- Questions. Search your inbox. If people ask the same thing twice, put the answer on the page.
7. Give strangers a reason to trust you
Shoppers don't know you. They've got your page, a small doubt and a card they'd like to keep safe.
Start with reviews. Northwestern's Spiegel Research Center found shoppers were far more likely to buy a product with five reviews than one with none. Reviews after the first five helped much less. So you don't need hundreds. Email your past buyers:
Hi Sam, how's the grinder working out? If you have a minute, an honest review would really help a small shop like mine. Good or bad is fine.
Show the average rating and the number of reviews together, next to the product name. Don't buy reviews or write your own.
Then cover the basics. Use your own web address, a real contact email and an About page with a person on it. Put links to shipping and returns at the bottom of every page, because Baymard found some shoppers head straight there.
8. Make the page load faster
Run your best product page through Google's free PageSpeed Insights tool, and look at the mobile score. The usual culprits are huge photos and apps you stopped using months ago. Shrink the first and delete the second. Then leave speed alone and go back to what your shoppers are telling you.
Ask the people who know
That list covers the common problems. Your own shoppers can tell you which parts apply to you. Here are three ways to ask.
Watch five people try to buy. Pick people who could really be customers. Have them use their own phone and say, "Buy something you'd actually want, and tell me what you're thinking." Then keep quiet, which is the hard part. Researcher Jakob Nielsen's long-standing advice is that after the fifth person you mostly see the same problems again.
Ask your buyers. Add one line to your order email: What almost stopped you from buying? They've just bought, so their memory is fresh.
Ask the ones leaving. A survey app can put one question on your busiest product page, shown after someone has been there a while: What's stopping you from buying today? Give them answers to tap: price or shipping, not sure it's right for me, need it sooner, not sure I trust this store, just looking.
Microsoft Clarity, a free tool, will also record real visits. Watch ten visits from people who added to cart and then left. You'll probably wince by the third one.
Some of those answers need a reply right there on the page. "Not sure it's right for me" is the hard one, because a page says the same thing to everyone and every shopper's doubt is different. That's what we built Possible for, so yes, this is our own product. Possible is an AI built on real sales psychology. It asks each visitor the right question at the right moment, remembers the answers and guides them to a sale. It's free to start, and you can read more about our AI sales assistant for online stores. If you sell three products and the choice is obvious, finish the fixes above first.
Watch your profit as well as your rate
A higher conversion rate can lose you money. Here's a made-up store that sells a $50 product and makes $20 on each one, trying 20% off for a month:
| Normal month | Month with 20% off | |
|---|---|---|
| Visits | 3,000 | 3,000 |
| Orders | 39 | 60 |
| Conversion rate | 1.3% | 2.0% |
| Sales | $1,950 | $2,400 |
| Profit | $780 | $600 |
The rate went up by about half. Sales went up too. The owner shipped 21 more boxes and made $180 less. That's because 20% off a $50 product takes $10 out of the $20 profit, so you have to sell twice as many just to stay level.
Your rate can also rise for dull reasons. Turn off an ad that brought in lots of window shoppers and your rate climbs, with not one more order to show for it. Returns can fool you as well. A page that sells a lot but gets a lot sent back is probably promising too much.
So keep one more number next to your rate: profit per 100 visits. Take your profit, after refunds, divide by your visits and multiply by 100. The store above went from $26 to $20. That's the number that pays you.
How to tell if a fix worked
Write down the date of each change. Make one big change at a time. Then compare four full weeks before with four full weeks after, and leave out any weeks with a sale or a holiday. Look at the step that fix should move. A checkout fix should show up in finished checkouts first.
Don't expect to see every fix. With 40 orders a month, plain luck can move you ten orders either way. So going from 40 to 46 tells you very little.
If you fixed something plainly broken, keep the fix and move on. You don't need proof that a working checkout beats a broken one. One clear problem, one fix and one honest look at the numbers is all that conversion rate optimization really comes down to. On Shopify, shopify conversion rate optimization follows the same steps, with its own reports and settings.
What to skip for now
- A full redesign. Unless your store looks broken, it rarely fixes the step your report points to. It also ruins your before and after numbers.
- Another app. Each one adds a monthly bill and more code to load. Add one when it solves a problem you've measured.
- Fake countdowns. A timer that resets on every visit gets noticed, and it makes the rest of your page harder to believe.
- Chat boxes that pop up unasked. In Baymard's testing, shoppers found them annoying. They did find chat useful when they had a specific question.
- Paid testing tools and agencies. Wait until your row in the table above says testing makes sense.
Your first two months
- Week 1: Buy from your own store and fix what gets in your way. Find your step in the report. Work out your profit per 100 visits, so you have a starting point.
- Week 2: Make the fixes for your step. Most of them are settings, and they take minutes.
- Week 3: Watch five people try to buy. Add the question to your order email. Fix the top of your best product page.
- Week 4: Email past buyers for reviews. Add the on-page question. Write down what you changed.
- Weeks 5 to 8: Leave things alone while the numbers build up. Then compare, and start again with your next worst step.
Questions about ecommerce conversion rate optimization
How can I improve my ecommerce conversion rate?
Find the step where shoppers give up, then fix that step first. For most small stores, the quickest wins are showing shipping costs early, letting people buy without an account, cutting checkout boxes and getting your first five reviews.
What is a good conversion rate for ecommerce?
About 13 orders per 1,000 visits (1.3%) is typical among established Shopify stores, and 27 per 1,000 (2.7%) puts you in the top fifth. It depends on what you sell. Food and skincare tend to run higher, and clothing and furniture lower.
What is the 80/20 rule in ecommerce?
It's a rule of thumb that most of your sales come from a small share of your products, often put at about 80% from 20%. It isn't an exact law. To use it, sort your products by sales for the last 90 days and fix the pages for your top few first. That's where a better page gets seen the most.
What is the difference between SEO and CRO?
SEO gets more people to your store from search engines. CRO, short for conversion rate optimization, gets more of the people who arrive to buy. You need both, and CRO makes every visitor you get from SEO worth more.
How long does conversion rate optimization take?
The fixes are quick. Many checkout settings take a few minutes. Seeing the result takes longer, so give each big change about four weeks before you judge it.
Do I need an A/B testing tool?
Not until you get about 10,000 visits a month, and even then only for big changes. At 20,000 visits a month, one test can take about three months. Below that, fix what's clearly wrong and ask your shoppers.
Go deeper on conversion tactics.