Most conversion advice starts with a list of apps. That is backwards. An app fixes one kind of problem, and until you know which step of your funnel is losing buyers, you cannot know which kind you have. Every tool in this guide is free, and they are used in a fixed order: confirm the numbers are real, find the weak step, see why buyers leave it, then fix that step and nothing else.
Step 1: Check that your numbers agree
Before reading any conversion rate, compare two counts for the same date range:
- Orders in Shopify (Orders, filtered to the dates). Shopify records these on its own servers, so this is the true figure.
- Purchases in GA4 (Reports, Monetisation, Ecommerce purchases), or purchase events in Meta Ads Manager.
If Shopify shows 100 orders and GA4 shows 80 purchases, a fifth of your buyers are invisible to GA4. Any conversion rate GA4 reports for that period is too low, and any test you run against it is judged on partial data. Shopify’s own reports are not affected, because they do not depend on a browser tag.
What to do with the result:
- Within a few percent: carry on to step 2.
- A large gap in GA4: check that GA4 is connected through Shopify’s Google & YouTube channel, then see the GA4 setup guide.
- A large gap in Meta or TikTok: the browser pixel is being blocked. A server-side connection sends the purchase from Shopify directly. CAPI Shield is a free Make.com scenario that does this for Meta, and the server-side tracking guide explains the approach for each platform.
This step does not change your sales. It changes whether you can tell if anything else you do has worked.
Step 2: Find the weakest step in Shopify’s funnel report
In Shopify admin, open Analytics → Reports, filter the Category to Behavior, and open Conversion rate breakdown. It is included with every plan and shows four steps:
| Step | What it counts |
|---|---|
| All sessions | Every visit in the period |
| Sessions with cart additions | Visits where something was added to the cart |
| Sessions that reached checkout | Visits that opened checkout |
| Sessions that completed checkout | Visits that ended in an order |
Each step is shown as a share of all sessions. The number that matters is the drop between two neighbouring steps, so work it out yourself:
Example: 10,000 sessions, 600 with cart additions, 300 reached checkout, 150 completed. Session to cart: 600 ÷ 10,000 = 6%. Cart to checkout: 300 ÷ 600 = 50%. Checkout to order: 150 ÷ 300 = 50%.
In that example, half of the people who open checkout leave without paying. That step is where the money is, because those visitors have already chosen a product. Raising checkout completion from 50% to 60% adds 30 orders a month without a single extra visitor.
Compare each step against your own previous months, not against another store. A step that has fallen since last quarter is a stronger signal than a step that looks low against a figure from someone else’s niche.
Set the report to compare mobile with desktop. If one device converts at a very different rate on the same step, the problem is usually specific to that device’s layout, not to your offer.
Step 3: Watch the sessions that leave that step
The funnel tells you where buyers leave. It cannot tell you why. For that, watch them.
Microsoft Clarity records sessions and builds heatmaps. Its Shopify App Store listing shows one plan, Free Forever, and installing it through that listing avoids editing theme code.
Once it has collected a week of sessions:
- Filter recordings by the page that matches your weak step (a product page,
/cart, or checkout). - Add Clarity’s rage clicks and dead clicks filters. These pick out visitors who tapped something repeatedly or tapped something that did nothing.
- Watch ten recordings on the device that converts worse.
Ten recordings is enough to see a pattern. Write down what each visitor did immediately before leaving. Common findings are a button below the fold on small screens, a size or variant picker that does not register a tap, or a delivery cost that only appears at the last step.
GA4 adds a second view. In GA4, open Explore → Funnel exploration and build the steps view_item, add_to_cart, begin_checkout, purchase. Break it down by Session default channel group to see whether one traffic source is dragging the step down. Paid social visitors often behave very differently from search visitors on the same page.
Step 4: Apply the fix that matches the step
Fix only the step you identified. Changing several steps at once makes it impossible to tell what worked.
If buyers leave before adding to cart
The product page is not answering their question. Check, in this order:
- Price and delivery cost visible near the add-to-cart button. Uncertainty about the final cost stops people adding.
- Reviews. Shopify’s own Product Reviews app is no longer available to install (its App Store page says so, checked 15 October 2026), so older guides that recommend it are out of date. Several reviews apps have free plans; check the plan’s monthly review-request limit before choosing. Ask recent customers for a review in your post-purchase email to build them up.
- Pre-purchase questions. If recordings show visitors scrolling up and down looking for delivery times or returns information, put it on the page. A chat widget can help too: Tidio has a free plan, but on its pricing page that plan covers 50 conversations a month and only the first 50 conversations with its Lyro AI agent, ever. Treat the free plan as a trial of whether chat helps, not as a permanent answer.
If buyers add to cart but do not reach checkout
The cart is creating doubt or friction:
- A surprise in the cart. Delivery cost or minimum order rules appearing for the first time is the classic cause. Show delivery cost, or a free-delivery threshold, on the product page.
- A slow or broken cart drawer on mobile. Recordings show this immediately.
- Discount code fields. An empty discount field prompts some buyers to leave and search for a code. If recordings show this, collapse the field behind a link.
If buyers reach checkout but do not order
Shopify’s checkout is hosted by Shopify, so most stores can change less here than they expect. What you can change on every plan:
- Express payment buttons (Settings → Payments). Turn on Shop Pay, Apple Pay, Google Pay and PayPal where available, so a returning buyer can skip typing.
- Branding (Settings → Checkout, then customise). Your logo and colours make checkout look like part of your store rather than a different site.
- Policies (Settings → Policies). Fill in refund and shipping policies; Shopify links them in the checkout footer, so buyers do not leave checkout to look for them.
Then recover the buyers who still leave. Shopify sends abandoned checkout emails through the Shopify Messaging app, using the automation Abandoned checkout emails by Shopify, where you set who receives it and how long after abandonment it sends (Shopify Help Center). Shopify Messaging includes 10,000 free emails a month, then charges from $1 per 1,000 (App Store listing). For a multi-email sequence with different content in each message, see the free abandoned cart recovery guide.
Step 5: Check speed only if recordings point to it
Slow pages hurt conversion, but speed is often blamed for problems that recordings show are something else. Look at it when recordings show visitors leaving before the page finishes loading, or when one device type converts far worse than the other.
- PageSpeed Insights reports Largest Contentful Paint, Interaction to Next Paint and Cumulative Layout Shift for your real visitors, where Google has enough data. Google’s Core Web Vitals documentation explains the thresholds.
- Shopify Theme Inspector for Chrome draws a flame graph of how long each part of your Liquid theme takes to render, so you can see which section is slow.
The fastest fix is usually removing apps you no longer use. Many apps add scripts to every page, and uninstalling an app does not always remove code it added to your theme.
How to know if a change worked
Run one change at a time and compare the same funnel step for the same number of days before and after, on the same device. Two weeks either side is a reasonable minimum for most small stores, because daily order counts swing a lot. If your numbers do not agree with Shopify’s order count (step 1), use Shopify’s report to judge the change, not GA4.
Record each change and its date somewhere simple, such as a spreadsheet. After a few months, that record of what worked on your store is worth more than any benchmark.
What this costs
Every tool in this guide is free at the time of writing: Shopify’s reports, Shopify Messaging up to 10,000 emails a month, Microsoft Clarity, GA4, PageSpeed Insights and Theme Inspector. Tidio’s free plan is free but limited, as described above. Prices and plan limits were read on each vendor’s page on 15 October 2026.
