Analytics setup for a Shopify store

Shopify's own analytics answer most trading questions, and merchants routinely install three external tools before reading them. Start with what's already there, add what's genuinely missing, and gate all of it behind consent.

Start with Shopify's own reports

Under Analytics. Depth varies by plan, but even the basics cover:

• Sessions, conversion rate, average order value.
• Sales by product, collection, channel and traffic source.
• The checkout funnel — added to cart, reached checkout, completed. This is the most valuable report in the admin.
• Returning versus new customers.
• Search terms, including ones returning nothing — search setup.

Because it's the commerce platform, the numbers are authoritative in a way a third-party tool's aren't. Where Shopify and GA4 disagree about revenue, Shopify is right.

Adding GA4

Install it through the Google sales channel, not by pasting a tag into your theme. The channel gives you:

• Storefront pageviews.
• Checkout and purchase events, which you cannot get from a theme-level tag — checkout is Shopify's, and script tags there ended with checkout extensibility.
• Consent handling that respects Shopify's customer privacy settings.

A hand-pasted GA4 tag gives you traffic and nothing past add-to-cart, which is the version most stores have and the reason their funnel looks broken.

Consent first

If you serve regions requiring consent before non-essential tracking, analytics must not fire until it's given.

Test it: private window, decline, then check the network tab for requests to Google or Meta. A banner that informs while everything fires anyway is a documented compliance problem rather than a fix.

Every analytics and advertising tool belongs in your privacy policy by name.

Don't install anything twice

The most common measurement error. A store ends up with GA4 via the channel, plus a pasted tag, plus a tag manager, plus an SEO app injecting one.

Duplicate events inflate sessions and conversions and make every downstream decision wrong. Check the rendered page for how many times each measurement ID appears, and keep one.

The same applies to the Meta pixel.

The four numbers to watch

Ignore the dashboards for a moment. These four, split by device, tell you almost everything:

1. Sessions — is traffic growing?
2. Conversion rate — mobile and desktop separately, always. They differ, and the mobile number is usually the problem.
3. Average order value.
4. Checkout completion rate — of the people who started, how many finished?

Whichever is worst relative to the others is where the work is. Redesigning the homepage when the problem is mobile checkout is the most common misdirected effort in ecommerce — conversion design.

Why the numbers disagree

They always will, and the reasons are known:

• Attribution windows differ between tools.
• Consent means analytics sees fewer sessions than Shopify does.
• Bots and previews inflate some tools.
• Currency — presentment versus settlement in multi-currency stores.
• Timezone — Shopify timestamps are UTC; a daily report built without converting disagrees with the admin.

Decide which tool is authoritative for which question. Shopify for revenue, GA4 for traffic sources and behaviour. Write it down, and stop trying to reconcile them to the penny.

Beyond the basics

Worth adding when you have a specific question:

• Session recordings — twenty mobile recordings tell you more than a month of dashboards.
• Search Console for organic performance and Core Web Vitals field data.
• A warehouse, once questions involve joining Shopify data with ad spend or cost of goods — getting data out.

Before you launch

Verify a test purchase appears as a conversion — not just a pageview — in every tool you've installed. Do this before launch, or the first weeks of data are gone and you can't get them back.

Read Shopify's checkout funnel before installing anything. It answers the question most stores buy three tools to avoid asking.