Most stores do not have an analytics problem. They have four analytics tools that disagree, nobody sure which one is right, and a weekly meeting spent reconciling numbers instead of deciding anything. The fix is deciding what each tool is for before you install it — and knowing that the one number you can trust is already in Shopify.
What follows is the stack we set up on client stores: Shopify as the source of truth, one behaviour tool, one attribution layer, and the server-side tracking that stops the ad platforms optimising blind.
Summary
- Shopify reports — revenue, orders, sessions by channel, cohorts. The source of truth. Free.
- GA4, installed properly — behaviour, landing pages, channel detail. Use a GA4/GTM app so events are consistent.
- Polar Analytics (and Triple Whale, Lifetimely in the same category) — blended ad performance, contribution margin, LTV.
- Server-side tracking — your own GTM container to Meta CAPI and Google. Not an app; an implementation.
1. Start with what Shopify already tells you
Shopify’s analytics cover more than most brands use: sales by product, variant, channel and location; sessions by referrer and landing page; conversion rate broken into added-to-cart, reached-checkout and converted; customer cohorts with repeat rate over time; and — on the higher plans — custom reports and ShopifyQL.
Two reports are worth a standing appointment. Cohort retention, because repeat purchase rate is the number that decides whether a retention programme is working (it is how True Glory Hair’s +27% was measured). And conversion by landing page, because it tells you which pages deserve the CRO work before any test is designed.
2. GA4, and why it never matches
GA4 is still worth having for landing-page and channel behaviour and because the ad platforms read it. Install it through a dedicated app such as GA4 Google Analytics 4 ‑ GTM rather than pasting tags into the theme, so purchase, add-to-cart and checkout events fire consistently and a theme update does not silently break them.
It will not match Shopify. Different attribution windows, different session definitions, browser blocking, and a checkout that used to be partially invisible to it. Reconcile monthly, expect a gap, and never average the two — Shopify is revenue, GA4 is behaviour.
3. The attribution and profit layer
Polar Analytics is the one we most often see doing this job well: it unifies Shopify, the ad platforms, the email tool and your costs into one dashboard, with blended ROAS, contribution margin, LTV by cohort and a survey-style post-purchase attribution option. Triple Whale and Lifetimely occupy the same space with different emphases — Triple Whale on media buying, Lifetimely on LTV and profit.
All three cost real money and none of them is worth installing before the basics are right: if your events are inconsistent and your costs are not loaded, a unified dashboard just makes the wrong numbers prettier. Install one, load real product costs and shipping costs, and use it to answer one question — which channel is profitable at the margin — rather than as a second general-purpose reporting tool.
4. Server-side tracking: the part that changes results
Browser pixels lose events to ad blockers, privacy settings and cross-domain journeys. Server-side tracking routes the conversion through your own tag-manager container to Meta CAPI and Google, enriched and deduplicated on event_id, so the platforms see more of what actually happened and optimise toward it. Hosting runs from about $20 a month; our server-side tracking guide compares Stape, TAGGRS, Addingwell and Google Cloud at your volume.
It matters most when a key conversion completes somewhere other than your storefront. On Gene Juarez, bookings completed inside Zenoti on Zenoti’s domain, and its native pixel fired straight to Meta — bypassing the tag manager entirely. Routing those events through the brand’s own server container moved event match quality from about 3.3 to 7.5 out of 10, which is the practical ceiling without passing personal data, and made thirteen locations’ worth of ad spend measurable.
This is implementation work, not an app install: a container, a data layer that fires the right events with the right identifiers, consent handling, and testing. Elevar is the common productised option; we build it directly as part of Plus development or a custom integration when the journey is unusual.
Quick comparison
| Layer | Tool | Answers | Cost | Install when |
|---|---|---|---|---|
| Source of truth | Shopify reports | Revenue, orders, cohorts, funnel | Free | Always |
| Behaviour | GA4 via a GA4/GTM app | Landing pages, channels, paths | Free + app fee | Always |
| Profit and attribution | Polar Analytics (or Triple Whale, Lifetimely) | Blended ROAS, margin, LTV | Subscription | Past ~$1m, with real costs loaded |
| Event delivery | Server-side GTM to Meta CAPI / Google | Match quality, recovered events | Implementation | Paid spend matters, or conversion happens off-domain |
| Session replay | Hotjar, Microsoft Clarity | Why a page fails | Free tiers | During a CRO sprint, then remove |
What to measure weekly
- Revenue, orders and AOV from Shopify — not from the dashboards.
- Conversion rate split into add-to-cart, reached-checkout, converted, so you know which step moved.
- Repeat purchase rate and the 90-day cohort curve.
- Blended contribution margin: revenue minus COGS, shipping, payment fees and ad spend.
- Core Web Vitals on your top landing pages — slow pages spend the ad budget twice. Check with the free Store Speed Test.
The three mistakes we find most
- Four tools, no definitions. Write down which tool owns which number before installing anything else.
- Costs never loaded. A profit dashboard without product costs is a revenue dashboard with extra steps.
- Tracking that breaks silently. A theme change removes a data-layer event and nobody notices for a quarter. Put a monthly check in the calendar, or let the performance retainer own it.
If the numbers disagree and you need one person to settle it, the free CRO audit covers tracking alongside the storefront and says what is actually broken.
FAQ
Are Shopify’s built-in reports enough?
For orders, products, sessions by channel and cohort retention, yes — and they are the source of truth for revenue. What they do not give you is cross-channel attribution, blended ad performance, contribution margin, or LTV by first-product cohort, which is why brands add a platform on top.
Do I still need Google Analytics 4 on Shopify?
Most stores keep GA4 for channel and landing-page reporting and because ads platforms read it, but it is no longer the system of record for revenue — Shopify is. Install it through a GA4/GTM app so events fire consistently, and reconcile GA4 revenue to Shopify monthly rather than expecting them to match.
What is server-side tracking and do I need it?
It routes conversion events through your own server container to Meta and Google instead of relying on the browser pixel, which browsers increasingly block. It matters most when a key conversion completes off your domain — a booking system, for example. On a 13-location salon group, routing bookings through a server container moved event match quality from about 3.3 to 7.5 out of 10.
Why do Shopify, GA4 and Meta all report different numbers?
Different attribution windows, different definitions of a conversion, and different loss to blocking. Pick Shopify as the revenue truth, use GA4 for behaviour, use the ad platforms for optimisation, and never average them.
