Using analytics to improve ecommerce conversion rates: a practical guide
Define conversion properly, find where shoppers drop out, use site search and Core Web Vitals data, and run A/B tests that are big enough to mean something.
How GA4, Matomo, Plausible and built-in platform analytics such as Shopify's differ on data hosting, consent, ecommerce reporting, cost, learning curve and integrations.
Choosing an analytics tool is less about features than about trade-offs: depth against simplicity, advertising integration against data control, and how much consent friction you are willing to accept. This comparison looks at Google Analytics 4, Matomo (cloud and self-hosted), Plausible and the analytics built into commerce platforms, using Shopify as the main example. It is based on vendor documentation and publicly available information as of November 2026, not on hands-on benchmark testing; features and prices change, so check the linked pages before deciding.
Google Analytics 4 (GA4) is the default for many shops: no licence fee for the standard version, deep event-based tracking and tight integration with Google Ads and Merchant Center. It is also the most complex of the four, and in the EU it normally sits behind a consent banner. A paid enterprise edition, Analytics 360, is sold on custom terms.
Matomo is an open-source analytics platform available as self-hosted software (Matomo On-Premise) or as a hosted service (Matomo Cloud). Its pitch is GA-like depth with data ownership and configurable privacy settings.
Plausible is a deliberately lightweight, cookie-free web analytics service hosted in the EU, with a free self-hosted Community Edition. It reports traffic, goals and revenue, not the detailed product-level ecommerce reports of the others.
Built-in platform analytics (Shopify Analytics, and the equivalents in other platforms and marketplaces) come with the store. Because the platform processes the orders, its sales figures are the system of record.
| GA4 | Matomo Cloud | Matomo On-Premise | Plausible | Shopify Analytics | |
|---|---|---|---|---|---|
| Hosting | Google infrastructure | EU (Frankfurt, Germany, per Matomo) | Your own server, wherever you choose | EU (per Plausible); self-hosting possible | Shopify’s platform |
| Data control | Google acts as processor; data subject to Google’s terms and settings | Vendor-hosted; customer owns data per vendor | Full control, full responsibility | Customer owns data per vendor | Held within the platform |
| Cookies by default | Yes (first-party) | Configurable, including cookieless | Configurable, including cookieless | No cookies | Platform cookies, governed by its consent tools |
| Consent (EU) | Normally required; consent mode can model gaps | Depends on configuration and country | Depends on configuration and country | Vendor says not needed; check local guidance | Managed through platform consent settings |
| Ecommerce depth | Detailed item, funnel and checkout-step reporting | Ecommerce, product and sales reports; funnels as paid add-on | Same as Cloud; paid plugins for some features | Revenue goals and funnels on higher tier; no product reports | Orders, products, customers, inventory from source data |
| Cost model | Standard: no licence fee; 360: paid | Subscription by monthly hits | Free Community edition; paid bundles and plugins | Subscription by monthly pageviews; free self-hosted edition | Included in platform plan; report types vary by plan |
| Learning curve | Steep | Moderate | Moderate, plus server administration | Low | Low |
| Integrations | Google Ads, Merchant Center, Search Console, BigQuery and many platforms | Plugins, tag manager, APIs | As Cloud, plus direct database access | Stats API, Looker Studio connector | App ecosystem; can connect GA |
GA4. Google says Analytics data is stored in its distributed file system and replicated across data centres, without naming specific regions, and that under the GDPR it acts as a processor for customers (Safeguarding your data). It also states that GA4 does not log or store IP addresses, though they are used in collection. Transfers to the US rely on mechanisms such as the EU–US Data Privacy Framework, whose adequacy decision was upheld by the EU General Court in 2025 and is subject to an appeal before the Court of Justice. Retention for user- and event-level data in standard properties can be set to two or 14 months; this affects explorations and funnel reports, not standard aggregated reports.
Matomo. Matomo states on its pricing page that Cloud data is stored in Frankfurt, Germany. With On-Premise, data sits wherever you host it, which gives full control and full responsibility for security, updates and backups.
Plausible. Plausible’s data policy says data is processed and stored in the EU, that customers own their data and that it is not sold or used for advertising. A self-hosted Community Edition is available under an open-source licence.
Shopify. Analytics data lives inside the platform alongside your orders and customer records, under the platform’s own terms and data processing commitments. That is convenient and coherent, but the data is only as portable as the platform’s exports and APIs allow, which is worth considering if you might move between a marketplace and your own store or change platform later.
In the EU, storing or accessing information on a user’s device generally requires consent under Article 5(3) of the ePrivacy Directive unless it is strictly necessary, and the GDPR applies to any personal data collected. Our guide to cookie consent and analytics for ecommerce covers the rules in detail.
GA4 uses cookies and is normally deployed behind a consent banner in the EU. Consent mode passes visitors’ choices to Google tags; in its “advanced” form, tags send cookieless pings when consent is refused, which Google uses to model missing conversions and behaviour. Modelling estimates; it does not recover the data.
Matomo can be configured to run without cookies and with anonymised IP addresses. Its own FAQ on consent-free use is notably candid: some regulators, including France’s CNIL, allow consent-exempt audience measurement under strict conditions, but Matomo recommends treating JavaScript-based tracking as requiring consent in strict ePrivacy regimes unless a supervisory authority says otherwise. It also warns that ecommerce order IDs and user-ID features can create personal data.
Plausible sets no cookies and, according to its data policy, counts unique visitors with a daily-rotating hash of a salt, the site domain, the IP address and the user agent, discarding the salt every 24 hours. Plausible says this means no consent banner is needed. That is the vendor’s position. The EDPB’s Guidelines 2/2023 take a broad view of what counts as accessing information on a device, which can include scripts instructing a browser to send data, so “cookieless” does not automatically mean “consent-free” everywhere. Check your national regulator’s position.
Shopify provides consent settings and a Customer Privacy API with categories for analytics, marketing, preferences and sale of data, which determine whether analytics processing is allowed for a visitor.
In November 2025 the European Commission proposed changes to the cookie rules as part of its digital omnibus package, including consent exemptions for certain audience measurement. As of November 2026 those proposals were still being negotiated; the current rules apply.
GA4 is the deepest of the web-analytics tools for ecommerce, provided it is implemented properly. Its recommended ecommerce events cover product list views, product views, basket actions, checkout steps, purchases and refunds, feeding item-level reports and funnel explorations. The catch is implementation: events must be sent with correct item data, and consent gaps mean figures will not match your order system.
Matomo offers ecommerce reports covering products and sales, abandoned-basket tracking and multi-channel attribution. Some advanced features, such as funnels and heatmaps, are premium plugins, as its pricing page shows.
Plausible handles ecommerce through revenue goals: a purchase event carries an amount and currency, and you can see conversions, total and average revenue by source, campaign, page, country or device. Its documentation notes there are no dedicated product-level reports (you add custom properties such as a product name instead), that refunds and later payments are not linked back, and that revenue tracking is a Business-plan feature.
Shopify Analytics is strongest where the others are weakest: orders, returns, products, customers and inventory come from the source data rather than from a browser script, so they are not affected by ad blockers or refused consent in the same way. Its reports include a customisable dashboard, detailed reports and a live view; the report types available depend on your plan. It is weaker on behaviour outside the store, such as a separate blog or landing pages on another domain.
As of November 2026:
GA4 has the steepest learning curve: the event model, explorations and attribution settings reward expertise and punish casual use. Its integrations with Google Ads, Merchant Center, Search Console and BigQuery are the main reason advertisers stay with it. Matomo’s interface is closer to traditional web analytics and many find it more approachable; self-hosting adds server administration. Plausible fits on a single dashboard and can be understood in an afternoon. Platform dashboards are the easiest of all because they speak the language of orders and products.
In our assessment, the common mistake is choosing the most powerful tool and then using a fraction of it. Pick the tool whose reports you will actually read, then use our guide to improving conversion with analytics to put it to work. Whatever you choose, remember the data-protection basics in our GDPR guide for online businesses.
This comparison is based on vendor documentation and publicly available information, not hands-on benchmark testing. It is general information, not legal advice.
Define conversion properly, find where shoppers drop out, use site search and Core Web Vitals data, and run A/B tests that are big enough to mean something.
Where AI product copy saves time, where it fails (invented specs, green and health claims), what Google says about AI content, and a review workflow that works.
How ePrivacy, the GDPR, Planet49, EDPB and CNIL guidance and the UK's 2026 PECR changes apply to cookie banners, analytics and Google Consent Mode.