A Google Analytics alternative without cookies: Seonavo and GA4 compared

01Side by side

The tinted cell marks the stronger side on that row. Google changes GA4 regularly, so check anything you plan to rely on against its own documentation.

02Cookies, consent and IP addresses

GA4 sets first-party cookies to recognise a returning browser. In places where consent is required before such cookies are set, that means a banner, and visitors who decline are either not measured or are estimated by modelling.

Seonavo sets no cookies at all. To tell one visitor from another within a day it computes a one-way identifier from the site, the date, the IP address and the browser, mixed with a secret key. The IP address itself is never written to disk and the identifier changes every day, so it cannot follow anyone from one day to the next or from one site to another.

The cost of that design is real: Seonavo cannot tell you how many of this week’s visitors also came last week. Someone who returns tomorrow is counted as a new visitor. If returning-user analysis matters to you, that is a reason to keep GA4.

Whether you need a banner depends on your country and on everything else running on your site, not on analytics alone. Seonavo does not add a reason to show one.

03Reading the numbers

GA4 gives you a set of standard reports and a tool for building your own. It can answer almost any question, provided you know how to ask it. Many site owners open it, see a wall of tables, and close it again.

Seonavo starts from the other end. The overview opens with sentences:

Below the summary it lists growth opportunities: a specific page, what to change, and roughly what the change is worth, worked out from your last 30 days. The same summary arrives by email once a week. The tables are still there when you want them.

04Scroll depth, rage clicks and AI traffic

Scroll depth. GA4’s enhanced measurement records a single scroll event, when a visitor gets 90% of the way down. To see 25%, 50% and 75% you add your own events, usually through Google Tag Manager. Seonavo measures the deepest point reached on every pageview, reports the average per page, and separately records whether the bottom of the page was actually reached.

Engaged time. Both tools try to count attention rather than an open tab. GA4 reports average engagement time while the page is in focus. Seonavo counts time only while the visitor is scrolling, moving, touching or typing.

Rage clicks. Three or more quick clicks in the same spot usually mean something looks clickable and is not. Seonavo reports them with the page and the element. GA4 has no such report; Microsoft Clarity does, and can be run next to either tool.

AI assistants. Visits from ChatGPT, Gemini, Perplexity, Claude and Copilot land in GA4’s referral list unless you define a custom channel. Seonavo groups them into an AI channel on its own. The guide to ChatGPT traffic explains how those visits are recognised and why the count is a minimum.

05Setup and page weight

GA4 setup means creating a property and a data stream, placing the Google tag, and for anything beyond the defaults, configuring events, often in Tag Manager. Seonavo is one line before </head>; goals are an attribute on the button you care about.

The Seonavo script is under 4 KB compressed and loads with defer, after the page. The Google tag is many times larger. On a fast connection neither is noticeable; on a slow phone the smaller script is kinder.

06When to keep Google Analytics

• You run Google Ads. Conversion import, attribution and remarketing audiences depend on GA4 or the Ads tag.
• You need revenue reports. Purchases, product performance and funnels across a checkout.
• You analyse returning users or cohorts. Seonavo’s daily identifier rules this out by design.
• You need more than 90 days of history or a raw export for your own warehouse.
• You need country or city reports. Seonavo does not record location.

07Running both

The two scripts do not interfere. A common arrangement is GA4 for advertising and long-term history, and Seonavo for the weekly read on what changed and which pages lose people. If you do this, compare trends rather than totals, because the two count visitors differently.

08Common questions