How to track conversions without cookies

You shipped a project. People visit. Some of them sign up, most don’t. You want to know the conversion rate, where people drop off, and what’s actually working. But every analytics tool you evaluate requires a cookie banner before you can measure anything.

Most developers accept this as the cost of analytics. They add Google Analytics, bolt on a consent manager, and move on. The cookie banner becomes part of the site’s furniture. But for side projects and indie SaaS, that banner isn’t just clutter. It erodes trust with privacy‑conscious users and adds a dependency you have to maintain.

That’s a false trade‑off. Cookies solve a specific problem: tracking users across multiple websites for ad targeting. If you’re only measuring what happens on your own site, you don’t need them. This guide covers how to set up real conversion tracking with Janus: signups, purchases, multi‑step funnels, and drop‑off analysis. No cookies, no fingerprinting, no consent banners.

Already have Janus installed? Skip ahead to the code examples. New to Janus? Add it in 5 minutes and come back.

That cross‑site tracking is the specific behavior that privacy regulations target. According to the ePrivacy Directive, cookies require user consent unless they are "strictly necessary" for a service the user requested. Analytics cookies don’t qualify. They are useful to you as a site owner, but not necessary for the visitor. That’s why Google Analytics requires a consent banner.

Product analytics has different requirements than ad tracking. You need to know whether someone visited your pricing page, clicked signup, and finished onboarding. All of that happens on your domain. None of it requires cross‑site cookies, browser fingerprinting, or third‑party scripts.

The consent requirement comes from the tracking mechanism, not from analytics itself. localStorage is first‑party storage scoped to your origin. Other websites cannot read it. It is not covered by the ePrivacy Directive’s cookie consent rules because it does not enable cross‑site tracking. GDPR still applies to processing of personal data, but a random anonymous ID in localStorage does not identify a natural person under GDPR’s definition.

Remove the cookies and fingerprinting, and the consent requirement goes with them. Read more about how Janus handles privacy and why it was built this way.

Sessions use a separate sessionId stored in sessionStorage. This ID is scoped to a single browser tab and expires when the tab closes. It tracks multi‑step flows within a visit without persisting anything beyond the session.

Neither localStorage nor sessionStorage is accessible to other websites. They are scoped to your origin, which means your analytics data cannot leak to third parties.

The difference between the two IDs matters for conversion tracking. Use distinctId to measure conversions that span days or weeks: a user visits your landing page Monday, reads docs Tuesday, signs up Friday. Use sessionId for single‑session flows: landing page to pricing to checkout in one sitting.

IP addresses are used transiently for country‑level geolocation via GeoIP lookup and then discarded. They are never written to the database. According to GDPR guidance, transient processing without storage does not constitute personal data collection.

The Janus loader is under 1 KB and loads asynchronously after the page paints. It is designed to have zero measurable impact on Core Web Vitals. No cookies, no fingerprinting, no third‑party data sharing.

<script src="https://addjanus.ca/janus.js"></script>
<script>
    Janus.init('jns_your_api_key_here');
</script>

Page views are tracked automatically, including SPA navigation. For conversions, you call Janus.track() at the moments that matter to your business.

// User completes signup
Janus.track('signup_completed', {
    plan: 'free',
    source: 'pricing-page'
});

// User makes a purchase
Janus.track('purchase_completed', {
    amount: 29,
    currency: 'USD',
    plan: 'pro'
});

// User hits a key activation moment
Janus.track('first_project_created', {
    template: 'blank'
});

Use verb_noun naming: signup_started, checkout_completed, payment_failed, trial_expired. This convention keeps your event log scannable, makes funnel steps easy to define, and simplifies querying via the HTTP API or through your AI coding agent.

Properties in the second argument go into the event’s payload. Put business‑specific data here: plan name, price, variant, referral source. Janus captures context separately and automatically, so don’t duplicate fields like referrer or page URL in your payload. Combined payload and context must stay under 10 KB per event. See the API reference for the full Janus.track() signature and the best practices guide for naming conventions.

function PricingCard({ plan, price }) {
    const handleSubscribe = () => {
        window.Janus?.track('checkout_started', {
            plan,
            price,
            source: 'pricing-page'
        });
        // proceed to checkout...
    };

    return (
        <button onClick={handleSubscribe}>
            Subscribe to {plan}
        </button>
    );
}

// Step 1: User lands on signup page
Janus.track('signup_started');

// Step 2: User submits the form
Janus.track('signup_submitted', {
    method: 'email'
});

// Step 3: User verifies their email
Janus.track('email_verified');

// Step 4: User completes onboarding
Janus.track('onboarding_completed', {
    steps_skipped: 0
});

// Step 5: User creates their first project
Janus.track('first_project_created');

Each event fires independently at the moment it happens. You don’t need to pass step numbers, reference previous events, or chain anything together. Janus connects them into a funnel using Journeys.

You choose the match scope when creating a journey. Match by distinctId for conversions that span multiple visits over days or weeks. This catches users who sign up on a different day than they first visited, which is common for SaaS products with an evaluation period. Match by sessionId for single‑session flows where you want to measure immediate friction.

Conversion windows range from 5 minutes to 30 days. The default is 24 hours. If a user completes step 1 but doesn’t reach step 2 within the window, they count as drop‑off. Wider windows capture more conversions but dilute the signal. For most SaaS signup funnels, 7 days covers the typical evaluation period.

A practical example: you define a 3‑step journey with signup_started → email_verified → first_project_created, matched by distinctId over a 7‑day window. Results show 200 users started signup, 140 verified their email (70% step rate), and 80 created a project (57% step rate, 40% overall conversion). The median time between steps tells you whether the drop‑off is friction or intent. Thirty seconds between steps means people are stuck on the form. Three days means they’re evaluating.

The value of funnel data is in the comparisons. Run the same journey across different time periods to see if conversion improves after a change. Run it filtered by referrer to see whether organic traffic converts differently than direct. Journeys also accept page view steps, so you can mix URL paths with custom events: /pricing → checkout_started → purchase_completed.

Journeys are available on Pro plans. The free tier tracks all events but does not include funnel analysis. See the Journeys documentation for setup instructions and step types.

For side projects and early‑stage SaaS, these gaps rarely affect the decisions you make. You need conversion rates and drop‑off points. You need to know which step loses the most users and how long people take between steps. You don’t need an identity graph to answer those questions.

Accuracy is high for the metrics that matter. Conversion rates are calculated from real events on your domain, not from probabilistic cross‑device matching. The numbers represent real actions from real visitors. The main limitation is that some returning visitors are counted as new if they cleared storage, which slightly inflates top‑of‑funnel numbers without affecting step‑over‑step rates.

If you run paid ad campaigns that require cross‑platform attribution, you can run GA alongside Janus on the pages where ads convert.

New to Janus? Add analytics to your side project in 5 minutes. Already tracking events? Connect Janus to your AI coding agent and check conversion rates from your editor.

Track conversions. Skip the cookie banner.

Free tier. No credit card. No consent popup.

Get started with Janus