Site analytics and A/B tests

Every site you host comes with analytics built in - no separate account, no cookie banner, nothing to install. Turn it on (or off) per site from the same place you manage that site's hosting settings; it's on by default.

What it tracks

• Views and sessions. How many people visited, and how many separate visits (sessions) that was, per day.
• Referrers. Which other sites sent you traffic.
• Campaign sources. If you link to your site with utm_source, utm_medium, or utm_campaign in the URL (the standard way ad platforms and newsletters tag links), those show up broken out by source.
• Goals. Form submissions, clicks on a phone number or email link, clicks to another site, and anything you tag yourself (see below) - each with a conversion rate: the share of visitors who did that thing at least once.
• Page-level performance. Which pages get the most traffic, and which convert best.

No cookies are used, and no IP addresses are stored. A visitor is identified only by a random id kept for the length of their visit.

Tagging your own goals

Add data-nx-goal="signup" (or any name you like) to a button or link in your site's HTML, and a click on it is tracked as that goal automatically.

If you want to track something from your own script, the page exposes a small helper:

<script>
  window.nx.track('newsletter_signup');
</script>

The weekly summary

Once a week, if your site has had any traffic, you'll get a short plain-language summary: what grew, what fell, your top referrer, and which page is converting best. It's the same summary an assistant reads before working on your site's marketing, so it always starts from real numbers rather than guessing.

A/B testing

You can run a simple A/B test on any single element of a page - a headline, a button's text, a call to action. Define it as a small list: which page, which element (by a CSS selector, like #hero-title), what counts as a win (a goal name from above), and two or more variants to try.

While a test is running, each visitor sees one variant consistently for the rest of their visit. Once you (or your assistant, if it's set to act on its own) decide there's enough data, "pick a winner" applies the best-converting variant to the live page and ends the test - no code changes needed on your end.

An assistant working on your site's marketing can set these up, check results, and pick a winner for you, always starting from your weakest-converting page.