Analytics
What Tailor measures about your traffic, and how to read it before you decide what to test.
Overview
Analytics is under Signals in the left sidebar. It is not trying to replace GA4 or Amplitude. It exists so that the thing choosing your tests and the thing measuring them are reading the same numbers, which is the part that usually breaks when those two live in different tools.
Everything on this page is also available to Tailor Agent and over MCP, so you can ask for it in words instead of building a report.
Traffic
Sessions broken down by channel, source, campaign, search term, ad content, device, referrer, and locale, each with its own conversion rate. Landing pages can be ranked by traffic or by spend.
The breakdown is the point rather than the total. A site-wide conversion rate is an average of segments that behave nothing alike, and the gap between two of those segments is usually where the test is.
Conversions
Conversion rates per configured goal, the most-clicked CTAs on each page, and form submissions per page. Goals are defined in Conversion Goals & Tracking.
Detected CTAs are derived from the page
The detected-CTA goal counts clicks on the buttons Tailor finds on the rendered page. That makes it useful with no setup, but it is not held constant across variants: if a variant adds or renames a button, the set of things being counted changes with it. For a comparison you can trust across arms, configure an explicit goal.
Engagement
Per landing page: scroll depth, time on page, and frustration signals, meaning rage clicks and dead clicks.
These move earlier than conversions, which is what makes them useful. A page that broke in a redesign shows up as a dwell-time collapse days before the conversion number has enough volume to say anything. Dead clicks are the cheapest bug report you will get: they are people clicking something that looks interactive and is not.
Content Influence
Ranks the pages that converting visitors tend to look at, showing each page's conversion rate among the people who viewed it and its lift against the site-wide baseline. For example, visitors who read a given case study might convert at 1.8 times the overall rate.
It answers "what do converters read before they convert", which turns into a test directly: surface that proof earlier, or link it from the page where people stall.
Read it as association, not cause. People who are already going to convert also read more. The lift tells you what correlates with converting inside the window, which is a good reason to run a test and a bad reason to skip one.
Visitor Journeys
For a single visitor, the full timeline: page views with the complete request URL and every query parameter on it, clicks, conversions, scroll depth, dwell time, and company-level enrichment where visitor identification is enabled.
The query parameters are the reason to reach for this. Aggregate reports show you the campaign a visit was attributed to; a journey shows you the URL the visitor actually landed on, which is how you catch a campaign whose parameters are being stripped by a redirect.
Any parameter whose name looks like a credential has its value redacted before it leaves Tailor, so a token that ended up in a landing URL is not readable here.
Date Windows
Relative ranges like "last 30 days" resolve on the server to the calendar days ending today in your account's timezone. That is why the same question asked twice gives the same answer, and why the agent's numbers match the dashboard rather than landing a day off.
When you are comparing a period against a launch or a campaign flight, prefer explicit dates. A relative window that slides forward each day will quietly change the comparison underneath you.