This review looks at how Userpilot captures user behavior, what you can learn from its analytics reports and dashboards, which features are available on each plan, and where the platform still falls short. I’ll also cover newer capabilities like autocapture, Agent Analytics, and Lia, and explain how they fit into the product today. Everything in Userpilot analytics starts with autocapture. Once it’s enabled, Userpilot begins recording supported user interactions automatically, so you don’t have to instrument every event before collecting data.<\/p>\n On web, autocapture records supported frontend interactions such as clicks, text inputs, and form submissions, with no developer wiring each one. On mobile, the Userpilot SDK automatically captures supported native interactions: taps, text changes, selections, switches, sliders, date and time picker changes, menu selections, and screen views. The result is a stream of raw behavioral data that exists before you have decided what matters.<\/p>\n Those auto-captured interactions arrive as raw events, and a raw event cannot be used across reports, segmentation, or in-app experiences until it is labeled. Teams review the raw stream and create labeled events<\/a>, which turn nameless interactions into named events you can reuse everywhere in Userpilot. This labeling step is the real work of setting up analytics, and it is worth doing deliberately rather than tagging everything.<\/p>\n On web, you create a labeled event with the Visual Labeler or by entering CSS selectors. On mobile, you build one by selecting captured raw events and defining matching conditions such as the screen, the UI element hierarchy, accessibility identifiers, and other platform-specific properties. A single labeled event can target Android, iOS, or both at once.<\/p>\n Tracked events are the other half of the picture: developer-defined events sent through code. They work across web and mobile, support custom metadata, and remain the right choice for backend events or anything autocapture cannot see. Custom events then group several tracked or labeled events into one business event when a single action spans multiple interactions.<\/p>\n One detail matters more than it sounds. A labeled event carries its historical data from before you labeled it in Event Overview and in analytics reports, so you can analyze behavior retroactively. Segments and content triggering, though, only use data collected after the event was labeled, which is the trade-off to remember when you plan a new segment.<\/p>\n The users dashboard is the overview of your entire user base in one place. You can filter by segment, by company, or by last-seen activity, stack multiple filters to narrow the result, run bulk actions on the users who match, and export the list as a CSV. It is the aggregate view, the altitude you start from before you go looking for a specific account.<\/p>\n Clicking an individual user opens a profile that shows the events and actions that person performed in whatever timeframe you choose. That is how you move from a number on a chart to the specific human behind it, which is usually where an investigation actually gets useful. The profile is the bridge between “retention dipped” and “this account stopped doing the thing that predicts retention.”<\/p>\n The company-level view is what makes this work for B2B. It supports account-based analysis, so you can move from overall company metrics down to the activity of individual users inside each account. For most B2B SaaS products, adoption is measured at the customer account level rather than the individual user level, and this view is where that measurement happens.<\/p>\n There is one constraint worth naming here, because it shapes how finely you can slice. Segments are capped at 10 on the Starter plan and are unlimited on Growth and Enterprise. If your analysis depends on many overlapping user segments, Starter will feel tight quickly.<\/p>\n The four analytics reports answer different product questions, so understanding what each one is built for matters more than simply knowing its name. All of them can be filtered by platform, either All Platforms, Web, or Mobile, so you can analyze your entire product or focus on one platform. Here’s what each report does and when to use it.<\/p>\n Trends<\/a> is the most flexible report in the set and the one worth learning first. You choose whether to analyze at the user or company level, then define up to 25 metrics drawn from events, pages, or content engagement. That single report can hold most of the questions a product team asks in a week.<\/p>\n Each metric can be measured several ways, and the range is wider than most people expect. You can count total events, unique users or companies, total sessions, active users on a DAU, WAU, or MAU basis, count per user including average, median, and the 90th, 95th, and 99th percentiles, or an aggregate of a numeric property such as sum, average, or distinct count. That percentile support is the difference between “average usage looks fine” and “the top 1% of accounts are carrying the number.”<\/p>\n Formulas let you build custom metrics using arithmetic operators, for example, a conversion rate calculated from two events. Breakdowns then split a metric by an event, user, or company property, or by a segment, with up to two breakdown properties per report, and numeric properties can be bucketed into ranges. A “For the First Time” filter restricts a metric to a user’s first performance of an action, which is how you isolate genuinely new behavior instead of repeat activity.<\/p>\n For visualization, Trends offers line, column, stacked variants, pie, single-metric, and scatterplot charts, viewed linearly or cumulatively. Pre-built templates cover Top 10 Events, Pages, Users, and Companies when you want an answer without building from scratch. This is the report where Userpilot’s analytics genuinely have depth.<\/p>\n Funnels<\/a> measure conversion and drop-off across a sequence of steps, which is the report you reach for when you want to know where users fall out. Each funnel shows stage-by-stage conversion and completion time, so you see both how many users make it and how long the surviving path takes. You can break the funnel down by user, company, event, device, browser, operating system, or other available properties to compare conversion across different groups.<\/p>\n Paths<\/a> reconstruct the sequence of actions before or after a chosen event, up to 10 steps in either direction. Where a funnel assumes a linear path you defined in advance, Paths show what users actually did, including the detours you would never have thought to chart. That is the report for discovering behavior rather than confirming a hypothesis.<\/p>\n Retention answers whether users return after a starting action, and you can view it as a trend, a single-period metric, or a cohort table. It is the report that tells you whether new behavior is sticking or evaporating a week later. I’ll keep this short here because we cover it properly in the dedicated cohort retention analysis<\/a> guide, including how to read the shape of a retention curve.<\/p>\n Two facts here decide who can use what. Of the four reports, Funnels, Paths, and Retention are Growth and Enterprise features, while Trends is available on every plan including Starter. Saved reports of any type can be viewed, edited, duplicated, deleted, sorted by type, filtered by the teammate who created them, and exported as a CSV.<\/p>\n This is the real reason to run analytics inside Userpilot rather than a standalone tool, so it is worth framing as a workflow rather than a feature list. The value comes less from any single report and more from the short distance between noticing something in a report and doing something about it.<\/p>\n When a report surfaces an unexpected drop-off, you can open Session Replay<\/a> directly from analytics and watch the affected sessions. Seeing what happened right before the drop usually tells you whether the problem is the product, the copy, or the timing. That inspection step is what stops a team from guessing at a fix.<\/p>\n From there, a finding becomes a segment built on user data, company data, or event behavior. That segment does double duty: it is an analytics filter for comparing its behavior against everyone else, and it is an audience for in-app experiences. On web that means flows, checklists<\/a>, tooltips, modals, and surveys, and the same behavioral segments can trigger mobile in-app experiences<\/a> too.<\/p>\n Event-based triggering is the part that makes the response feel timed rather than broadcast. Content fires at the moment a user does, or does not, perform an action, so the message lands in context instead of on a schedule. The practical loop is: spot a problem in a report, watch the affected sessions, build a segment of the impacted users, and trigger an experience for them, with no export and no engineering involved.<\/p>\n This anti-feature-factory habit is how I try to work in practice, and it is the kind of fix that moves a number without a sprint. When we launched Userpilot’s own email feature, the funnel showed a sharp drop at the domain verification step, and I did not open an engineering ticket for it.<\/p>\n Within a few hours, I just created a targeting tooltip and showed it to users and highlighted the correct steps for them to make it clear what to do next. That helped a lot on reducing friction and supporting users in real time without involving our dev team.<\/p><\/blockquote>\n One caveat belongs here so the loop isn’t oversold. Event-based content triggering<\/a>, called Event Occurrence, along with A\/B testing, are Growth and Enterprise features. The analyze half of the loop starts on Starter through Trends, but the act half depends on the plan you are on.<\/p>\n Feedback analytics get lighter treatment here because we cover them fully in the Userpilot NPS<\/a> guide. What matters for a user-analytics review is where the feedback lives, not every survey setting. It lives in the same platform as the behavioral data.<\/p>\n Every survey has its own analytics showing the share of respondents who chose each option, the most common answers, and the qualitative responses, filterable by segment, company, or time period. The NPS dashboard then consolidates the score data: total shown, responses, response rate, qualitative responses, and how the score trends over time. You can read individual responses per user when a number needs a face.<\/p>\n Here is the point for analytics, specifically. Because qualitative feedback sits beside behavioral data, a low NPS score can be cross-referenced against what that user actually did in the product. That is a different kind of answer than a survey tool alone can give you.<\/p>\n User analytics assumes the population you are measuring is people, and for most of its history that assumption held. When part of a product’s activity comes from AI agents calling APIs or MCP tools, headline metrics like DAU and event counts can be inflated by activity no human performed. Analysts at a16z have taken to calling one version of this the “AI tourist” problem<\/a>, where low-friction sign-ups distort the numbers you trust.<\/p>\n Userpilot’s answer is to separate agent-originated events from human-originated ones at the instrumentation layer, so any report can be filtered by source. Agent Analytics<\/a> is the dedicated view for agent activity, kept apart from the human stream rather than blended into it. If a meaningful share of your traffic is non-human, that separation is what keeps your human metrics honest.<\/p>\n
\n
\n![\"demo](\"https:\/\/blog-static.userpilot.com\/blog\/wp-content\/uploads\/2026\/06\/CTA-blog-banner-1-1.png\")
<\/a><\/p>\nHow Userpilot captures behavioral data<\/h2>\n
![\"Raw](\"https:\/\/blog-static.userpilot.com\/blog\/wp-content\/uploads\/2026\/07\/Userpilot_raw_events_optimized_1600.jpg\")
The users dashboard and user profiles<\/h2>\n
The analytics reports, and how each one is built<\/h2>\n
![\"Userpilot](\"https:\/\/blog-static.userpilot.com\/blog\/wp-content\/uploads\/2026\/05\/Product-Analytics-hp-userpilot.png\")
Trends: How a metric changes over time<\/h3>\n
![\"trend](\"https:\/\/blog-static.userpilot.com\/blog\/wp-content\/uploads\/2026\/07\/WhatsApp-Image-2026-07-29-at-00.09.08.jpeg\")
Funnels: Conversion and drop-off across a sequence<\/h3>\n
![\"funnel](\"https:\/\/blog-static.userpilot.com\/blog\/wp-content\/uploads\/2026\/07\/Screenshot-2026-07-29-at-12.12.29-AM-scaled.png\")
Paths: What users actually do<\/h3>\n
![](\"https:\/\/blog-static.userpilot.com\/blog\/wp-content\/uploads\/2026\/07\/new-new.jpeg\")
<\/p>\nRetention: Whether users come back<\/h3>\n
![\"retention](\"https:\/\/blog-static.userpilot.com\/blog\/wp-content\/uploads\/2026\/07\/Screenshot-2026-07-29-at-12.15.49-AM.png\")
Turning analytics into action<\/h2>\n
![\"the](\"https:\/\/blog-static.userpilot.com\/blog\/wp-content\/uploads\/2026\/07\/IMAGE-2-analytics-loop-1.png\")
![\"Session](\"https:\/\/blog-static.userpilot.com\/blog\/wp-content\/uploads\/2026\/05\/session-replay-powered-analytics-userpilot.png\")
Survey and feedback analytics<\/h2>\n
![\"Userpilot-nps\"](\"https:\/\/blog-static.userpilot.com\/blog\/wp-content\/uploads\/2026\/07\/userpilot-nps-dashboard-GIF.gif\")
Reading analytics when some of your users aren’t human<\/h2>\n
![\"Userpilot](\"https:\/\/blog-static.userpilot.com\/blog\/wp-content\/uploads\/2026\/05\/AI-Agent-Analytics-General-view-Userpilot.png\")