PostHog + Kite

Kite reads PostHog's web analytics to give you a current picture of your site's traffic. It pulls visitors, pageviews, sessions, average session duration, and bounce rate for any period, compared with the period before. It breaks those numbers down by page, UTM source and campaign, device, browser, and country. It also reads Core Web Vitals page by page, so it can spot the pages that load slowly.

Here's how that plays out. Say sessions from a LinkedIn campaign doubled last week, but signups didn't move. Kite sees both numbers and finds the landing page the campaign sends people to. Before it suggests rewriting anything, it checks whether that page is slow on phones.

Kite builds funnels in PostHog from the events you already track to find the exact step where people give up. A typical funnel might run from a pricing page view, to a click on "Start free trial," to a completed signup, to a first project created. Kite reads the conversion rate at each step and finds the one that loses the most people. It can then list the actual people who dropped off at that step, along with their session recordings when PostHog captured them.

Paths fill in the rest of the story. Kite reads the most common sequences of pages and events, so it can tell you what people do right before they buy, or where they go instead of finishing signup. If visitors who abandon the signup form keep heading to your security page, that's a strong hint about what the form is missing.

When a step has no event to measure it, Kite can create one. It defines a PostHog action, such as a "Sign up click" that matches clicks on your signup button. That way, the funnel ends on the conversion you actually care about.

Funnels show where people leave, and PostHog's heatmaps and recordings show what happened just before they did. Kite reads click, rage-click, and scroll-depth heatmaps for a page, and it can filter them by screen width to compare phones with desktops. That's how it learns that most mobile visitors never scroll far enough to reach your pricing table. It can also open the sessions behind a hotspot. PostHog keeps heatmap data for 90 days, so Kite reads it within that window.

Kite searches your session recordings by page, date, person, or console errors. It can also ask PostHog to review a specific set of recordings against a question, such as why these visitors abandoned checkout. It saves the recordings worth watching into a playlist your team can open.

Errors belong in this picture too. Kite reads PostHog error tracking, so a signup button that breaks in one browser shows up as a named error with the pages and users it affected, rather than as an unexplained dip in signups.

Kite checks the tracking behind your campaign reports, because attribution only works when UTM tags match up. It runs PostHog's marketing analytics checks on your project. It audits recent UTM tags, finds campaigns with mistyped or unmatched sources, and lists the conversion goals you've set up with their last 30 days of results. It can break down the signups that count toward a goal by UTM source and campaign. It also suggests custom events that would make good conversion goals.

This is the work that keeps a report honest. Imagine half your paid signups are sitting under "non-integrated" because someone tagged a campaign "Linkedin" instead of "linkedin." Kite finds the mismatch, shows you which campaigns are affected, and suggests the mapping that puts those signups back under the right channel.

Once Kite knows which step to fix, it can test the fix as a PostHog experiment. First, it estimates how long the test needs to run to detect a meaningful lift with your current traffic. Then it creates the experiment with its feature flag and variants and adds the conversion it will measure. After you approve, Kite launches the experiment, reads the results as they come in, and recommends a winner. When you agree, it ships the winning variant to everyone, or it ends the test without shipping if neither version won.

A PostHog experiment decides who sees which version. Your site or app then shows that version by checking the experiment's feature flag. When Kite works on your website code through an imported GitHub repository, it can build the new version too. Otherwise, your developers connect the variant to the flag and Kite handles the test from there.

Sometimes the fastest way to learn why people stall is to ask them, and Kite does that with PostHog surveys. It can create a survey that pops up inside your site or product, or a hosted form with its own link. Kite writes the questions and targets the people you want to hear from, such as visitors who reached the pricing page but didn't start a trial. It launches the survey once you approve it. Then it reads response rates, individual answers, and a summary of the themes, and it stops the survey when it has heard enough.

A survey answer becomes evidence for the next change. If a third of pricing-page respondents say they couldn't tell which plan includes single sign-on, Kite has a specific rewrite to propose.

Kite builds PostHog cohorts, which are saved groups of people defined by their properties or behavior. One example is everyone who signed up in September and hasn't come back since. A dynamic cohort updates its membership on its own, while a static cohort holds a fixed list. Kite uses cohorts to break down funnels, to target surveys, and to keep certain users out of an experiment.

Kite also reads how people behave over time:

• Retention shows how many new signups come back in the weeks after they join.
• Lifecycle splits your active users into new, returning, resurrecting, and dormant.
• Stickiness shows how many days in a period people use a feature.

Together, these tell Kite whether a campaign brought in people who stay or people who leave after a day.

If your team sends messages through PostHog workflows, Kite can build those emails and the workflows that send them. It writes the email template in your brand voice. Then it creates the workflow with its trigger, such as a signup event or a scheduled batch, and runs a test before anything goes live. For a batch send, Kite counts exactly how many people the audience matches and tells you that number before you approve. Once the workflow is live, Kite reads its delivery stats.

That means a dormant-user cohort can turn straight into a win-back email without leaving PostHog. Kite picks the audience, writes the message, and then checks the lifecycle numbers to see how many of those people came back.

Kite keeps the analysis it does visible in PostHog, where your team already looks. It saves the insights it builds, such as a funnel or a retention chart, and adds them to new or existing dashboards. It marks important moments on your charts with annotations, so the day a new pricing page went live appears right on the trend line. It can also set up subscriptions that deliver a chart or dashboard snapshot on a schedule. And it reads PostHog's weekly web analytics digest, which compares visitors, top pages, top sources, and goal conversions with the week before.