What PostHog is and how Kite uses it
PostHog is a product analytics platform that records what people do on your website and in your product. It stores each pageview, click, and signup as an event, and it adds session recordings, heatmaps, experiments, surveys, and feature flags on top of that data. An AI marketing agent uses PostHog the way a growth marketer would. It reads the events to find where people stop converting, looks at what they did at that moment, and then tests a change and measures it in the same project.
Kite runs that whole loop. It finds the step that loses people, works out why, and sets up the experiment, survey, or email that addresses it. Then it reads the same numbers again to see whether more people got through. You connect your own PostHog account through OAuth.
Marketing jobs Kite can do with PostHog
PostHog holds both halves of conversion work: the record of what people did, and the tools that change what they see next. The jobs below follow that order. They start with reading your traffic, move through finding and explaining a leak, and end with fixing it and keeping your team informed.
See where your traffic comes from and where it lands
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.
Find the step where people drop out of your funnel
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.
Watch what visitors did on the page that lost them
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.
Make sure your campaigns get credit for the signups they bring
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.
Test a fix with a PostHog experiment
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.
Ask visitors and users why they stall
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.
Group users and follow them over time
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.
Send lifecycle email through PostHog workflows
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.
Keep your team's PostHog dashboards current
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.
What Kite reads and changes in PostHog
Information Kite can read
- Web analyticsVisitors, pageviews, sessions, bounce rate, session duration, top pages, UTM sources and campaigns, devices, countries, Core Web Vitals, and the weekly digest. Kite uses these to spot traffic changes and the pages behind them.
- Funnels, paths, and trendsStep-by-step conversion rates, the people who converted or dropped off at each step, common navigation paths, and event counts over time. These tell Kite which step to fix first.
- HeatmapsClicks, rage clicks, mouse movement, and scroll depth on a page, plus the sessions behind a hotspot. These show Kite how people actually use a page.
- Session recordingsEach recording's duration, starting page, activity, and console errors, plus PostHog's review of selected recordings. Kite uses these to explain a drop-off.
- ErrorsError tracking issues with the pages, users, and sessions they affected. Kite checks these when a conversion falls for no visible reason.
- Marketing analyticsConversion goals, UTM tagging quality, and unmatched sources. Kite uses these to keep campaign credit accurate.
- ExperimentsEach experiment's status, exposures, and results for every variant and metric. Kite reads these to recommend a winner.
- SurveysQuestions, targeting, response counts, individual answers, and theme summaries. Kite turns these into the next change to test.
- People and cohortsPerson profiles and properties, cohort membership, retention, lifecycle, and stickiness. Kite uses these to judge whether new users stay.
- Your existing setupSaved insights, dashboards, and the events and properties your project tracks. Kite reads these first so it builds on what your team already measures.
Information Kite can create or change
- Actions and event descriptionsKite creates actions that turn clicks or pageviews into named conversion events, and it adds descriptions to your events and properties.
- Insights, dashboards, annotations, and subscriptionsKite saves charts, builds and updates dashboards, marks launches on charts, and schedules snapshots for your team.
- Experiments and feature flagsKite creates draft experiments and their flags, adds metrics, launches and pauses tests, refreshes results, and ships the winning variant.
- SurveysKite creates pop-up surveys and hosted forms, then launches and stops them.
- CohortsKite creates dynamic and static cohorts and changes who belongs to them.
- Recording playlistsKite collects the session recordings your team should watch.
- Workflows and email templatesKite writes email templates, builds workflows, runs tests, and turns a workflow on after you approve it.
- Person propertiesKite sets properties on a person's profile, such as a lifecycle stage, so later cohorts and funnels can use them.
How Kite connects to PostHog
Connect your PostHog account through OAuth. Kite then works with the records and permissions available to that connected account. If your organization has more than one project, such as one for your marketing site and one for your product, you choose the project Kite works in, and it can switch projects when a question spans both. You can start with a goal, like "find out why trial signups dropped last week." Kite proposes the funnel, the date range, and the recordings it will look at before it runs anything.
This connection is separate from the analytics built into Kite. A website Kite builds and hosts comes with its own traffic and conversion reporting. Connect PostHog when your site or product already sends its events there.
Playbooks using PostHog
- How to find where users drop out of your funnel with PostHogKite builds the funnel from your PostHog events, finds the step that loses the most people, explains why from the recordings and paths behind it, and proposes the change to make first.
- How to find broken or leaky funnels with Segment, PostHogSegment sends your events into PostHog. Kite checks whether a weak step is a real drop-off or an event that stopped arriving, so you fix the tracking or the page, whichever is actually broken.
- How to see which product behaviors lead to paid customers with PostHog, StripePostHog supplies what people did in your product, and Stripe supplies who went on to pay. Kite lines the two up on a shared customer identifier to find the actions that paying customers have in common.
Show 2 more playbooks
- How to see which Typeform submissions become conversions in PostHogTypeform supplies the form responses, and PostHog supplies what those people did next. Kite follows each submission into PostHog to see which forms produce people who go on to sign up or buy.
- How to get a weekly summary of your PostHog dashboard in SlackPostHog's own subscriptions post snapshots of a dashboard's charts to Slack on a schedule. Kite goes further: it compares last week with the week before, checks which changes are real, finds the cause, and posts a short summary with what to do next.
Where PostHog fits in Kite's marketing work
PostHog can carry the whole conversion loop for a site or product that Kite doesn't host. Kite reads the funnel to find the weak step, uses recordings, heatmaps, and surveys to explain it, and tests the fix as a PostHog experiment. Then it measures the result with the same events it used to find the problem. Because every stage happens in one project, the number Kite set out to move is the same number it reports back to you.
PostHog holds the funnel itself when your site or product sends its events there. Kite builds the funnel from the events your project already tracks, finds the step where visitors give up before signing up or buying, and breaks that step down by device, country, and UTM source to see which visitors it loses. It then reads the paths, heatmaps, and session recordings behind that step to work out the cause before it proposes a fix.
PostHog supplies the before-and-after numbers when a change can't run as an experiment, because the traffic is too thin or the flag setup isn't there. Kite marks the day the change went live with an annotation, then compares conversion on the same PostHog events before and after that day, so the verdict rests on the record your team already trusts.
PostHog supplies the traffic numbers when your site already sends its events there. Kite reads visitors, pageviews, sessions, and bounce rate against the period before, breaks each change down by page, source, and device, and explains in plain language what moved and why.
PostHog holds the answers when you have a question about your funnels, retention, or groups of users. Kite starts from the events and properties your project already tracks, builds the funnel, retention chart, or cohort that answers your question, and explains the result in plain language. When the answer touches your marketing, Kite breaks it down by UTM source and campaign, so you can see which channels brought the users you're asking about.
PostHog holds both your traffic and what each signup did afterward, which is what Kite needs to tell bots from people. Kite breaks your visits down by source, country, browser, and device against the period before, reads the session recordings behind a suspicious slice, and then rebuilds your funnel without the suspected bots so you can see your real conversion rate.
PostHog supplies the daily visitor counts when your site sends its events there. Kite compares each day's visitors with your site's normal level, and when a day falls well below it, Kite breaks the drop down by page, UTM source, and device so the alert says which pages and sources lost the visits. From there, Kite can find out why your website traffic dropped by checking the pages and campaigns behind the drop.
PostHog shows where a sudden change in signups or traffic is concentrated when your site or product sends its events there. Kite first compares the period with the one before to confirm the change is real, then breaks it down by page, UTM source, device, and country. It also checks error tracking for anything that broke at that moment, such as a signup button that stopped working in one browser.
PostHog supplies the visit record when your site runs somewhere Kite doesn't host. Kite breaks down your visits by referring domain and UTM source to pull out the ones ChatGPT, Perplexity, Claude, and Gemini send, sees which pages they land on, and builds funnels that show whether those visitors sign up or buy compared with everyone else. It then reports the numbers every week.
PostHog holds the conversion record when your site or product sends its events there. Kite creates the actions that turn a signup click into a named conversion event, audits your UTM tags so each signup is credited to the campaign that brought it, and reads your conversion goals with their recent results. That way, the signup count and the sources behind it are numbers you can trust.
PostHog's marketing analytics checks give Kite a direct read on how your campaign links are tagged. Kite audits the recent UTM tags in your project, finds the sources and campaigns that are mistyped or don't match a channel, and suggests the mapping that puts those visits back under the right one. It also lists your conversion goals with their last 30 days of results, so a goal that stopped counting shows up in the same audit.
PostHog supplies the signup side of the match when your site or product sends its events there. Kite reads the people and events in your PostHog project to see which channel, campaign, and landing page brought each signup and what they did before paying. It then matches those people to the deals in your CRM or the customers and payments in Stripe, so you can compare your channels on paying customers and revenue instead of visits and signups.
PostHog can be one of the tools whose signup count disagrees with your CRM or billing system. Kite reads how your PostHog project defines each conversion, including the actions and conversion goals behind it, so it can explain why the PostHog number differs from the others. Once your team settles on one definition, Kite writes it into the descriptions of your PostHog events and properties, so everyone who opens the project reads the same meaning.
PostHog measures the funnel when your pages or product send their events to it. Kite builds a PostHog funnel across the steps from first visit to purchase, finds the step where people drop out, and uses that step's recordings and paths to decide which step to improve first.
PostHog shows what new users do inside your product between signing up and reaching their first result. Kite builds the funnel from signup to that result, finds the step where new users stall, and uses recordings, paths, errors, and surveys to work out why. It then tests a fix, such as a new onboarding screen, as a PostHog experiment and rebuilds the same funnel to see whether more new users get there.
PostHog shows the screens where new users stall, which is where onboarding and upgrade copy matters most. Kite finds those screens from your funnels and paths, reads the recordings and survey answers behind them, and writes the onboarding screens, empty states, and upgrade prompts for each one. It can then test the new copy as a PostHog experiment to see whether more new users take the next step.
PostHog measures each step between clicking Sign up and becoming a paying customer when your signup flow and product send their events there. Kite builds the funnel across those steps, finds the one where people give up, and tests one change at a time to that step or to an upgrade prompt as a PostHog experiment. With Stripe connected through a shared identifier, Kite judges each version by how many people go on to pay, then ships the version that wins.
PostHog can both send the sequence and measure it when your team sends messages through PostHog workflows. Kite writes each email template, builds the workflow with a signup event as its trigger, and runs a test before anything goes live. It then reads the workflow's delivery stats alongside your PostHog funnels and retention, so each email is judged by how many new signups go on to take the action that makes them active users.
PostHog supplies the signup and activity data a cohort analysis needs. Kite groups your users by when they signed up and reads PostHog retention to see whether recent groups come back as often as earlier ones. With Stripe connected through a shared identifier, it also compares how well each group goes on to pay.
PostHog shows what your paying customers did in the product before they paid. With Stripe connected through a shared identifier, Kite finds that moment, then tests upgrade prompts and paywall changes around it as PostHog experiments and judges each one by who actually pays in Stripe. If your team sends messages through PostHog workflows, Kite can also write the trial emails that lead people toward that moment.
PostHog supplies the product usage that shows who is ready for a sales conversation. With Stripe connected through a shared identifier, Kite compares what paying customers did in PostHog with what everyone else did, then builds a dynamic cohort of free and trial users who reach that same usage. Because the cohort updates its membership on its own, Kite sees each new user who qualifies and reads their person profile before it drafts the email your founder sends.
PostHog shows what your expanding customers did in the product before they paid more. With Stripe connected, Kite compares the habits of customers who upgraded with everyone else on the same plan, then builds a dynamic cohort of current customers who match that pattern, so your team always has a list of the customers who are ready for an upgrade or add-on now. It can also test an in-product upgrade prompt as a PostHog experiment.
PostHog shows what customers did in the product in the weeks before they cancelled. With Stripe connected through a shared identifier, Kite compares those customers with the ones who stayed, then builds a dynamic cohort of paying customers who show the same warning signs. Because the cohort updates its membership on its own, Kite checks it every week and drafts a message for each customer who joins it.
PostHog owns the website and product stages of a full-funnel dashboard when your site or product sends its events there, from first visits through signups to the first actions new users take. Kite pulls those stages from PostHog alongside the later stages from the tools that own them, and it can save the funnel as an insight on a PostHog dashboard, so your team reads the top of the funnel where it already looks.
PostHog supplies the website and product side of the weekly report. Kite reads PostHog's weekly web analytics digest and the dashboards your team already keeps, compares the week with the one before, and brings the changes that matter into the report.
PostHog is one of the sources a public metrics page can draw from, alongside Stripe and Google Analytics. When your site or product sends its events to PostHog, Kite pulls the figures you choose to share from your PostHog project, gives each one a clear definition on the page, and keeps them current on the schedule you set.
Review and approval
Kite proposes work by default. A team approves it with a reply in the conversation or the Approve button on a proposed plan. Teams that want Kite to act first can turn that on. In PostHog, approval comes before the steps that change what real people see or receive: launching an experiment, shipping a winning variant, launching a survey, and turning on an email workflow. Reading data, building funnels, and saving charts happen as part of the analysis.
Common questions
Can Kite connect PostHog behavior to revenue in Stripe or my CRM?
Yes, when your tracking gives both tools a shared identifier, such as the user's email address or account ID. With that in place, Kite matches each person in PostHog to their Stripe customer or CRM contact and compares what paying customers did with what everyone else did. Without a shared identifier, Kite compares the two sides as trends over the same period instead of person by person.
Does Kite check the ad platform data that PostHog pulls in?
Yes. If you've connected ad platforms to PostHog's marketing analytics, Kite checks each one's sync status, when it last synced, and any errors it hit. It then confirms that the campaign names coming from the ad platform match the UTM tags on your events, which is what lets PostHog put ad spend and signups in the same report.
