What Amplitude is and how Kite uses it
Amplitude is a product analytics platform. It records what people do in your app or on your website as events, such as "Signup Started," "Account Created," or "Project Shared," and lets you study those events as funnels, retention curves, and user timelines. An AI marketing agent uses Amplitude to connect marketing to what happens after someone arrives. It finds the step where new users give up, works out which campaigns bring people who stay, and checks whether a change it shipped actually moved the number.
Kite does that work in your own Amplitude project. It reads your funnels, retention, revenue, and individual user activity, then acts on what it finds, both in Amplitude and in the pages, emails, and campaigns it runs elsewhere. You connect your project with its API key and secret key.
Marketing jobs Kite can do with Amplitude
Amplitude holds the record of what users do after they sign up, so Kite's work with it runs from reading that record to acting on it. The jobs below start with finding where people stall, move through judging which channels pay off, and end with the changes Kite makes inside Amplitude to keep the numbers trustworthy.
Find the step where users drop out of a journey
Kite builds funnels from the events your product already sends to Amplitude to find the exact step where people give up. A self-serve signup funnel might run from "Pricing Page Viewed," to "Signup Started," to "Account Created," to "Project Created." Kite reads how many people complete each step within the time window you'd expect, such as seven days for a trial, and how long the ones who convert take to get there.
The overall rate rarely tells the whole story, so Kite splits each funnel by one property at a time, such as platform, country, plan, or first UTM source. That's how a vague "activation is down" becomes something specific, like "new mobile users from one paid campaign stop at account creation." Kite can also limit a funnel to new users only, which keeps returning customers from hiding a problem with first-time visitors.
Before Kite trusts a sudden drop, it checks the events themselves. It reads the event types in your project with this week's volume for each, so it can spot an event that stopped arriving or was renamed in a recent release. A step that broke because of tracking gets a tracking fix, not a page rewrite.
See which channels bring in users who stay
A campaign that fills your signup chart with people who never come back costs more than it looks. Kite reads Amplitude's retention data to see what share of new users return in the days and weeks after they sign up, and it can split that by a property such as acquisition source or plan.
Here's how that plays out. Say a LinkedIn campaign and a Google search campaign each brought in 300 signups last month. If 35 percent of the search signups are still active four weeks later and only 8 percent of the LinkedIn signups are, Kite flags the gap and recommends where to shift effort. It can then act on that recommendation, for example by rewriting the LinkedIn landing page so it sets clearer expectations, or by drafting an onboarding email for the users who signed up and went quiet.
When your product sends revenue events to Amplitude, Kite also reads revenue per user, revenue per paying user, and total revenue over time. That lets it judge channels by what their users pay, not just by how many sign up.
Trace how individual users found you
Sometimes the useful question is about specific people. Kite can look up a user by the ID your product gives them and read their event stream from the first event forward, including the UTM parameters and referrer recorded when they arrived.
That's useful when a handful of accounts matter a lot. If five teams upgraded to your top plan this month, Kite can read how each one first arrived, which pages they saw before signing up, and what they did in their first week. Patterns across those five timelines often point to the campaign or page worth more budget.
Watch the numbers that show growth and engagement
Kite reads the headline numbers your team checks every week. These include daily, weekly, and monthly active users, new users, and how those numbers break down by platform, country, or app version. It also reads session counts per user, average session length, and today's live activity in five-minute intervals.
Kite uses these numbers to notice when something changes and to explain it. If new users jump 40 percent on a Tuesday, Kite checks which country, platform, or source the jump came from before it tells your team what happened. It can deliver that readout once or put it on a schedule, such as every Monday morning.
Mark launches and campaigns on your charts
When Kite ships something, it can add a note to your Amplitude charts on that date. A note might read "Pricing page rewrite" or "Spring webinar email sent," filed under a category such as Marketing, and it can cover a single day or a date range. Kite can attach the note to one chart or show it across every chart in the project.
Those notes make before-and-after comparisons readable for everyone. Anyone who opens your signup chart a month later can see why the line moved, without asking Kite or digging through old conversations. Kite also reads the notes your team has already added, so it knows about a release or outage before it explains a change.
Save and update cohorts of users
A cohort in Amplitude is a saved group of users, such as everyone who started a trial in September but never created a project. Kite reads the cohorts your team has already saved, and it can pull the members of one, along with the user properties you choose, to use in its own work. For example, it can use that list to plan an email to the trial users who stalled.
Kite can also create a cohort from a list of users it identified, or add and remove members from an existing one. If Kite's research marks 40 free accounts as ready for a sales conversation, it can save them as a cohort so your team can open the same group in Amplitude, chart it, and send it to the tools your Amplitude account already syncs with.
Tag users and record events
Kite can update user properties in Amplitude without sending a new event. For example, after an email campaign it can mark each recipient with the campaign's name, so later funnels and retention charts can be split by who got the email. It can also send events into your project, which is useful when a step in the journey happens somewhere Amplitude doesn't see.
Keep your tracking plan clean
A funnel built on the wrong event gives the wrong answer, and tracking plans get messy as teams grow. Kite reads your project's event types, event properties, and user properties, along with their categories and descriptions, so it uses the exact names your product sends.
Kite can also tidy that plan. It writes descriptions for events that have none, groups events into categories your team will recognize, and updates or removes event types that are no longer used. It can restore an event type that was deleted by mistake. When a funnel needs a step your product doesn't track yet, Kite can add the event to your tracking plan with a clear description, so your developers know exactly what to build.
What Kite reads and changes in Amplitude
Information Kite can read
- FunnelsStep-by-step conversion between two to four events, the time people take to convert, and a breakdown by one property such as platform or UTM source. These tell Kite where to focus.
- RetentionHow many users return after a starting action, by day, week, or month, and split by a property. This shows which channels bring users who stay.
- Event trendsUnique users, totals, and averages for any event over time, grouped by a property. Kite uses these to explain spikes and drops.
- Active users, new users, and sessionsDaily, weekly, and monthly counts, live activity for today, sessions per user, and session length. These are the headline numbers Kite watches.
- RevenueRevenue per user, revenue per paying user, and total revenue over time, when your product sends revenue to Amplitude. These let Kite weigh channels by what users pay.
- Individual user activityOne user's profile and event stream, including the UTM parameters and referrer from their first visit. Kite reads these to trace how important accounts found you.
- CohortsYour saved cohorts and, on request, the members of one with selected user properties. Kite uses them to compare groups and plan targeted work.
- Chart notesThe notes your team has added to charts and the categories they belong to. These give Kite context before it explains a change.
- Your tracking planEvent types with this week's volume, event properties, user properties, and event categories. Kite reads these first so every query uses the right names.
Information Kite can create or change
- Chart notes and releasesKite adds, edits, and removes chart notes, creates note categories, and records product releases, so changes show up on the date they happened.
- CohortsKite creates a cohort from a list of users and adds or removes members from an existing one.
- User propertiesKite sets or updates properties on users, such as the campaign they received, and links different IDs that belong to the same person.
- EventsKite sends single events or larger batches into your project.
- Event types and categoriesKite creates, renames, describes, categorizes, removes, and restores event types, and manages event categories.
- Account propertiesKite sets properties on a group such as a company, including its name, industry, or plan, when your Amplitude plan supports account-level reporting.
How Kite connects to Amplitude
Add the API key and secret key from your Amplitude project's settings. Kite then works with that project's data and the access those keys grant. If you keep separate Amplitude projects, such as one for your marketing site and one for your product, connect the one that holds the journey you care about most. You can then start with a question like "why did trial activation drop last week?" Kite looks up the exact event names in your project, runs the funnels and retention queries, and comes back with what it found and what it proposes to do about it.
This connection is separate from the analytics built into Kite. A website Kite builds and hosts records its traffic, form submissions, and conversions in Kite's own analytics. Connect Amplitude when your product or website already sends its events there.
Playbooks using Amplitude
- How to find where users drop out of an Amplitude funnelWhen you want to know which step of a journey is losing people, Kite charts the steps of that journey in order in your Amplitude project, checks that each step's event is tracked properly, and breaks the weakest step down by who reaches it, then turns what it finds into a fix.
- How to find broken or leaky funnels with Segment, AmplitudeWhen Segment sends your events into Amplitude, a funnel step can drop because people leave or because tracking broke. Kite checks delivery in Segment, then checks the event names, volumes, and user identities in Amplitude, so you fix the tracking faults and the real leaks separately.
Where Amplitude fits in Kite's marketing work
When your product sends its events to Amplitude, Amplitude is where Kite answers your questions about funnels, retention, and groups of users. You ask in plain English, and Kite looks up the exact event names in your tracking plan, then runs the funnel, retention, or event trend that answers the question. It checks this week's volume for each event first, so a tracking fault isn't reported as a real change, and it ties the answer back to your marketing by splitting the result by a property such as first UTM source.
When the journey from first visit to signup or purchase is tracked in Amplitude, Amplitude is where Kite finds the step that loses the most people. Kite builds the funnel from the events your site or product already sends, such as "Pricing Page Viewed" through "Account Created," and splits the weakest step by one property at a time, such as platform, country, or first UTM source, to see which visitors that step loses. It checks this week's volume for each step's event before it trusts a drop, so a tracking fault gets a tracking fix and a real leak gets a change to the step that causes it.
Amplitude supplies the measurement for conversion work on a product or website that Kite doesn't host. Kite reads the funnel to find the weak step and uses user timelines to understand why people stall there. It then follows signups forward with retention to see whether the people who convert actually stay. After Kite ships a change, it marks the date on your Amplitude charts and reads the same funnel before and after, so the number it set out to move is the number it reports back to you.
When your product sends its events to Amplitude, Amplitude shows how far new signups get before they reach the first result that turns them into paying customers. Kite builds the funnel from signup to that result, limits it to new users, and splits the step where they stall by a property such as platform or first UTM source. If a step has no event to measure it, Kite adds the event to your tracking plan so your developers know what to build. After Kite rewrites the signup page or sends an onboarding email, it tags the users who got the email and marks the change on your charts, then reads the same funnel again to see whether more new users got there sooner.
When your product sends an event for each onboarding step, Amplitude shows the screen where new users stall. Kite builds the funnel through those steps for new users only and reads what individual users did before they stopped, so the onboarding screens, empty states, and upgrade prompts it writes speak to what each user has done so far. After the new copy ships, Kite marks the date on your charts and reads the same funnel again to see whether more users take the next step.
When the funnel Kite builds ends in a signup for your product, Amplitude measures what happens after that last step. Kite reads how far new signups get through your product's key journey and whether the people each campaign brings in keep coming back, and it reads revenue per user when your product sends revenue events. That tells Kite whether the funnel is bringing in customers who stay and pay, so it improves the step that matters for revenue, not only the one that collects the most signups.
When your product or website is measured in Amplitude and you can't run an A/B test on it, Amplitude holds the before-and-after record. Kite marks the date a change shipped with a note on your Amplitude charts, then reads the same funnel for the period before and the period after, so the verdict on whether the change worked comes from the charts your team already uses.
When the number that moved lives in Amplitude, such as signups, new users, or active users, Amplitude is where Kite checks the change. Kite first reads the event's volume and your tracking plan to confirm the change is real and not an event that was renamed or stopped arriving. It then splits the number by a property such as country, platform, or source to find where the change is concentrated, and reads your team's chart notes to match it to what shipped at that moment.
When your product sends its signup and activity events to Amplitude, Amplitude holds the history a cohort analysis compares. Kite reads retention from the signup event by week or month to see whether people who joined recently come back as often as the people who joined before them. When your product sends revenue events, Kite also reads revenue per user over time, so the comparison covers what people pay as well as whether they stay.
When your product records in Amplitude how paying customers use it, Amplitude is where the warning signs before a cancellation show up. Kite reads the activity of customers who left to find what they did less often, or stopped doing, before they cancelled, then looks for the same signs among the customers you still have. It can save the customers it flags as a cohort and tag them with a user property, so your team can open the same group in Amplitude while Kite drafts the message to each one.
When your product records upgrades in Amplitude, Kite builds the funnel from signup to upgrade and reads the activity of users who upgraded to learn what they did before they paid. That moment shapes the upgrade prompts and trial emails Kite designs. Kite tags the users who got each email so the funnel can be split by who received it, while the final answer on who actually paid comes from Stripe.
When your signup steps and upgrade prompts live in a product measured in Amplitude, Amplitude shows which step loses the most people between clicking Sign up and paying. Kite builds the funnel through those steps for new users only and splits the weakest step by a property such as platform, plan, or first UTM source, so each test aims at one step. After a change ships, Kite marks the date on your charts, reads the same funnel before and after, and reads retention to check whether the people the new version brought in keep coming back.
When your product records what people do in Amplitude, Amplitude is where Kite finds the free and trial users worth a personal email. Kite reads each user's activity to learn what paying customers did before they paid, then watches for free and trial users who reach the same point. It saves the users it flags as a cohort and updates their user properties, so your team can open and chart the same group in Amplitude while Kite drafts the personal email your founder sends each one.
Limits and product-specific details
- Kite works with the events your product sends. Amplitude only holds the events your app or website records. Kite can add a missing event to your tracking plan, but your developers still need to add the tracking code before the step shows up in a funnel.
- Some features depend on your Amplitude plan. Reading, pulling, creating, and updating cohorts goes through Amplitude's Behavioral Cohorts API, which Amplitude offers as an add-on on its Growth and Enterprise plans. Setting properties on companies or other groups requires Amplitude's Accounts add-on. Revenue numbers appear only if your product sends revenue events.
- Kite's analysis lives in Kite. Kite reads Amplitude's data and adds notes and cohorts there, but it doesn't save new charts or dashboards in Amplitude. It reports its findings in the conversation and can publish a report or dashboard of its own.
Common questions
Do I need to connect Amplitude if Kite hosts my website?
No. Kite's built-in analytics measure traffic and conversions on any site Kite hosts. If a site Kite publishes already loads Amplitude's standard snippet, Kite's analytics also record the same event names from the next publish onward, and your Amplitude tracking keeps working as before. Connect Amplitude when you want Kite to read your history there or work with what happens inside your product.
Can Kite run tests in Amplitude Experiment?
Not through this connection. On a site Kite hosts, Kite runs its own A/B tests: it sets the goal, splits traffic between versions, and asks you to approve the winner. For a product or site measured in Amplitude, Kite measures a change by comparing the funnel before and after it shipped, with a note on your charts marking the date.
How is this different from Amplitude's own AI features?
Amplitude's AI features work on your Amplitude data. Its AI Agents run analyses, investigate why a number moved, build dashboards, flag problems, and recommend actions. Kite starts from your business and your marketing goals, uses Amplitude as one of its sources, and then does the work the answer calls for. That might mean rewriting a landing page, sending an onboarding email, building a lead list, or publishing a report. It then comes back to Amplitude to check whether the change worked.
