Product Updates

🔍 Conversion Drivers in Funnels

Pinpoint what’s helping or hurting your conversions. Conversion Drivers automatically highlight the most significant factors influencing whether users drop off or complete a funnel, so you don’t have to guess where to dig deeper.

What You Can Do Now

• Identify high-impact drivers of conversion or drop-off without needing a hypothesis in advance

• Analyze correlations across event properties, user properties, and intermediary events

• View performance summaries for each driver, including conversion rate and share of funnel participants

• Drill into a driver to access a conversion matrix and correlation coefficient

• Group the funnel by any surfaced driver with one click to explore broader trends

How It Works

When viewing a funnel, click on any step and select “View Drop-Off & Conversion Drivers.” A modal will appear showing a ranked list of the most statistically significant factors associated with conversion or drop-off at that step.

You can configure which types of data we analyze:

• Event properties like plan type, referral code, or platform

• User properties like country, account age, or signup method

• Intermediary events, which occurred between the two funnel steps

Each surfaced driver includes:

• How much more or less likely users with the factor were to convert, expressed as a multiple (for example, users with platform = Android were 1.2x as likely to convert)

• Conversion rate for users with the factor

• Share of funnel participants who had the factor

Clicking into a driver opens the drilldown view, where you can explore:

• A conversion matrix that compares outcomes for users who had the factor versus those who did not

• A correlation coefficient measuring how strongly the factor is associated with completing the funnel

If a pattern looks meaningful, you can group your funnel by that property with a single click. This reconfigures the chart to show step-by-step conversion performance broken down by the selected value.

Impact on Your Analysis

Funnels tell you what your conversion rate is. Conversion Drivers help explain why.

This feature is especially useful when:

• You are exploring a new funnel without a clear hypothesis

• You notice a drop-off and want to identify potential causes

• You want to validate whether specific user groups or behaviors are influencing conversion

• You are monitoring changes in funnel performance and need to explain what shifted

By surfacing statistically significant behaviors and attributes, Conversion Drivers gives you a starting point for deeper investigation and helps you move faster from observation to insight.

Available Now

Conversion Drivers are available on any funnel for customers on the Pro plan or Enterprise customers with the Advanced Analytics package. Click on a step and select “View Drop-Off & Conversion Drivers” to get started.

Multi-Variant Support for Dynamic Configs

Dynamic Configs just got more powerful. You’re no longer limited to a single config value and a fallback. With multi-variant support, you can define multiple named JSON variants and control which one is served - all from the Statsig console, no deploys required.

🧠 Real-world Example

Let’s say you’re using a third party messaging service. You use Twilio, but want a fallback in case it goes down, and you’re testing AWS SES for cost savings.

With Multivariate Dynamic Configs, you can:

• Define config variants for each provider

• Roll traffic to AWS SES for 5% of users to validate integration

• Failover to Sendgrid if Twilio has an outage

• Tune timeouts and retry logic on the fly - no redeploys

⚡️ Available Now

We will be rolling out Multivariate Dynamic Configs to all Statsig customers over the next few weeks. If you'd like to get an early access, please reach out.

📼 Playlists in Session Replay

You can now group replays into Playlists to curate and share the sessions that matter most.

What You Can Do Now

• Add a replay to an existing playlist or create a new one directly from the replay viewer

• Access all your saved playlists from the new “Playlists” tab next to “Sessions”

• Cycle through playlist sessions without returning to the full list view

How It Works

From any replay, click the “Add to Playlist” button to save it. You can either select an existing playlist or create a new one on the spot. The Playlists tab lets you view and play through a sequence of sessions in one place.

Impact on Your Analysis

Playlists make it easy to gather replays related to:

• Bugs or usability issues

• Experiment test groups

• Feature launches or onboarding flows

They’re ideal for sharing context across teams without needing to repeat yourself.

Try building a playlist for your next launch review or bug triage session.

🧑‍💻 Revamped Users Tab

We’ve redesigned the Users tab in Metrics Explorer to make it more useful out of the box and more powerful when you need to dig deeper.

What You Can Do Now

• View recent users instantly

  Landing on the Users tab now shows a live sample of recent users with basic info, so you can immediately start exploring.

• Filter to specific user sets

  Narrow the list by applying filters based on:

  • User properties (e.g. country, plan type)

  • Events performed (e.g. signed_up, clicked_button)

  • Experiment or feature gate group (e.g. users in treatment vs control)

How It Works

You’ll see a sample of recent users as soon as the tab loads with no query required. Use the filter panel to drill down into any segment you care about. For example:

• Find users who dropped out of a funnel after the second step

• See only users in the “new-nav-rollout” feature gate

• Inspect event history for users on a specific pricing tier

Impact on Your Analysis

Filtering to specific sets of users helps you move from high-level trends to concrete user behavior. You can:

• Debug feature exposure by confirming which users were actually in a gate or experiment

• Investigate drop-offs by checking what users did before and after a key event

• Validate hypotheses about certain user groups, like whether trial users behave differently than paid ones

🎯 Segments in Funnels and Metrics Explorer

You can now create and analyze Segments directly from Funnels and use them in Metrics Explorer for deeper, follow-up analysis.

What You Can Do Now

• From Funnels: Create a new Segment from users who did or didn’t convert at a given step.

• In Metrics Explorer:

  • Filter your chart to only include users in a Segment.

  • Break down results by whether users are in a Segment.

• Segment Type: Currently limited to ID List-based Segments (max size: 1000 users).

How It Works

When viewing a funnel chart, you’ll see a new option to “Create Segment” next to conversion results. Clicking it will generate an ID List-based Segment you can name and save.

In Metrics Explorer, you’ll find Segment filtering under the filter panel, and breakdown by Segment in the group-by dropdown. These options appear only when the Segment is based on an ID list.

Impact on Your Analysis

This gives you a fast, integrated way to dig deeper into interesting groups of users, like understanding what users who dropped off at step 2 did before or after, or comparing behavior between converters and non-converters over time.

Looking Ahead

We’re planning to:

• Support additional segment types beyond ID Lists.

• Remove the 1000-user limit for ID List-based Segments.

Hypothesis Advisor

Writing good experiment hypotheses is key to a strong experimentation culture. Statsig now gives instant feedback on experiment hypotheses—flagging what’s missing. Admins can set custom requirements, which Statsig uses to guide experimenters toward stronger, more complete hypotheses.

This is gradually rolling out; reach out in Slack or to vm at statsig dot com if you'd like early access. When available, enable it from Settings -> Experimentation -> Statsig AI. See docs here.

🧩 ID Resolution in Funnels

Track users across the point where they go from anonymous to identified. Funnels can now connect actions taken before login with those that happen after, giving you a more complete view of conversion paths that span identity states.

What You Can Do Now

• Cross the anonymous-to-identified boundary in funnels

  Connect pre-login behavior (e.g. browsing, adding to cart) with post-login events (e.g. checkout, onboarding completion)

• Toggle ID resolution per funnel

  Click the gear icon when editing a funnel to enable or disable ID resolution for that analysis

• Configure identifiers

  In Settings → Analytics & Session Replay, choose which identifiers represent anonymous vs. identified users. Defaults are Stable ID and User ID

How It Works

When enabled, ID resolution stitches together events across anonymous and identified IDs if they’re seen on the same device. This turns fragmented journeys into a single user flow, even when a user logs in midway.

Example:

1. User views a product (Stable ID)

2. Signs up (User ID)

3. Completes checkout

With ID resolution on, these events are treated as a single funnel path.

Impact on Your Analysis

Funnels that previously showed drop-off at login steps may now show full completion. You’ll see higher true conversion rates, more accurate attribution, and better insight into how anonymous traffic behaves before converting.

🎬 Create a Statsig Sample App

You can now create a simple sample app to try out Statsig with - we've partnered with SampleApp.ai to let you easily create one with only a single prompt. This might help you out with exploring how Statsig works if you're a marketer or non-technical persona, that would still like to see what Statsig might look like once integrated in your app.

On statsig.sampleapp.ai - just enter a prompt or pick one of the samples, and we'll create a simple website for you to play with feature gates, analytics, and more.

Experiment Summary Customization

We're excited to announce the ability to add new custom sections and reorder sections in the experiment summary tab for greater customization of your experiment reporting.

These capabilities will also be available for experiment templates, giving you the ability to preconfigure summary sections to standardize formatting across your organization. These changes make the summary section a great place to store your experiment metadata like product research docs, links to design, or details on rollout plans.

🧪 Experiment Exposure Events Metrics Explorer (Cloud)

Overview

You can now treat experiment exposure events like any other event in Drilldown (time-series) and Funnel analyses. Exposure events include properties such as group, as well as user properties logged with the exposure. We currently only show first exposures to the experiment.

What You Can Do Now

• Pick exposure events in Drilldown charts to track how many users saw each variant over time.

• Add exposure as the first step of a funnel to measure post-exposure conversion paths.

• Group or filter by exposure properties, for example, break down results by variant, region, or device.

• Overlay exposure counts with key metrics in Drilldown to check whether metric changes align with rollout timing.

How It Works

1. Exposure logging

   The first time a user is bucketed into an experiment, an exposure event is recorded with contextual properties.

2. Event selection

   In both Drilldown and Funnel charts, exposure events appear in the same event picker you already use.

3. Property handling

   Any custom fields travel with exposures, enabling the same group-by and filter controls available for other events.

Impact on Your Analysis

Drilldown

• Validate rollout health by confirming traffic splits and ramp curves over calendar time.

• Catch logging issues early—spikes, gaps, or duplicates stand out immediately.

• Align timing with metrics by viewing exposure and conversion lines on one chart.

Funnels

• Measure post-exposure journeys starting the moment users see a variant.

• Pinpoint variant-specific drop-offs by breaking down each step.

• Ensure clean attribution because exposure proves the user entered the test.

• Segment by exposure fields (e.g., region or device) to uncover cohort-level insights.

This feature is now available on Statsig Cloud and coming soon to Warehouse Native. Give it a try the next time you validate an experiment. Seeing exposure data side-by-side with core metrics speeds up debugging and sharpens your reads on variant performance.