Analyze your Flows drop-off performanceUpdated 15 days ago

After you’ve built and published a Flow, you can use Analysis mode to evaluate the effectiveness of your Flow.

With Analysis, you can see what’s working well for shoppers during the Flow and identify areas where you can make helpful changes.

By addressing areas in the Flow where shoppers drop-off or create tickets, you can make meaningful improvements to help resolve common questions — and ultimately increase your number of automated interactions.

Requirements

To access automation features like Flows, you must have an active AI Agent subscription
You must have Admin or Lead permissions to see Flow Analysis

Open your Flow Analysis

You can find and access the Analysis for a Flow from the Performance by feature report in Analytics. Alternatively, you can open a Flow in the Flow Builder then select Analysis.

Go to Analytics
In the menu, locate AI & automation, then select Performance by feature.
Select Analyze Flow next to the Flow that you want to review

Selecting the Analyze Flow button next to a Flow in the performance by feature report to analyze in more detail

Filter Analysis results by date

Use the calendar to filter the results of your Flow’s Analysis by date. You can select dates based on when data for the latest version of the Flow is available.

Your date range may cover a period of time where you had a different version of the Flow published. When this happens, we only show data for the most recent version of the Flow.

In the Performance by feature report, selecting the calendar to filter results by date range

How to read your Flow’s Analysis

Your Analysis always shows data for the latest version of the flow. Analysis data resets whenever you make a change to the Flow so you can always evaluate how the current state of the Flow performs.

If you’ve published your Flow in multiple languages, the Analysis includes data from the latest version of all published languages.

Overall Flow performance

Use the main Analysis stats to see how your Flow is performing overall. The overall Analysis shows cumulative stats for the entire flow, from beginning to end.

In the Analysis mode for a Flow, looking at the overall performance stats in the top-right corner of the page

Here’s a breakdown of the stats for overall Flow performance:

Stat	Description
Total starts	The total number of shoppers who started the Flow from the beginning
Automation rate	The % of shoppers whose inquiries the Flow resolved. This stat also includes the change in automation rate. For example, if your Analysis is filtered to 7 days, you can see the change in automation rate in comparison to the preceding 7 days.
Automated	The number of shoppers that completed the Flow and who are considered automated. This includes: - Shoppers that selected “Yes, this was helpful” at the end of a branch in the Flow OR - Shoppers who left the Flow at the last step and did not contact support in the following 72 hours
Drop off	The total number of shoppers who started a Flow and left before reaching the end
Tickets created	The total number of shoppers who started a Flow and either: - selected “No, I need more help” to create a support ticket OR - reached the end of a Flow where the action is Create ticket

Stats for Flow steps

Use the stats for each step card in the Flow to understand which steps shoppers select most frequently. You can also identify where shoppers drop-off and fail to continue with the Flow.

Steps with a high % of drop off may indicate an opportunity to revisit the step and make changes.

In a Flow's Analysis mode, looking at the performance stats for a specific step in the Flow to understand views and drop offs

Here’s a breakdown of the stats for the individuals steps in a Flow:

Stat	Description
View	the total number of times shoppers viewed the step (these are not unique views)
Drop off %	the percent of shoppers that reached the step and then left the Flow

Note: views for each step in the Flow are not unique. This means that a shopper may view the same part of the Flow more than once, including if they use Back and then continue forward in Flow. This may increase the total # of views for some steps.

Stats for Flow ends

By looking at the end cards in your Flow, you can understand where shoppers frequently conclude and whether the end successfully resolved their inquiry.

In other words, did the branch of the Flow automate their interaction, or did they need additional support from a member of your team?

In the Analysis for a Flow, looking at the performance stats for end steps in the Flow to understand drop offs, tickets created and the number of automated interactions with shoppers

Here’s a breakdown of the stats for ends in a Flow:

Stat	Description
Automated	The number of shoppers who reached the endpoint in the Flow and who are considered automated. This includes: - Shoppers that selected “Yes, this was helpful” at the end of a branch in the Flow OR - Shoppers who left the Flow at the last step and did not contact support in the following 72 hours
Drop off	the % of shoppers who reached the endpoint and then either: - select “No, I need more help”, but do not create a support ticket OR - if Chat is offline, they leave the Flow without providing their email address to create a ticket
Tickets	The total number of shoppers who reach the endpoint in the Flow and created a ticket. The shopper either: - selected “No, I need more help” to create a support ticket OR - the action at the end of the Flow creates a ticket by default

Was this article helpful?

Yes