Skip to main content
ComponentsEngineRUM JS

Reach Customers Before They Leave

Detect, alert and respond

Connect the Signals

One customer across channels

Find Why Customers Cancel

One customer or all customers

Set It Up

What to deploy and configure

Overview

A customer who cancels a subscription, a contract or a policy rarely does it on impulse. First they look for an answer on your portal, open a chat that doesn’t help, or leave a frustrated comment. Then they open the cancellation page. The signals are there, but they sit in different tools, and your retention team hears about the customer only after they’re gone. GermainUX watches the whole journey in real time: the portal, the cancellation flow, the chat with your agents or chatbot, the feedback and the services behind them. When a customer starts to cancel, can’t get help or tells you they’re unhappy, GermainUX alerts your retention team with the customer’s context, so the team can reach out while there’s still time. At scale, it shows which journeys, frictions and failures drive cancellations.
Portal (Web, Mobile, Self-Service)Chat and AgentsFeedbackCancellation JourneyRetention Team
The same approach works for any business where customers can cancel online: insurance, telecom, utilities, banking, media or software subscriptions. This page uses a Salesforce Experience Cloud portal with Genesys chat as the example. An affinity-insurance provider uses this approach to email its client service team, in real time, the details of clients who started but didn’t complete a cancellation, with a link to their recorded session.

Questions It Answers

  • Which customers started to cancel today but haven’t finished?
  • Which customers asked for help and didn’t get it?
  • What did a customer try to do before opening the cancellation page?
  • Which frictions, errors or unanswered questions come before most cancellations?
  • Is the chat or the contract lookup failing for customers right now?
  • Did our outreach keep the customer?

How Each Signal Is Monitored

GermainUX monitors each part of the journey with the components that fit it. All the data lands in one place, linked to the customer’s session.

Reach Customers Before They Leave

Each signal can raise an alert to the team that can still change the customer’s mind. The response follows the same three steps for every trigger:
  1. Detect. GermainUX recognizes the trigger in the customer’s session, chat or feedback.
  2. Alert. An alert by email or SMS, Microsoft Teams or Slack tells the team who the customer is, what they tried to do, the likely reason, and links to their recorded session.
  3. Respond. Someone from the team calls or emails the customer with help or a better offer. An automated action can also open a follow-up task in your CRM or a ticket in ServiceNow, so no customer is forgotten.
Send each trigger to one team, and alert on each customer only once, so the alerts stay actionable. Alerts and replays contain personal data: mask or exclude what your teams don’t need, such as payment details, and send only the context the team needs to reach the customer.

Connect the Signals to One Customer

A useful alert needs to know who the customer is. GermainUX links the signals through the customer’s session and their business identifiers: To capture the customer or contract ID, add a metadata extractor that reads it from the page or from an API response. Without it, the alert says what happened but not to whom.

Prevent Avoidable Cancellations

Some customers cancel because something on your side failed. GermainUX tests the paths they rely on and watches the signals that come before failures.

Test the Paths Customers Rely On

A synthetic user scenario signs in to the portal on a schedule, opens a contract and starts a chat, the way a customer does. Add HTTP checks for the contract lookup, billing and chat services. When a step fails or slows down, GermainUX alerts you before customers meet it.

Watch the Early-Warning Signals

Set SLAs and outlier detection on these signals, so a growing problem raises an alert before it turns into cancellations.
Analysis of failed portal requests during a two-hour outage, with the contract lookup API and 503 and 504 responses among the Leading Factors, and user-facing errors at 4 times their normal level

Failed Contract Lookups During an Outage - GermainUX

To see what sets the failed requests apart, compare them with the successful ones in the same window. In this example, a failed request takes 11.4 seconds against 0.18, most of it waiting for the API, and customers see 102 user-facing errors against 13.
Compare page of failed and successful portal requests during the contract lookup outage, 11.4 seconds against 0.18 seconds on average, with wait duration and download duration at the top of Top Differences, followed by user-facing errors and the contract lookup API

Failed and Successful Portal Requests Compared During an Outage - GermainUX

Analysis of chats that no agent answered during one evening, 13 against 5 at the same hours on the previous days, with Leading Factors by device, operating system and page

Chats Without an Agent on an Understaffed Evening - GermainUX

Find Why Customers Cancel

The alerts help one customer at a time. The analysis shows what drives cancellations across all of them.

One Customer: Session Replay and AI Summary

Open the customer’s session from the alert. The AI session summary tells you what the customer tried to do and where they struggled. Session Replay shows the pages they visited, the errors they saw, the chat they had and the moment they opened the cancellation page. The person who calls the customer knows what went wrong before picking up the phone.

All Customers: Journeys, Conversations and Feedback

Open the Analysis Dashboard on the cancellation journey:
  • Leading Factors show which regions, devices, browsers or pages cancel more than the others. To compare plans or products, filter the analysis on each one, or put them side by side on the Compare page.
  • Customer journeys and drop-offs show the paths customers took before they cancelled, and where the ones who stayed went instead.
  • Conversation categorization groups the chats before a cancellation by topic, such as prices, billing or a problem nobody solved.
  • Feedback analysis groups the comments the same way.
For example, after a price increase on one plan, filter the analysis on the customers of that plan who opened the cancellation page. In this example, 150 Plus customers open it in the week after the price increase, against 65 the week before.
Analysis of Plus customers who opened the cancellation page in the week after the price increase, 150 against 65 the week before, with daily starts on the Relationships chart

Plus Cancellation Starts After a Price Increase - GermainUX

To check that only Plus moved, compare the customers of the other plans with the week before. In this example, 94 Basic and Premium customers open the cancellation page, against 98 the week before.
Compare page of Basic and Premium customers who opened the cancellation page, 94 in the week after the Plus price increase against 98 the week before

Cancellation Starts of the Other Plans Before and After the Price Increase - GermainUX

To check why, filter the negative feedback on its category. In this example, 132 customers leave negative feedback about price in the week after the increase, against 9 the week before. Nothing technical is broken, so this answer goes to the pricing and retention teams.
Analysis of negative feedback in the Price category, 132 in the week after the Plus price increase against 9 the week before, with daily feedback jumping from 11 November on the Relationships chart

Negative Feedback About Price After the Plus Price Increase - GermainUX

The investigation moves from the population to one customer and back:
Cancellations RiseAggregate AnalysisLeading FactorOne CustomerSession Replay

Set It Up

Each step adds value on its own. Start with the portal and the cancellation journey, then add the other signals.
1

Monitor the portal

Deploy monitoring for Salesforce Experience Cloud, or RUM JS for your website. Configure privacy, and add a metadata extractor for the customer or contract ID.
2

Model the cancellation journey

Configure a business process for the cancellation flow, keyed on the customer or contract ID, with a timeout that marks the customers who start but don’t finish. Send the customer’s plan or product with each step, so the analysis can compare them.
3

Monitor the chat

Turn on chat and CTI monitoring for Genesys or your chat platform, and AI conversation analysis for the transcripts. Add Phrases of Interest for the words customers use when they want to leave.
4

Turn on the feedback popup

Enable the Feedback Popup on the portal, and decide who owns each kind of feedback.
5

Alert the right team

Create alerts for each trigger, with the customer’s context, and send each one to the team that owns it. Add synthetic scenarios and SLAs for the paths customers rely on.

Additional Resources

Chat Analysis

Agents, chatbots and outcomes

Salesforce Experience Cloud

Portals, journeys and feedback

Use Cases

More Use Cases

Get Help

Every business has its own cancellation flow, channels and retention offers. The GermainUX team can help you model the journey, choose the triggers and set up the alerts that reach your retention team in time.

Contact GermainUX Support

The GermainUX team helps you get up and running faster, empower your team, turn data into actionable insights, and optimize performance to deliver measurable business outcomes.

Can't Find What You Need?

Search all GermainUX docs with Ctrl + K, or use the feedback buttons at the bottom of this page to tell us what is missing.

The content of this document is proprietary to GERMAIN SOFTWARE.