---
title: "Customer journey optimization: How to find and fix customer drop-off"
description: "Customer journey optimization finds where customers stall in a live journey and changes the timing, channel, or offer. See how to measure and fix yours."
canonical_url: "https://www.braze.com/resources/articles/customer-journey-optimization"
language: "en-US"
datePublished: "2026-10-28T20:20:00.000Z"
publisher: "Braze"
authors:
  - "Team Braze"
---

# Customer journey optimization: How to find and fix customer drop-off

Customer journey optimization is the practice of measuring how customers move through a journey, finding the points where they stall or leave, and changing the timing, channel, message, or offer at those points to improve the outcome. Optimization picks up where [customer journey automation](https://www.braze.com/resources/articles/customer-journey-automation) leaves off; once a journey is built and live, you can see how customers respond and begin to fine-tune the experience to boost results.

If your journeys are live but aren't performing the way you'd hoped, this guide will help you work out why and what to do about it. We cover where journeys tend to break, how to measure them, and what to change.

**TL;DR**

- Customer journey optimization measures and improves a journey that's already live, while automation builds and runs the journey. Optimization allows you to change what isn't working.
- While a drop-off point shows where customers leave, the cause usually sits two or three steps earlier, in decisions like a welcome series that overpromises or a segment that has drifted since launch.
- Measure before you change anything: track conversion and time in each stage, compare cohort retention curves, state the denominator behind every rate, and use control groups to confirm what a change achieved.
- Personalized customer journeys keep the same structure for everyone while adjusting timing, channel, content, and offer for each customer. Generative AI creates the content variants, and decisioning picks which one each customer receives.
- The right customer journey solution combines a unified customer profile, shared data for measurement and execution, and built-in holdouts. With those in place, optimization can lift stage conversion, retention, and time to value while reducing opt-outs.

## What is customer journey optimization?

Customer journey optimization is the practice of measuring how customers move through a journey, finding the points where they stall or leave, and changing the timing, channel, message, or offer at those points to improve the outcome. The term often gets mixed up with journey mapping and journey automation, but they each describe different stages of a customer journey.

**Journey mapping** is a design exercise done before a journey exists. A team plots the stages, touchpoints, and decisions a customer is likely to move through and agrees on what the journey should achieve. The output is a plan.

**Journey automation** is where you actively build the journey and how it will run. Triggers, branches, delays, and messages are set up so customers move through it automatically based on what they do. The output is a live journey.

**Journey optimization** is the ongoing work that follows. It begins once a journey is already running and instrumented, which means the journey is sending to real customers and recording what they do at each step: who entered, who converted, who dropped off, and how long each stage took. Without that data, there's nothing to measure and no way to tell whether a change helped.

Since optimization starts from evidence about how a live journey performs, the output is a repeating loop with four steps:

1. **Measure:** Track conversion, time in stage, and retention at each journey stage against a consistent baseline.
2. **Diagnose:** Find the drop-off points and trace each one back to the step that causes it.
3. **Change:** Adjust the timing, channel, message, or offer at that step.
4. **Verify:** Compare results against a control group to confirm whether the change improved results, and then start the loop again.

## Where journeys break, and why the break is rarely where the drop-off shows

When a journey underperforms, the obvious move is to fix the step where customers leave. The trouble is that a drop-off point tells you where customers gave up, but not necessarily the reason why. Sometimes that traces back to an earlier [lifecycle stage](https://www.braze.com/resources/articles/customer-lifecycle-management), to a decision made when the journey was built but not later revisited, as we’ll explore in the sections below.

This can happen when teams build journeys around the campaign calendar, with each team scheduling its own sends. A customer-centric journey, in comparison, is planned around what each customer has done (or neglected to do), and what they need next.

### The welcome series overpromises

If your [welcome series](https://www.braze.com/resources/articles/email-marketing-customer-journey) promises quick wins but setup takes a week, new customers may run into friction they weren't expecting. Some might cancel around day 30, and your report would record it as a retention problem.

### Onboarding picked the channel, and it never changed

A customer who read every email in their first week might now mostly use your mobile app. If the journey still leads with email, engagement could drop at the next step, making the step appear as the problem, rather than the channel.

### Your segments have drifted

A segment like "viewed pricing twice" may have matched a clear group of customers at the time of launch. But over time, as new acquisition channels bring in different people, the same behavior may no longer represent the same level of intent. Conversion built around the initial segment won’t necessarily apply down the line.

### Too many messages in total

Each team might set a send frequency that makes sense for its own program, but a customer that qualifies for several programs could get multiple messages a day, and if they unsubscribe, it’s logged against the most recent one.

## Customer journey insights: How to measure a customer journey before you change it

Measure before you optimize. If you don’t know how a journey is performing, you don’t know whether a change made a difference, for better or worse. You need a baseline of insights to compare against once you begin optimizing. Here are some things to consider as you measure.

### Stage-level conversion and time in stage

Look at each journey stage separately. Track the number of customers who enter it, the number who move on to the next stage, and the average time they spend there. A low conversion rate shows you the stage where customers leave, but the time they spend in a stage can send you an earlier warning. If customers are taking much longer than usual to get through a stage, something there may be slowing them down, and you can look into it before they drop out.

Time to value (TTV) is how long it takes a new customer to move from entering the journey to the first moment your product is useful to them. You decide what that moment is. For B2C companies, it’s usually an early action that customers who stay tend to take, like placing a first order, saving a first playlist, or finishing their profile. Measure the time from entry to that action, and track whether it gets shorter as you make changes.

### Choosing the right denominator

The denominator is the number you divide by to get a conversion rate. Journey rates are often worked out using sessions, the customers who entered the journey, or the full addressable base as the denominator, and the same results can produce three very different rates.

**Example:** A hypothetical journey has 10,000 eligible customers. Of those, 4,000 enter the journey, generate 12,000 sessions between them, and 400 convert.

| Denominator | Sum | Rate | What it measures |
| --- | --- | --- | --- |
| Sessions | 400 ÷ 12,000 | 3.3% | How often a single visit leads to a conversion |
| Entrants | 400 ÷ 4,000 | 10% | How well the journey converts customers who started it |
| Addressable base | 400 ÷ 10,000 | 4% | The share of all eligible customers who converted |

When you report a journey rate, name the denominator alongside it, and stick with the same one whenever you compare results before and after a change.

### Cohort retention curves

A cohort is a group of customers who entered the journey around the same time, such as everyone who signed up in the first week of March. A retention curve shows what percentage of that group is still active at day 7, day 30, day 90, and beyond.

Cohorts are useful for checking whether a change lasts. Say you update your welcome series in April. If April sign-ups are more active at day 30 than March sign-ups were, the update is likely helping. If the two groups look the same by day 30, the update may have sped customers through the first few steps without making them more likely to stay.

Keep in mind that other things, like seasonality or a big promotion, can also change how a cohort behaves.

### Holdouts and control groups

A control group, also called a holdout, is a randomly selected group of customers who don't receive a journey, or a specific change to it. Because they're otherwise the same as everyone else, comparing the two groups tells you how many conversions the journey caused, which is also known as incremental lift. Set this up before you make any changes so that you have a fair comparison.

You can create a holdout group of customers from a single journey, or a version of it, to see whether the journey itself is worth running. But for a wider view, a [global control group](https://www.braze.com/resources/articles/using-global-control-groups-to-measure-the-impact-of-customer-engagement) holds a small percentage of customers out of all your messaging, so you can measure the overall impact of your engagement program. It's a useful next step as the maturity of your engagement program grows, and you move from testing individual journeys to measuring everything together.

## Creating personalized customer journeys without a rebuild

When a journey underperforms, your first instinct might be to redesign it. Yet a redesign takes time to build and launch, and once the structure changes, new results become difficult to compare with previous ones. Changing what varies _inside_ the journey you’ve already built is a more strategic move. It keeps the parts that work and gives you a clear before-and-after comparison. Personalized customer journeys use the same structure for everyone and adjust the details for each customer.

### Timing

A reminder sent at 9 am could be buried by the time a customer who only opens your app in the evening gets to it. Sending at the time each customer usually engages, or changing how long the journey waits after a behavioral trigger, can improve results at that stage.

### Channel

Say a customer has stopped opening your emails but still opens your app most days. Sending the next step as a push notification or in-app message gives it a better chance of being seen, and it updates a channel choice that may not have changed since onboarding.

### Content

Each step in a journey can hold several versions of the same message. Two customers might both receive a "come back" email, one featuring the running shoes they browsed and the other showing hiking boots. With [real-time personalization](https://www.braze.com/resources/articles/real-time-personalization), those details can update based on what each customer did most recently.

### Offer

It's easy to give every customer the same discount. A customer who was going to buy anyway doesn't need 20% off, while someone close to canceling might respond better to a trial extension. Matching the offer to each customer helps you save discounts for the people they're most likely to sway.

### Generative AI and decisioning

Combine the four dimensions and the possibilities multiply fast. Five content variants, three channels, four send times, and two offers already make 120 possible versions of a single step, which is far more than a team could test or assign by hand. AI can help with this, though it involves two different capabilities.

Generative AI writes and designs. Give it a brief and guidelines, and it can produce subject lines, body copy, and image variations, building a finite library of content variants in your brand's tone of voice for each journey step.

Decisioning takes action, selecting from that library for each customer. It picks which variant to send, on which channel, and at what moment, based on the outcomes it has observed so far. A customer who keeps ignoring evening emails, for example, might start getting push notifications in the morning instead.

## What to look for in a customer journey optimization solution

Tools that automate a customer journey handle the build and launch. Optimization relies on a few more capabilities, and the criteria below can help you assess your current customer journey solution or compare customer journey optimization software.

| Criteria | What to look for | What to ask |
| --- | --- | --- |
| Unified customer profile | One record per customer, updated continuously | Is the profile built ahead of time or at send time? |
| Shared data | Journeys and reporting on the same data | Can I trace a result to the step that caused it? |
| Holdouts and control groups | Control groups built in, at journey and program level | Can I hold customers out without building lists by hand? |
| Individual adaptation | Choices made per customer | Does it learn from results, or rely on fixed rules? |
| Changes without engineering | Marketers can edit and test journeys themselves | Can my team add a branch or variant without a developer? |

### A unified customer profile

A unified profile keeps each customer's activity and message history in one record that updates continuously. Every decision in a journey draws on that record, so gaps in it show up in what customers receive. Some systems assemble a profile from separate sources at the moment a message is sent. In that setup, a customer who bought something an hour ago might still get a discount for the same item.

### Shared data for measurement and execution

Shared data allows you to trace a result back to the exact step and message behind it. That gets difficult when the system running your journeys and the one reporting on them are separate. Customers may be counted in different ways or events may arrive late, and the conversion rate in your report won't line up with what the journey sent.

### Built-in holdouts and control groups

A/B tests compare versions of a message against each other. To find out whether a journey impacts customer behavior, you need a control group that receives nothing. Look for a solution that lets you set these up for a single journey or across all your messaging, without building lists by hand.

### A customer-centric journey that adapts to each individual

Segment-level journeys treat everyone in a group the same way. That works for a while, but segments change over time, and the journey doesn’t adjust unless someone updates the rules. The contrast is adapting a solution that adapts to individual customer behavior and changes what each customer receives based on the results of earlier sends and what they do next.

### Journey changes without engineering

Optimization involves frequent small edits and tests, so it helps if your marketing team can make them directly. Adding a branch or launching a new message variant shouldn't need a developer each time. Engineering still sets up data connections and integrations, but the day-to-day journey changes stay with the people and system running them.

## Customer journey benefits: What optimization actually returns

Optimizing customer journeys improves results in four main areas. Here's what each of these customer journey benefits looks like and what causes it.

### Conversion within a stage

The goal is to remove the friction point and move more customers to the next stage of the journey. Since each stage feeds the one after, early improvements can increase the overall number of customers along the conversion path. In eCommerce, checkout alone loses most of the shoppers who reach it: an analysis of 50 studies puts the average [cart abandonment rate at 70.22%](https://baymard.com/lists/cart-abandonment-rate), and the [global average visit-to-purchase conversion rate is 2.66%](https://www.shopify.com/blog/ecommerce-conversion-rate).

### Retention across stages

Improvements in earlier stages pay forward. A customer who sees value sooner, and isn't overwhelmed with messages from your brand, has more reason to continue their journey across your [lifecycle marketing](https://www.braze.com/resources/articles/customer-lifecycle-marketing) program. Mobile app engagement is generally most in jeopardy during the first month, where day-30 app retention averages [5.3% on iOS and 3.8% on Android](https://www.businessofapps.com/data/app-retention-rates/).

### Fewer unsubscribes and opt-outs

The more you keep customers from opting out, the more you can still reach in later journeys and campaigns. Unsubscribes tend to fall when the total number of messages a customer receives feels reasonable, and each one is relevant to them. An analysis of 3.6 million campaigns found a [median unsubscribe rate of 0.22%](https://www.mailerlite.com/blog/compare-your-email-performance-metrics-industry-benchmarks) per campaign, which applies to every send a customer receives.

### Time to value

A shorter TTV gives customers fewer opportunities to lose interest before they see what your product can do for them. Removing unnecessary waiting periods between steps can help, but so can prompting customers who stall during a particular set-up step. There's no reliable cross-industry benchmark here, because each team defines its own value moment, so compare against your own starting point.

## How Braze approaches optimizing customer journeys

Each part of the optimization loop needs something different: accurate data to measure and diagnose, flexible channels to make changes, and control groups to verify them. Braze brings all of that into one customer engagement platform.

If your customer data already sits in a warehouse such as Snowflake or Google BigQuery, the [Braze Data Platform](https://www.braze.com/product/braze-data-platform) can build audiences and trigger journeys straight from it, without duplicating the data. Profiles also pick up app, web, and backend events as customers act. When you're measuring a journey or tracing a drop-off to its cause, your numbers and your audiences come from the same place, which makes it easier to tell a real problem from a data mismatch.

Changing a channel can be slow when each one runs in a different tool. In Braze, email, push, in-app, SMS, and WhatsApp are all built into the platform and draw on the same customer profile. So, if your diagnosis shows customers ignoring email at one stage, you can switch to push or in-app without moving anyone to another system.

To verify a change, you can test a whole journey, or a single new version of it, against a control group that doesn't receive it. Braze reports the difference as incremental lift, so you know how much of the result came from the journey itself.

[BrazeAI™](https://www.braze.com/product/brazeai) helps you make more changes with less manual effort. BrazeAI Operator™ can build on-brand content, campaigns, and Canvases, and Intelligent Timing and Intelligent Channel choose when and where to reach each customer based on their past engagement.

For enterprise teams managing journeys across many products, offers, or regions, [BrazeAI Decisioning Studio™](https://www.braze.com/product/brazeai-decisioning-studio) makes an individual decision for every customer, optimizing toward a business KPI you define, like revenue or retention. Its driver trees show which parts of that KPI moved, such as conversions per customer or revenue per conversion. That helps with diagnosis and verification, because you can see whether and why a result changed.

_Ready to see the full loop in one place? Explore how [journey orchestration](https://www.braze.com/product/journey-orchestration) works inside a single customer profile._

## Customer journey optimization FAQs

### What is customer journey optimization?

Customer journey optimization is the practice of measuring how customers move through a journey, finding the points where they stall or leave, and changing the timing, channel, message, or offer at those points to improve the outcome. Each change is treated as a test, measured against outcomes such as conversion, retention, or satisfaction.

### How is customer journey optimization different from customer journey automation?

Customer journey automation is how you build and run a journey; customer journey optimization improves it after it’s live. Automation uses triggers and predefined workflows to send messages or take actions without manual effort, such as a reminder after an abandoned cart. Optimization measures how an [automated customer journey](https://www.braze.com/resources/articles/customer-journey-automation) performs and changes what isn't performing.

### What metrics show that a customer journey needs optimizing?

The clearest signs are high drop-off at a specific stage or step, falling conversion rates between stages, and a lengthening time to convert. Rising churn, fewer repeat purchases, and declining satisfaction or effort scores also point to friction. Track each stage separately, because a single overall conversion rate shows that a problem exists but not where customers leave.

### How often should a customer journey be reviewed?

Review each core journey at least quarterly, and check a change at day 7, day 30, and day 90 as each cohort reaches those points. Review sooner when something significant changes, such as a new product, price, channel, or market. Monitor conversion and drop-off continuously so problems show up between reviews.

### Do you need a separate tool to optimize a customer journey?

Not necessarily, but you do need two distinct layers. Journey analytics measures how customers move across channels and shows where they stall or leave. A customer engagement platform acts on those findings by changing the timing, channel, message, or offer customers receive. Some platforms combine both, while others focus on one layer and connect to the other.

---

*Published 2026-10-28T20:20:00.000Z · canonical lives at https://www.braze.com/resources/articles/customer-journey-optimization*