How to Build a Hyper-Personalized Customer Journey with Braze

A hand tying green raffia around cream and green stems wrapped in kraft paper on an oak workbench, with a small blank gift label beside them.

A hyper-personalized customer journey with Braze combines Liquid profile data, catalogs and Connected Content, then adds defaults, aborts and retries so every message still makes sense when data is missing. This guide sets out which data layer suits each decision, what happens when a call fails, and the order CustomerIK recommends building it in.

Key Takeaways

  • A hyper-personalized journey chooses content from data about the person and still behaves sensibly when that data is missing.
  • Profile data is read with Liquid, reference data sits in catalogs, and Connected Content suits data that must be current at the moment of sending.
  • A default value shows for empty values but not for blank values, so a field holding only a space passes through.
  • Abort when the personalised element is the reason for the message. Aborted sends appear in Messaging Observability as Liquid abort.
  • With retry set, a failed Connected Content call aborts a Message step but removes the user from the Canvas in a Context step.
  • We recommend sorting every personalised element as garnish or reason before anyone builds the journey.

What makes a hyper-personalized customer journey with Braze work in practice?

Braze documents Liquid as an open-source template language created by Shopify and used by Braze to power dynamic personalization. When a message is sent, Braze scans the text for Liquid. It pulls the data for that user and swaps in the actual value.

We use a working definition. A journey is hyper-personalised when each step chooses its content from data about that person, and still behaves sensibly when that data is missing. The build effort belongs in the second half of that sentence.

Which data layer should each personalisation decision use?

Braze gives you three places to hold what a message needs. Liquid reads profile data, catalogs hold reference data, and Connected Content fetches information accessible via API. The catalogs documentation puts the split plainly: catalogs hold the data, and Liquid pulls it out.

The last column below is our suggestion, not Braze guidance. The middle column is documented behaviour.

LayerHoldsDocumented behaviourOur rule of thumb
Profile attributes read with LiquidData about one personResolved for that user before the message is sentUse for anything the person told you or did
CatalogsReference data such as products or locationsStorage that Liquid reads fromUse for anything shared across many people that changes on your schedule
Connected ContentAnything an API can returnRuns at render time, response under two seconds, 1 MB response limitUse only when the data must be current at the moment of sending

Personalisation uses personal data. This post describes documented mechanics only and does not settle consent or data protection questions, so involve your privacy team before you pass profile fields to an external endpoint.

What should a message show when the data is missing?

Braze states that without a default value, a missing field will be blank in the message. The fix is the default filter. It works for strings, booleans, arrays, objects and numbers.

Hi {{ ${first_name} | default: 'Valued User' }}, thanks for using the App.

One documented detail catches people out. The default value shows for empty values, but not for blank values. A field holding only a space passes straight through.

Our suggestion: check imported attributes for whitespace before you trust a default. If blank values keep arriving, the cause sits upstream, and CustomerIK's customer data management service covers data governance and identity resolution.

When should a message abort instead of falling back?

Braze lets you stop a send with abort logic, and you can attach a reason. The aborted send appears in Messaging Observability with the outcome Liquid abort, and Braze does not count aborted messages toward the send count in your account or in Currents.

We sort every personalised element into one of two kinds. If it is garnish, such as a first name, use a default. If it is the reason for the message, such as the recommended item, abort when it is missing.

Because aborted sends drop out of Currents counts, reconcile your reports against Messaging Observability. Do that before anyone reads a gap as a delivery fault.

How should Connected Content be used inside a journey?

Connected Content lets you insert information from an API directly into messages. It runs when the message is rendered, and for in-app messages that means at impression time. Braze also documents that content is not inserted if the server takes longer than two seconds to respond.

Some behaviours shape how you plan the endpoint. Get requests are cached by default. Braze applies no separate rate limit to Connected Content, so request volume scales directly with your delivery speed, and the endpoint owner should see your planned send speed before launch.

What happens when a Connected Content call fails in a Canvas?

The retry documentation separates Canvas step types. Connected Content calls in a Canvas retry only when the call includes the retry tag, and the difference between step types matters when you decide where a risky call lives.

Step typeDocumented behaviour when every retry fails
Message stepRecipients are delayed on the send queue while Braze retries up to five times. The message is aborted and the user advances to the next step
Context and User Update stepsBraze retries at the step level, up to five times. The user exits the Canvas

Two further facts are worth holding together. If a Connected Content call uses abort logic for the same condition as the retry logic, the abort logic takes precedence, so retries never start. A 404 from your URL renders an empty string in its place.

We read these together as a placement rule. A failed call in a message step costs you one message. A failed call in a Context step costs you the user's place in the journey, so decide which loss you can accept before choosing where the call sits.

How do catalogs and selections keep recommendations relevant?

A selection filters a catalog so a message only draws from items that meet criteria you define. Braze allows up to 30 selections per catalog and up to 10 filters per selection. When a geolocation filter is applied, results are sorted by distance with the nearest item first.

Our suggestion: name each selection after the decision it serves. Retire it when the journey that used it is retired.

What does a worked example look like?

This example is hypothetical and is not client work. Take a hypothetical retailer sending a price-drop message from a Canvas. The first name is garnish, so it gets a default value.

The price is the reason for the message. If the catalog returns nothing, the message aborts with a plain reason such as "no price data".

Suppose the retailer also pulls a live stock figure from its own endpoint. That call sits in a message step with retry, so a failure skips one message instead of removing the user from the Canvas.

Showing a generic offer instead of aborting is a business decision. The retailer's owner should confirm it before launch.

In what order should you build it?

This sequence is our practice, not a Braze procedure.

  1. List every personalised element in each step and mark it as garnish or reason.
  2. Read profile data with Liquid and add a default to every attribute you mark as garnish.
  3. Move shared reference data into catalogs, then define selections for the decisions that need them.
  4. Add Connected Content last, only where data must be live, and choose retry and abort per step type.
  5. Test every message that uses Liquid, because Braze recommends testing all of them before sending.
  6. After launch, check Messaging Observability for Liquid abort outcomes and investigate any that you did not expect.

How does this connect to the rest of the programme?

This post covers the mechanics of a single journey. For the wider approach, read the hyper-personalization guide for Braze campaigns and the personalization strategy from planning to execution.

For the data side, see the guide to Braze Connected Content and the piece on what resolves when in real-time personalisation. If a journey misbehaves, start with the common personalisation pitfalls and the guide to complex Canvas workflows.

For fallback logic across channels in a Canvas, a CustomerIK case study on unifying guest profiles for a hotel group covers consolidating profile data and an automated fallback sequence.

As a Braze implementation partner, CustomerIK recommends deciding abort or fallback for every personalised element before anyone writes a template.

Frequently Asked Questions

1. What is the difference between a catalog and Connected Content?

A catalog stores data inside Braze and Liquid reads it. Connected Content fetches information from an API when the message is rendered.

2. Do default values replace a field that contains only a space?

Braze documents that the default value shows for empty values, but not for blank values, so a space passes through.

3. What happens to a user if Connected Content fails in a Message step?

Braze retries up to five times when retry is set. If every attempt fails, the message is aborted and the user advances to the next step.

4. What happens if Connected Content fails in a Context step?

Braze retries at the step level, up to five times. If every retry fails, the user exits the Canvas.

5. Where can I see aborted sends?

They appear in Messaging Observability with the outcome Liquid abort, and they are not counted toward the send count in your account or in Currents.

Sources

CustomerIK is a Braze implementation partner working across onboarding, technical integration, marketing operations and customer data management.

If you want a journey reviewed for what it does when data is missing or a call fails, let's talk.

Preferences

Privacy is important to us, so you have the option of disabling certain types of storage that may not be necessary for the basic functioning of the website. Blocking categories may impact your experience on the website.

Accept all cookies
Accept all cookies

These items are required to enable basic website functionality.

Always active

These items are used to deliver advertising that is more relevant to you and your interests.

These items allow the website to remember choices you make (such as your user name, language, or the region you are in) and provide enhanced, more personal features.

These items help the website operator understand how its website performs, how visitors interact with the site, and whether there may be technical issues.

Thank you! Your submission has been received!
Oops! Something went wrong while submitting the form.