
Retention surveys in Braze are built as survey landing pages or in-app message surveys, and email reaches them through a Net Promoter Score block. Answers can be logged as custom attributes, used to segment or trigger follow-up messaging, and exported through Currents. Plan the question order, the logged attributes and the follow-up before you build anything.
Key Takeaways
- Braze surveys run on landing pages and in-app messages, and email reaches a landing page survey through a Net Promoter Score block.
- Only answers logged as custom attributes can be used to segment users, and long text answers can't be logged.
- We recommend asking for the reason in a radio button group, with long text as an optional last question.
- An email score tap is a partial response until the final submit, and partial submissions have a six-hour grace period before they appear in results.
- Survey Response events reach Currents and a SQL table, and you derive promoter, passive and detractor groups from the score.
- Braze's release notes listed the rating scale as early access on 23 Jul 2026, so confirm it's on in your workspace.
How do you run retention surveys in Braze?
Braze surveys let you collect first-party feedback from your users and act on it in follow-up messaging. You never leave the dashboard to do it. Surveys run on landing pages and in-app messages, and email can send people to a landing page survey with a score block.
The order below is our recommendation, not Braze's documentation. Steps one, five and six come before the build because they decide whether a survey changes anything.
- Write down the decision the answers will change, before you open the editor.
- Choose the channel: a landing page, an in-app message, or an email block that links to a landing page.
- On the message composition page, select Survey as your message type before you open the editor.
- Add the form blocks you need and give each one an Identifier for Reporting.
- Decide which answers you will log as custom attributes, because that is what lets you segment by them later.
- Build the follow-up campaign or Canvas, triggered when someone completes the survey.
- After launch, review the survey analytics, and export through Currents if you need to join the answers to other data.
Braze documents the mechanics behind steps three, four and seven. The rest is our sequencing advice.
Which survey channel suits a retention question?
We recommend choosing the channel by where the person is when they answer. Landing pages and email reach people outside your app. In-app message surveys ask them while they're inside it.
| Channel | How a respondent answers | Where you review results | Documented detail |
|---|---|---|---|
| Landing page | Opens a survey landing page, which can be a single page with multiple linked steps | The landing page analytics view | Long text answers can't be logged as custom attributes |
| In-app message | Answers inside the message, built in the drag-and-drop editor | The Responses tab | Minimum SDK versions: Swift 5.0.0+, Web 3.3.0+, Android 20.0.0+ |
| Selects a score tile in the NPS Survey block, then lands on a published survey landing page | The linked landing page's survey reporting, and Messaging then Surveys | One NPS Survey block per email; the landing page needs an NPS question |
Braze's release notes list the email Net Promoter Score block as generally available, with an entry dated 15 Oct 2026. If you don't see Surveys under Messaging in your workspace, the documentation says to contact your Braze account manager.
Which question types suit retention feedback?
The Net Promoter Score block is a separate form block from the Rating block, not a setting inside it. Braze groups the answers into promoters, passives and detractors for you. It shows those groups and the resulting score in the survey analytics view.
A number scale question captures sentiment, satisfaction or likelihood to recommend as a single number. You can record the rating as a survey response, log it as an integer custom attribute, or do both. Braze's release notes listed the rating scale as early access on 23 Jul 2026, so confirm it's switched on in your workspace before you plan around it.
Long-form text can carry the reason behind a score, and Braze lets you set a maximum of up to 1,000 characters. Long-form text fields in iOS in-app message surveys are temporarily limited to 250 characters. Keep the maximum at or below that for iOS audiences.
For retention work we recommend asking for the reason in a radio button group. Long text works as an optional last question. A radio group can drive a segment or a branch, while long text can only be read.
Turn on Randomize choice order for the reason list. Randomisation changes only the display order for each respondent. Reporting and exports stay mapped to the choices you configured.
What happens when someone answers only the first question?
An email score tap creates a response before the person has finished anything. Braze opens the landing page and records the score on arrival. The recipient can then answer any remaining questions.
The response stays partially complete until they select the final submit control, even when the score is the only question. Braze defines partially complete as users who submitted some data but did not complete all required questions. Partial submissions have a six-hour grace period before they appear in results.
If a recipient changes their score and then submits the rest of the survey, Braze uses the later score. The email score is applied to the first Net Promoter Score question on the page. In a multi-step survey, that question must be on the first step.
We recommend treating a partial response as a sign that someone engaged, not as a reason. Don't trigger a message that assumes a reason you haven't collected.
How do survey answers drive follow-up messaging?
You can segment users by survey responses that are logged as user attributes. You can also segment by survey completion status. Campaigns and Canvases can be triggered when a user completes a survey, on a landing page or in an in-app message campaign.
Each form block takes an Identifier for Reporting, and that identifier appears in survey reporting and exports. Braze documents that responses can be logged to custom attributes, with long text called out as one that can't.
If you choose not to log a response, you can't segment users by that response value. For in-app message surveys, Braze notes that question-and-answer triggering that doesn't rely on logged user attributes is not available.
In practice, that makes the logged attribute your branching key. We recommend naming it and planning its values at build time.
How do you analyse retention survey results beyond the dashboard?
Braze exports each survey answer through the Survey Response event in Currents. Each event represents one respondent's answer to one survey question. The same data is available as a SQL table in Query Builder, SQL Segment Extensions and Snowflake Data Sharing.
The promoter, passive and detractor groups appear in the dashboard. They aren't separate Currents fields or SQL columns, so you derive them from the score.
The query below is Braze's documented example. Put your own question reporting identifier in place of the placeholder.
SELECT
question_reporting_id,
CASE
WHEN answer_single_number >= 9 THEN 'promoter'
WHEN answer_single_number >= 7 THEN 'passive'
ELSE 'detractor'
END AS nps_segment,
COUNT (*) AS response_count
FROM USERS_MESSAGES_SURVEY_RESPONSE_SHARED
WHERE answer_type = 'nps'
AND question_reporting_id = 'REPLACE_WITH_YOUR_NPS_QUESTION_REPORTING_ID'
GROUP BY question_reporting_id, nps_segment;Landing Page Impression events are anonymous. You can't join them to survey responses on user ID to calculate a response rate.
Completing a landing page survey writes a Survey Response event. It doesn't fire the generic Landing Page Form Submission event.
Open-ended answers can carry personal data. Braze's documentation says to consider your own PII handling policy before exporting them.
What does a retention survey look like in a worked example?
Take a hypothetical subscription fitness app whose retention team wants to know why people go quiet after their first month. This is an illustration, not client work, and it carries no results.
The team builds one multi-step survey landing page. Step one holds the Net Promoter Score block, because the email score can only apply to a question on the first step. Step two is a radio button group of reasons with randomised choice order, and step three is an optional long text box.
They send an email to a segment of quiet users, with the Net Promoter Score block linked to that page. The reason answer is logged as a custom attribute with a clear Identifier for Reporting. A Canvas triggered on survey completion then branches on that attribute, with one path for each reason.
Someone who only tapped a score has a partial response and no logged reason. The trigger is survey completion, so they never enter the reason branches. Nobody gets a message about a problem they haven't named.
Where do retention surveys go wrong?
The first failure is asking for the reason only in long text. The answers are readable and exportable. You can't segment users by them, so the follow-up ends up generic.
The second is a build-order slip. The email block only lists published survey landing pages that include a Net Promoter Score question. A page that is still unpublished won't appear in the picker.
The third is reading completions as the whole audience. Partial responses are real people who engaged. Look at them separately before you decide the survey didn't land.
How does this connect to the rest of the programme?
A survey is only as useful as the measurement around it. Our guide to retention metrics covers what to track alongside survey answers. For acting on a logged attribute, see segmentation for retention campaigns.
Timing and content come next. See triggering retention messages at the right time for send timing. Personalising retention messaging covers the follow-up itself.
The earlier post on feedback surveys for onboarding covers the first-session version of this method. Retention campaign mistakes to avoid lists the errors that tend to follow a survey.
As a Braze implementation partner, CustomerIK recommends writing the follow-up for each survey answer before you write the survey. Every question you ask then has a message waiting for it. Braze's own survey documentation is at Braze Docs: Surveys.
Frequently Asked Questions
1. Can I run a retention survey from email alone?
Not on its own. The email block links each score tile to a published survey landing page. That page needs a Net Promoter Score question, so the survey itself lives on a landing page.
2. Where do I find survey results across channels?
Go to Messaging then Surveys to see surveys across landing pages, campaigns and Canvases. If that page isn't in your workspace, the documentation says to contact your Braze account manager.
3. Can I segment users by what they wrote in a long text answer?
Long text responses can't be logged as custom attributes. You can't segment users by a long-form response value directly.
4. How long before a partial response shows in results?
Partial submissions have a six-hour grace period before they appear in results. People can finish the rest of the survey first.
5. Does randomising answer choices change my reports?
It changes only the display order for each respondent. Reporting and exports use the same response data whichever order a person saw.
Sources
- Braze, Surveys. Retrieved 09 Nov 2026.
- Braze, In-app message surveys. Retrieved 09 Nov 2026.
- Braze, Landing page surveys. Retrieved 09 Nov 2026.
- Braze, Email NPS survey block. Retrieved 09 Nov 2026.
- Braze, Release notes. Retrieved 09 Nov 2026.
- Written by CustomerIK, a Braze implementation partner.
CustomerIK is a Braze implementation partner working across onboarding, technical integration, marketing operations and customer data management.
If you're planning a survey and want help mapping each question to its follow-up, let's talk.




