> For the complete documentation index, see [llms.txt](https://docs.askyazi.com/llms.txt). Markdown versions of documentation pages are available by appending `.md` to page URLs; this page is available as [Markdown](https://docs.askyazi.com/study-types/surveys.md).

# Surveys

How a Yazi survey works on WhatsApp, what participants see, and how to design one that people finish.

A survey asks every participant the same set of structured questions, one after another, inside their own WhatsApp chat. Use a survey when you need numbers you can count, compare and track: usage and attitudes, satisfaction, screeners, brand tracking or concept scores.

{% hint style="info" %}
**In short:**

* Participants answer in WhatsApp with no app, browser or login. Every answer is saved the moment it is sent, so a partial complete still gives you data.
* People can stop and come back later. When they return, the survey carries on from the next question they have not answered.
* Aim for 22 to 28 questions. Drop-off rises noticeably beyond about 32 questions, and long grids are the biggest cause.
  {% endhint %}

## What the participant sees

{% stepper %}
{% step %}

### They join

The participant taps your study link or QR code, or taps a button on a WhatsApp invite you sent them. Joining by link opens WhatsApp with a short starter phrase already typed, for example "Send to start #4F9A1", which they send to begin. See [Study links](https://docs.askyazi.com/reach-participants/study-links.md).
{% endstep %}

{% step %}

### They see your opening message

The first message explains who is running the research and why. If your study runs in more than one language, they pick their language first. See [Languages and translation](https://docs.askyazi.com/study-types/languages-and-translation.md).
{% endstep %}

{% step %}

### They answer one question at a time

Each question arrives as a WhatsApp message. Depending on the question type, they tap a button, pick from a list, fill in a short WhatsApp form, type a reply, or send a voice note, photo, video, document or location pin. Some questions can show as a WhatsApp Flow, a form screen that opens on top of the chat. See [In-chat vs WhatsApp Flow](https://docs.askyazi.com/build-a-study/in-chat-vs-whatsapp-flows.md).
{% endstep %}

{% step %}

### Wrong answers are re-asked

If someone sends the wrong kind of answer, for example a photo when you asked for a video, or a letter that is not one of the options, Yazi tells them the answer was not accepted and asks again, in their own language. These attempts show in the transcript but are not saved as answers.
{% endstep %}

{% step %}

### They receive your closing message

When they reach the end, they get your completion message. You can use it for a thank-you, payment instructions or a link. Instead of a completion message, a survey can also end with a closing broadcast, a redirect to another survey, or a reward.
{% endstep %}
{% endstepper %}

<figure><img src="https://358667285-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2FzMRQUXiGQlBwWdg6WEmG%2Fuploads%2F1e15xRoYL7719MpRpxxt%2Fwhatsapp-flow-question.webp?alt=media&amp;token=318ecb52-d112-4e04-ae50-61f4dd70c2ff" alt="A survey question shown on the participant&#x27;s phone as a WhatsApp Flow screen with a progress bar"><figcaption></figcaption></figure>

## Build a survey

You build a survey on the **Build** tab. Each question is a row with its number, text and type, plus icons for logic, settings, preview, duplicate and delete. Use the **+** buttons between rows to insert a question, and drag rows to reorder them. For every question type and what it looks like on the phone, see [Question types](https://docs.askyazi.com/build-a-study/question-types.md).

Survey-wide settings live in the **Study configuration** panel on the right of the **Build** tab, under **Settings**:

| Setting                    | What it controls                                                                                                                                                                                                   |
| -------------------------- | ------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------ |
| **Participant experience** | **Builder Mode** (**In-chat survey** or **Flow-based survey**), **Campaign Language** (the language of system messages), **Translate responses into**, **Multi-Language Support** and **Hide Question Percentage** |
| **Completion & follow-up** | How the survey ends: **Send a completion message**, the **Completion message** text and an optional button, plus advanced endings such as a closing broadcast, redirect or reward                                  |
| **Quota**                  | Caps per answer option. See [Quotas](https://docs.askyazi.com/build-a-study/quotas.md)                                                                                                                             |
| **Response quality**       | **Response Trust** scoring, its **Sensitivity** (Lenient, Balanced, Strict) and AI answer checks. See [Response Trust](https://docs.askyazi.com/monitor-fieldwork/response-trust.md)                               |
| **Privacy & screening**    | **Privacy Protection** (hides who said what), the **Screen-Out Message** sent when someone is disqualified, and the **Repeat Attempt Message** sent when someone tries to retake a survey they cannot retry        |

<figure><img src="https://358667285-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2FzMRQUXiGQlBwWdg6WEmG%2Fuploads%2FJiYmLK08y9zg4BzndOHI%2Ffs-build-tab-survey.webp?alt=media&amp;token=f7695c99-34b5-4033-8dd5-c17195441e43" alt="The Build tab of a survey: questions on the left, Study configuration on the right, and the Build, Share, Responses and Reporting tabs at the top"><figcaption></figcaption></figure>

### Opening and closing messages

* **Opening message.** Say who you are, why you are asking, roughly how long it takes, and what the participant gets for taking part. Naming the brand or organisation they know helps people trust the message and reduces the chance they block your number.
* **Completion message.** This appears as a locked row pinned to the bottom of your question list. Click it to edit the wording. It cannot be moved or deleted because it always comes last. You can also set it in **Study configuration > Completion & follow-up**: turn **Send a completion message** on or off, edit the **Completion message** (clear the text to use the default), and choose **Add a button**: **No button**, **Link button** or **Quick-reply**. If the survey ends a different way (a closing broadcast, a redirect or a reward), the row shows that ending instead, read-only, with a note saying where it is set. For Flow-based surveys with the completion message turned off, the last question becomes the final step.

<figure><img src="https://358667285-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2FzMRQUXiGQlBwWdg6WEmG%2Fuploads%2FKSiNGUD5vZxRzySiJTjw%2Ffs-completion-followup.webp?alt=media&amp;token=2beea651-b46e-4ba2-b553-809be9581a67" alt="Study configuration > Completion &#x26; follow-up: the completion message and an optional button"><figcaption></figcaption></figure>

{% hint style="warning" %}
Write your completion message in the builder (the locked row or **Completion & follow-up**), not in the dialog you see when going live, so it is saved with the study's settings.
{% endhint %}

### Statements

A **Statement** is a message with no answer, for example instructions or a section introduction. The participant taps **Continue** to move on, so a run of statements arrives one at a time rather than all at once. You can convert a statement into a question, and a question into a statement, from the question type.

### Optional questions

Mark a question as optional and participants can reply "Skip" (in their own language). The skip is recorded as a skip, not as a blank.

## Progress shown to participants

Yazi can show participants a running percentage of how far through they are. The percentage counts up steadily through the whole study, including when a survey hands over to an AI interview, and moves further along when logic skips a section.

You control it with the **Hide Question Percentage** toggle in **Study configuration > Participant experience**. Check this setting on every new study so participants see, or do not see, the percentage as you intend.

{% hint style="info" %}
Do not type your own counter into question text, such as "Q5 of 12". It never updates when you add or remove questions, so it soon tells participants the wrong total. The builder warns you when it spots one.
{% endhint %}

## Pausing and coming back

WhatsApp chats do not expire, so a participant can leave halfway and return hours or days later. When they come back they get a welcome-back message with a **Continue** button. Tapping it takes them to the next question they have not answered, with all their earlier answers kept.

Two exceptions:

* A **looping question** (one that repeats per item) restarts the loop rather than guessing where the person was, so nothing in the loop is skipped by mistake.
* A small number of very old half-finished sessions may restart from the first question. Their earlier answers are still kept.

If you need to reach people who stopped partway, use [Reminders](https://docs.askyazi.com/reach-participants/reminders.md), or filter the **Share** tab by **In survey** and send them a broadcast. See [Send a broadcast](https://docs.askyazi.com/reach-participants/send-a-broadcast.md).

## Design guidance

### Length

| Guide                              | Questions |
| ---------------------------------- | --------- |
| Short pulse or screener            | 8 to 12   |
| Standard survey (the sweet spot)   | 22 to 28  |
| Point where drop-off rises sharply | about 32  |

In time, aim for about 15 minutes of answering, with 20 as a ceiling. Because people answer in short bursts and come back, elapsed time is often longer than answering time. There is no hard limit on question count, so these are guidelines, not rules.

### Order

* Put the questions that matter most early. Treat everything after them as a bonus, because every answer is saved as it arrives.
* Put classification questions (age, region and so on) at the end, unless you need them for routing or quotas. If you recruit through a panel, some profile fields may already come with the participant. See [Screening and incidence](https://docs.askyazi.com/build-a-study/screening-and-incidence.md).
* Open text and voice note questions have higher drop-off. Place them where the participant is most engaged.

### Wording for a phone screen

* Use the participant's everyday words, not your internal product names or acronyms.
* Keep question text short. Long statements and multi-part questions are hard to read on a phone.
* Ask for one thing per question. Where you expect a long answer, add "please answer in one message", because WhatsApp users often send several short messages and only the first is captured as the answer.
* If a media question needs a specific format, say so. Yazi adds a default instruction line to media questions (for example "Please reply with a video") so people know how to answer.
* Turn on option randomisation for lists where order could bias the answer.

### Stimulus

Images attached to a question are sent as WhatsApp messages in the chat. They are not pinned on screen while the participant answers later questions, but the participant can scroll back to them. If several questions refer to one image, open with a statement saying "the next few questions are about this image", and resend the image every second or third question.

## Limits and things to know

* **WhatsApp's own format limits apply.** Button counts, option lengths, multi-select and matrix limits are on [WhatsApp limits](https://docs.askyazi.com/build-a-study/whatsapp-limits.md).
* **No ranking question.** Use the piping workaround described on [Question types](https://docs.askyazi.com/build-a-study/question-types.md).
* **Editing a live survey has consequences.** New questions only reach people who have not passed that point yet, and deleting a question loses the answers already collected for it. See [Publish and go live](https://docs.askyazi.com/study-types/publish-and-go-live.md).
* **The conversation stays on the participant's phone.** It does not disappear after the survey ends, unless the participant deletes it.
* **Changing study type.** You can change a survey into another study type from the type chip at the top left of the study, keeping its questions and answers, but not once it has gone out to people.

## Related

* [Choose a study type](https://docs.askyazi.com/build-a-study/choose-a-study-type.md)
* [Question types](https://docs.askyazi.com/build-a-study/question-types.md)
* [Logic and routing](https://docs.askyazi.com/build-a-study/logic-and-routing.md)
* [AI interviews](https://docs.askyazi.com/study-types/ai-interviews.md)
* [Test your study](https://docs.askyazi.com/study-types/test-your-study.md)
* [Publish and go live](https://docs.askyazi.com/study-types/publish-and-go-live.md)
