Collecting unstructured audio diaries on WhatsApp is a longitudinal qualitative method: participants record open-ended voice notes inside WhatsApp, usually on a schedule or after a trigger event, without following a rigid question list. Researchers download, transcribe, code, and analyse those entries for themes, sentiment, and change over time. WhatsApp is the capture channel because it removes the friction of downloading a separate app or clicking an email link.

Quick answer: collect unstructured audio diaries on WhatsApp by sending participants an open-ended prompt (kickoff, event-triggered, or reflective), asking for a 60 to 180 second voice note in reply, then automatically downloading, transcribing, and coding those OGG/Opus files for themes over the course of the study. The three setup paths are the free Business App for small pilots, the Cloud API for custom builds, or a purpose-built research platform for speed.

What "unstructured" actually means

The "unstructured" part is what separates this from a survey. Instead of choosing from a list or answering narrow questions, participants speak freely. Academic work on audio diary methods consistently finds this approach captures richer emotional data, reflective pauses, and contextual detail that structured instruments tend to miss. The method sits between qualitative research and mobile ethnography, and works best when the goal is understanding lived experience over time rather than measuring a single moment.

When to use unstructured audio diaries

Audio diaries earn their place when you need in-the-moment emotional context, when a behaviour is hard to recall accurately after the fact, when participants are more comfortable speaking than typing, or when you want longitudinal texture across days or weeks rather than a single snapshot.

They're a weaker fit when tight comparability across a large sample matters more than depth, when heavy probing is essential (you can't interrupt a voice note to follow up), or when participants are in environments where speaking aloud isn't practical. If a study needs both the depth of unstructured voice capture and the ability to probe adaptively, pairing diary studies with AI-moderated interviews on WhatsApp can cover that gap, and combining the two methods with a survey layer is covered in our guide to running surveys, diaries, and AI interviews in one study.

Setup: three technical paths

Three options exist, and the right one depends on scale and where the team wants to spend its time.

Prompt patterns that produce useful diary data

Good prompts are the backbone of useful diary data. "Unstructured" doesn't mean "no guidance." The participant drives the content, but a clear starting point and a sense of what's useful still matter.

• 01
  Kickoff prompt. "In a quick voice note, about 1 to 2 minutes, tell us what stood out today about [topic]. Anything surprising, frustrating, or delightful works, there's no wrong answer."
• 02
  Event-triggered prompt. "Right after you [event], record what happened and how you felt. No need to be formal, just talk like you're telling a friend."
• 03
  Reflection prompt. "In 1 to 2 minutes, describe one moment today that changed your mind about [X]. What led up to it?"
• 04
  Mid-study check-in. "We're halfway through. Record a voice note about what's become easier or harder about [experience] compared to the start."
• 05
  Closing prompt. "Looking back on this week, talk through what you'd keep, change, or stop about [experience]. Anything goes."

Set expectations early: aim for 1 to 2 minutes per voice note, shorter is fine if the point has been made, and try not to go past 3 minutes. Without this guidance, some participants send 8-minute monologues that become unreviewable, while others send 5-second clips with nothing usable. The 60 to 180 second range balances richness against analyst feasibility; roughly a minute of unstructured speech produces about 130 to 170 words once transcribed.

Media, transcription, and language realities

What file types you'll actually receive

When participants send voice notes through WhatsApp, the Cloud API delivers them as OGG/Opus files, with webhook payloads showing the MIME type audio/ogg; codecs=opus. The pipeline needs to handle this format for download, storage, and transcription. Standard audio messages carry a 16 MB size limit, rarely an issue for voice notes in the 1 to 3 minute range.

A common API pitfall

A frustrating error shows up often: "this audio is no longer available." It tends to appear when businesses try to send audio back to participants, or when media URLs expire before download. The fix is ensuring audio files are properly encoded as OGG/Opus with correct MIME headers, built into the pipeline as standard practice rather than a one-off patch.

Built-in transcription is not enough

WhatsApp rolled out on-device voice-note transcription in late 2024. iOS supports a broad set of languages; Android initially limited support to English, Portuguese, Spanish, and Russian. For research purposes, this isn't sufficient: you need transcripts that can be exported, searched, coded, and analysed with consistent quality across every participant, not variable on-device processing. Working across African markets makes this sharper still, since you'll likely hit languages and dialects where built-in transcription simply doesn't work.

Low-resource language challenges

Automatic speech recognition has improved dramatically for major world languages but remains uneven for many African languages, and code-switching (mixing languages within a single voice note) compounds the problem. The practical response: use the best available ASR for bulk transcription, then spend human QA time on key segments. Tag code-switching explicitly in the codebook, and keep original audio files so analysts can listen back when a transcript looks off. This challenge runs through most multilingual fieldwork; our complete guide to multilingual qualitative research covers translation and transcription workflows in more depth.

Ethics, privacy, and consent

Voice notes are personal data, and unstructured audio diaries are especially sensitive because participants speak freely, which means they might disclose health conditions, political views, sexual orientation, financial difficulties, or other topics that qualify as special category data.

GDPR (EU and EEA)

Under GDPR Article 9, processing special category data requires explicit consent or a suitable research basis with appropriate safeguards under Article 9(2) and Article 89. The consent flow needs to clearly explain what data is being collected, why, where it will be stored, who can access it, how long it will be kept, and how participants can withdraw.

POPIA (South Africa)

South Africa's POPIA treats "special personal information" similarly. Processing typically requires consent or a specific public interest or research basis, with safeguards proportionate to how sensitive the data is. For the retention side of this, see our guide to secure retention and deletion policies for participant data.

• 01
  Plain-language opt-in at enrolment, with separate confirmation for voice and video capture.
• 02
  Specific data residency disclosure: where the audio is stored, who can access it, and for how long.
• 03
  Withdrawal mechanics participants can use at any point during the study, not just at sign-up.
• 04
  Special-category data warning if prompts could elicit health, political, or otherwise sensitive content.
• 05
  Retention schedule defining when raw audio, transcripts, and exports will be deleted or anonymised.

Common pitfalls and field fixes

• 01
  Overlong voice notes. Without guidance, some participants treat every prompt as a 10-minute session. Set expectations at onboarding and send mid-study reminders; always provide transcripts to analysts, and where useful, back to participants for member-checking.
• 02
  Drop-off after day three. Diary fatigue is real. Respond to at least some entries with a brief acknowledgment. Scheduled template messages outside the 24-hour window keep the cadence going, but template quality ratings matter: if participants report messages as spam, Meta throttles sending capacity.
• 03
  Language and ASR failures. Across multiple African markets, expect ASR to stumble on certain languages and dialects. Plan human review for critical excerpts and budget the time; it's a requirement for trustworthy analysis, not a nice-to-have.
• 04
  API and media errors. Media URLs from the Cloud API expire, so a system that doesn't download voice notes promptly will lose them. Build automated download into the webhook handler and normalise audio to a consistent format for downstream processing.
• 05
  Single-device bottleneck. A study that started on the Business App and grew ends up stuck with one phone, one researcher, and no API access. The fix is migrating to the Cloud API or a purpose-built platform, a common story.

What "good" looks like

Frequently asked questions