Traditional methods often struggle in emerging markets, where internet access can be inconsistent and mobile data is genuinely expensive relative to income. WhatsApp is optimised for exactly this environment. Here's how to design, launch, and analyse a study that keeps costs low without sacrificing data quality.

Why WhatsApp is a game changer for affordable research

A survey run over WhatsApp is a genuinely different cost structure from the alternatives. SMS carries character limits and a per-message fee; phone calls need real time and personnel. WhatsApp messages use a small amount of internet data instead, and a message sent through the WhatsApp Business API can cost under two cents. There's no charge for the recipient to receive a message either, removing a real barrier to participation in markets where people are careful with their data plans.

Designing your WhatsApp survey for maximum engagement

A great WhatsApp survey feels less like a questionnaire and more like a conversation, which is the whole reason it works. People are already comfortable on the app: recent estimates put the average user opening WhatsApp somewhere between 20 and 28 times a day. Survey design should tap into that natural, chat-based behaviour rather than fight it.

Keep it conversational and multimedia-rich

Send questions one at a time to create a natural flow, in simple, clear language, rather than one long list. Yazi's survey question bank can help phrase WhatsApp-friendly prompts that reduce confusion and drop-off. The real power comes from going beyond text: multimodal questions with images or short audio clips aid comprehension, and letting participants reply with a voice note or a photo of how they actually use a product adds a layer of qualitative depth text alone can't capture.

Setting up your study for success

Pilot testing and iteration

Always start with a pilot. Use a sample size calculator to scope the main study before launching at scale, and run a small trial with 5 to 20 people from the target audience first, enough to reveal confusing questions, technical glitches, or flow problems. BFA Global ran exactly this kind of pilot for a WhatsApp survey in Ghana, testing question clarity and even the best time of day to send messages; a later field study saw a 72% completion rate among people who started the survey, a strong result that reflects the value of iterating before scaling.

Finding participants and getting consent

Surveying people on WhatsApp requires their phone number and explicit opt-in; unsolicited messages aren't an option. Recruit through existing customer lists with consent, social media campaigns, community partnerships, or a vetted panel provider. Once you have a contact, the first message matters most: a clear, polite introduction stating who you are, the survey topic, how long it takes, any incentive, and a direct consent request ("reply YES to start"). This isn't just good manners, it's required under privacy regulations like GDPR and POPIA.

Building credibility and trust

People are rightly cautious about messages from unknown numbers. Establishing credibility immediately, a verified business profile, a clear explanation of how you got their number, and a consistent, professional tone, makes the difference between a completed survey and a message left on read.

Reaching your audience and maximising responses

Sending strategies: broadcast vs. API

For smaller studies on the standard WhatsApp Business app, a broadcast list sends a message to up to 256 contacts at once, each received as a private, one-to-one message, though recipients must have your number saved in their contacts to receive it. For larger studies, the WhatsApp Business API through an automation platform removes that contact cap entirely, provided messages use pre-approved templates and only reach people who've opted in.

Timing and reminders

When you send a survey matters. Avoid busy morning hours and late nights, and let pilot testing reveal the optimal time for a specific audience. Not everyone answers the first message, a polite reminder roughly 24 hours later can meaningfully lift response rates by catching people who were simply busy. WhatsApp's delivery and read receipts, the grey and blue checkmarks, give real-time feedback on whether a message was seen, which makes deciding when to send that reminder far less of a guessing game.

Incentives, data, and best practices

Compensating for data and time

Even though WhatsApp itself is data-efficient, a small incentive like airtime credit is good practice. It compensates participants for the data they used and thanks them for their time, a gesture that can meaningfully improve willingness to participate in markets where people track mobile data spend closely.

Privacy and data security

WhatsApp provides end-to-end encryption in transit, but responsibility for the data doesn't stop there, you're accountable for how it's stored and used once received. Platforms built for research, like Yazi, offer EU or South African data residency options to help meet these compliance requirements; see Yazi's Data Security Executive Summary for details.

Understanding sampling bias

No method is perfect. WhatsApp research is limited to smartphone users with internet access, which can skew a sample younger and more urban. Quota sampling or weighting against known population demographics helps correct for this, and integrating a WhatsApp survey with phone or SMS outreach for people not on the platform creates a more inclusive, multi-modal design overall.

Managing and analysing your data

A well-designed workflow automates as much of this as possible, ideally including AI-moderated interviews for qualitative depth alongside the quantitative survey. Automation turns messy, multi-format chat data into structured, actionable insight and dramatically speeds up the analysis phase compared with manually reconciling text, voice notes, and images by hand.

Frequently asked questions