Learnings from running WhatsApp surveys

We have been running surveys on WhatsApp for a while now - about 50+ WhatsApp surveys and more than 10,000 responses to date. These surveys have spanned different categories, questionnaire lengths, audiences, and languages. However, one can clearly see patterns emerge around what the successful ones (i.e., those with high response rates) did better than the rest:?

1. Successful surveys felt like a conversation

The key strength of messaging apps like WhatsApp is their conversational nature. These conversations tend to have a more casual and friendly tone than that of a typical business survey. Keeping this tonality goes a long way in getting better responses.

How do you make a survey conversational? Consider the following two versions of a segment from an ad test survey:

Version 1:

Q. How did you like this ad?

A. I liked it a lot

Q. Why?

A. It was funny

Q. How would you rate this ad? Please rate it on a scale of? 1 to 5, where 1 stands for very bad and 5 stands for excellent?

A. 4

Version 2:

Q. Thank you, Nitin, for watching the ad. How did you like it?

A. I liked it a lot

Q. Good to know that. Why did you like it?

A. I found some parts of it very funny and relatable. Especially the part where the boss walks in. Happened to me once.

Q. How would you rate this ad?? Like you see the movie rating with 1 to 5 stars- 1 being very bad and 5 being excellent

A. 4

Version 2 is clearly more conversational. It addresses the respondent as a person and by their name - an undervalued hack! It validates their response, thereby making them comfortable with sharing their opinion and refers to something the respondent is very familiar with - the movie rating process. It is hardly a surprise that we saw a 15% improvement in response rate with version 2.?

2. Successful surveys localized and improved question comprehension

A big advantage of messaging apps is that all of them come with built-in language support. WhatsApp, for example, supports over 60 languages. It is a no-brainer then to take advantage of this feature and localize the surveys in the language of the respondent. Not only does it help get better completion rates, but it also increases comprehension of the questions. Thus, getting better quality responses. In fact, when we conducted a WhatsApp survey in a rural part of India for a major consumer product company, we added audio support along with each question, i.e., if the respondent could not understand the question by reading, they could hear the audio version of it.

3. Successful surveys avoided long questionnaires

When you talk to your customer once a year, the temptation is to ask everything one can think of (aka FOMO). The result is 40-50 minute long surveys, tired respondents, and questionable data. Not just WhatsApp, such surveys are not advisable for any medium- online or offline.

WhatsApp is not built for such long surveys. The strength of WhatsApp surveys is that you don’t need one-time long surveys. It allows you to keep doing several, short and focused surveys throughout the year. Shorter surveys have higher completion rates and improve data quality. Additionally, it avoids the blind spots that infrequent customer contact creates.

4.?Successful surveys avoided long questions

We live in an era of short attention spans. Twitter allows 140/ 280 characters. 140 characters roughly translate to 15-20 words. And yet users barely read an entire tweet. The average text length on WhatsApp is about 5-6 words. Granted, this average is pulled down by emojis, one-word responses, etc. but it does provide some guidance on how long the questions should be.

5. Successful surveys gave updates on completion progress

Unless your survey is about a subject that is a high involvement one, most respondents start to wonder where the conversation is going and how long it would take. It helps a great deal to give out this information clearly and frequently. Additionally, when the survey creator gives out information to the respondent, the conversation feels more equitable where both sides exchange information.

6.?Successful surveys ?assured privacy

There is an increased awareness and concern around privacy. With customers being constantly nagged by communication sent by marketers, not all of which is ethical, there is skepticism around sharing information. Part of the job of the survey creator is to assure the respondent that the data they share would be used ethically and for the purpose that it has been collected. If you are collecting personally identifiable information, make sure you take their consent on how this data can be used

7. Successful surveys offered relevant incentives

Perhaps this is stating the obvious. In most cases, anyone who participates in a survey does so because they get some form of payback. Sure, some might do it for altruistic reasons but that is rare. It helps to state the incentive clearly and upfront. Incentives work only if they are relevant, accessible, and valuable to the respondent. Starbuck coupons for rural research in India is not an incentive as it would be irrelevant and inaccessible to the audience. A Rs. 50 voucher is not an incentive if you are researching high-end cars- since it is not valuable enough for the type of audience you might be targeting. Also, if you care about your brand (if not karma), pay up! Do not try and short-change the survey respondent by making it hard to redeem their incentive.

Of the seven points mentioned above, the first six are about reducing the cost of responding to a survey. The cost can be in the form of time and effort required or in the form of information given out, especially personal information. The last point, incentives, is about the upside of incurring this cost. Unless the survey promises to be net-positive for the respondent, the response rates would be low. In a nutshell, assure privacy to overcome the resistance to sharing information, incentivize to get respondents started, and reduce the effort required to complete the survey.?

?