Engaging the next billion consumers

Vinay Ahuja talks to Nathan Eagle, co-founder and CEO of Jana (formerly txteagle) and a leading expert on mobile research in emerging markets.

Nathan, tell us about Jana. I believe it emerged from the heart of Africa with blood written all over it?
Jana started through my work at the University of Nairobi. I started an MIT initiative called EPROM (Entrepreneurial Programming and Research on Mobiles) to teach African computer science students how to develop their own mobile apps. One of the first applications we developed was for rural hospitals in Kenya. Our SMS Bloodbank app enabled rural nurses to send us a text message with the blood supply levels in their local blood bank, and we built a visualisation that enabled people working in Kenya’s centralised blood repositories to see in real time where blood was needed. The first week this system was a big success. However, the second week the service was live, half the nurses stopped texting in the data, and by the end of the month virtually no nurses were texting data in.

Ultimately the project failed because of a lack of insight on my part. I didn’t realise that the price of an SMS represents a substantial fraction of a rural nurse’s daily wage. By asking them to send us daily text messages we were essentially asking them to take a pay cut, something that was not fair. But because my work at MIT generally involves helping mobile phone operators analyse large amounts of data, I had access to several backend billing systems from local East African mobile operators. This enabled me to develop an airtime reward platform that we used to credit these rural nurses with a small amount of air time in exchange for the day’s data in a text message. We gave them ten cents worth of airtime as compensation, enough to cover the cost of the SMS and about a penny extra, to say thank you for participating. Virtually every rural nurse re-engaged the platform, and it is now under review by the Kenyan Ministry of Health for nationwide deployment.

The takeaway here was that being able to incentivise people to provide data over their mobile phone is a really powerful way to engage with a huge number of people. We’ve now installed this airtime reward platform into the backend billing systems of 232 mobile operators across almost 100 countries. This allows us to compensate over 2.1 billion pre-paid mobile phone subscribers with airtime in denominations as low as ten cents.

That is probably half the world’s number of mobile phones in use, right?
It’s a bit under 45%.

Is this the size of your active panel?
We haven’t compensated 2.1 billion people, but we have permission to. It’s all opt-in.

Is the service now expanding beyond emerging markets?
We are really focused on emerging markets. And there’s a reason for this. There are a lot of different mechanisms to engage with and get information about individuals in places like the United States. But when you engage with “the next billion” consumers, the mobile phone is, in many instances, the only viable channel.

What kind of training do you provide to people to potentially do those tasks that one of your clients might want them to do?
The majority of our revenue currently comes from subjective surveys: “Have you used laundry detergent in the past week? If so, what brands?” “What brands come to mind when you think about the beverage industry?” You don’t need much training, you just need to be literate.

That said we’re launching a major project with the World Bank that involves some training because of the nature of the data we want to collect. We’re asking people around the world to go to their local markets and tell us the price, for example, of a kilogram of a particular type of rice. This provides the Bank with a way to calculate Purchasing Power Parity (PPP). Before we began the project, the World Bank hired professional price collectors once every four years to collect the prices of 600+ commodities around the world. We’re actively collecting the same data with our system for the Bank, and this means that we do need a bit more training of our members. And for this project we’re not using SMS at all any more, we’re using smart phones. In fact in many of these countries, Kenya included, well over 30% of the mobile phones have an, at least limited, web capability coupled with a camera. We invite members with those types of phones to go out and collect price data. Generally, the training is more along the lines of – try answering a question that we already know the answer to. If you give us the right answer you continue to the next level. If you don’t, then you’re no longer qualified but we’re still very interested in engaging you for other surveys.

That’s pretty smart. I think this could potentially turn into an opportunity for more real-time tracking as well.
Absolutely. Instead of once every four years, we’re providing the World Bank data every month about how these prices change. One of our largest clients at the moment is the United Nations, which conducts disaster surveys across a lot of different countries. Last year they did it across 33 countries using what we call the Land Rover methodology: they flew people out, rented Land Rovers and asked individuals about whether their local communities were prepared for disasters. This took half a year to do, it cost five million dollars, and ultimately they could only get 7,000 respondents. We did this study for the UN this year and we got close to 50,000 respondents. We did it in over 50 countries and got a much wider demographic because we got so many respondents. Perhaps more importantly, we have the ability to repeat our polls in these regions with high frequency to extract real-time trends.

It sounds like there is real power behind this technology to reach people who otherwise would not be willing – or perhaps able – to provide input.
Exactly. We can begin to reach out to individuals who traditionally don’t have their voices heard. Being able to start engaging with rural women, making sure that we can have their voices heard, is an important thing. Mobile phone penetration was originally dominated by males. Now, the growth is actually faster in the female demographic, across the board. It’s really exciting to enable people’s voices to be heard, especially those people who are often underserved and understudied.

What are some of the constraints in the application of mobile research?
This is not a direct replacement for web or face-to-face surveys. The mobile phone is something unique and you have to cater the research accordingly. We target to have no more than five questions in a survey, simply because there’s a real fatigue about answering long lists of questions, especially on lower-end hand sets or when people aren’t used to the idea of a survey. The questions are, generally, quite concise and multiple choice. We steer well away from things like matrix questions or long drop-down boxes.

In terms of thinking about the quality of the data that comes back to you – I believe that you have also developed an accuracy inference engine. Tell us a little bit about how that works.
The math is relatively straightforward. We rate respondents with a confidence score that continuously gets updated. If an individual gives us ten correct answers in a row for, as in the World Bank survey, the different prices in their market, we no longer need five other people pricing the eleventh product in that market as well. We can reduce the number of people who have to be involved in order to get a correct answer. So people’s confidence scores are really critical, because that’s how much we value, essentially, the data that they provide. And the more accurate and reliable they are, the more they get compensated, because it means that fewer people need to be involved in the study and it costs us less money.

ESOMAR has drafted guidelines on mobile research to ensure and promote professional standards and good practices. Are they helpful, and do you have any thoughts on how these might evolve in the future?
We really are in the extremely early stages of mobile research, especially in emerging markets. As we learn more about what works and what doesn’t work, these guidelines will inevitably have to change. Also, how we use phones for research is to some degree out of our hands. It comes down to the mobile operators themselves. Tariff plans tend to change relatively quickly, based on the operator. The whim of the operator dramatically changes what we can and can’t do from a research perspective.

Which specific areas will we need to be especially mindful about when it comes to maintaining and further developing standards around research with mobile technology?
It’s absolutely critical that individual respondents are at no risk of financial liability. Asking rural nurses to send us a text message every day with current blood-supply levels was not the right way to do research. Having people pay out of pocket to provide data gives you a massive bias towards those who can afford it, and ultimately it’s just not fair.

It’s critical that we develop methodologies that ensure respondents incur no cost for participating in the research. And while that’s an easy thing to say, in practice it’s harder to do than one may think.

Is privacy an issue?
Absolutely, and it’s critical that we maintain the standard opt-in methods for research that we use everywhere else as well. Privacy needs to be taken seriously, and it’s going to become an increasingly big deal going forward. There hasn’t been a major privacy blow up in most markets where we’ve been working, but I think it’s just a matter of time. We need to remain vigilant and make sure that the research community as a whole behaves appropriately in these new markets.

The full interview is available as an audiocast here.

Nathan Eagle, the co-founder and CEO of Jana (formerly txteagle), serves as an Adjunct Professor at Harvard University. His Ph.D. from the MIT Media Laboratory on reality mining was declared one of the ten technologies most likely to change the way we live by the MIT Tech Review. Nokia has named him as one of the world’s top mobile phone developers and he’s been elected to the TR35, the exclusive group of top innovators under 35. Nathan is also speaking at ESOMAR’s 3D Digital Dimensions event taking place from 26 to 28 October in Miami.

Vinay Ahuja is Associate Director for Consumer & Market Knowledge at Procter & Gamble with over 18 years experience. He has lived and worked in India, China, Belgium and Switzerland, and is currently located in Dubai with responsibility for Consumer & Market Knowledge for the Middle East, North Africa and Pakistan region.

ESOMAR has developed Guidelines for Mobile Research and its Online Research Guideline covers interactive mobile.