Txteagle in Flight: Mobile Data Collection for Disaster Preparedness

Posted by MelissaUlbricht on Feb 07, 2011.

Txteagle is a data collection and engagement platform that leverages mobile airtime compensation for data collection and customer engagement. It is currently used by one nonprofit organization to survey constituents about disaster preparedness. We looked into the company and one of its customers. 

Txteagle widely advertises its ability to reach 2.1 billion mobile subscribers currently. MobileActive.org spoke with Nathan Eagle, co-founder of the service, to learn more about how it works, how many active participants there are and where some of the large numbers come from. We also spoke with Terry Gibson, project manager for the Global Network of Civil Society Organisations for Disaster Reduction, who is currently using txteagle to collect mobile data on 40,000 respondents in 48 countries.

When Engagement Drops Off, Incentivize!

While working as a Fulbright professor at the University of Nairobi a few years ago, Nathan Eagle received funding from Nokia to set up EPROM, a set of mobile application development labs across sub-Saharan Africa. One of the first developments was an application that enabled rural nurses in Kenya to text in blood supply levels at local hospitals. The app included an interface so that central blood repositories could also see the blood supply levels in real time. The first week was a big success for the app, Eagle said. But by the second week, about half of the nurses stopped texting. By the end of the month, virtually no data was coming through the system, and it was deemed a failure.

At that time, Eagle was working with Safaricom, a mobile operator in Kenya, where he had access to the back-end billing system. He was able to send small denominations of airtime through the billing system to the rural nurses in exchange for a properly formatted text message with current blood levels. “For the opportunity to earn a penny of airtime, suddenly all the nurses wanted to re-engage with the platform and send in data,” Eagle said.

This led Eagle to start thinking about how monetary incentives can be used in other mobile data collection endeavors: to provide data about local market prices, to send in blood and hospital supply levels, or to conduct surveys for market research.

 Txteagle Grows Up

The compensation platform that Eagle hacked together in Kenya for Safaricom was eventually integrated with the billing systems of the company’s partners: 220 mobile operators around the world.

This is where the rather impressive number of 2.1 billion consumers comes from: The company has access to 2.1 billion active numbers across the 220 operators in 80 countries. Eagle said that nothing outside of country and operator is initially known merely from the phone number; it is only through an opt-in process that more information is learned, in exchange for airtime.

Within the txteagle platform, there are clients (organizations or companies use the platform to collect information) and members (individuals use their mobile phones to complete surveys and receive airtime in return for filling out surveys). A typical txteagle client has a developed survey in hand, and is often conducting expensive, face-to-face survey or interviews, Eagle said. The individual survey respondents are the txteagle “members,” who have primarily been learning of the service through word of mouth.

Txteagle taps its subscriber base to reach individuals in a particular area via SMS or USSD messaging, and members who choose to complete the survey receive airtime compensation in varying amounts. The current number of members who have taken surveys and received compensation through txteagle is in the “hundreds of thousands,” and the company has a goal to reach 1 billion people in five years, Eagle said, though he refused to disclose more detailed number.

 Collecting Information on Disaster Preparedness

The Global Network of Civil Society Organisations for Disaster Reduction is a recent txteagle client. The network consists of about 300 not-for-profit organizations around the world, and is concerned with reducing the impact of disasters and increasing people’s resilience to disasters.

The Network formed four years ago in response to the United Nations International Strategy for Disaster Reduction  Terry Gibson, project manager for the network, said that it was created “to help mobilize local people to press for more effective translation of the high-level policy into practical action.”

As part of a community-level survey called Views from the Frontline, the network sends members into the field to survey people on specific measures of progress on the UN disaster reduction program.  In the past, such measures were done by way of face-to-face, paper surveys with vulnerable communities around the world. The surveys ask two main questions:

  1. How vulnerable do you think your community is to disasters?
  2. Over the last five years, have a seen a decrease, no change, or an increase in disaster losses?

The data is then taken and shared at UN assessment meetings; current data collection is being compiled ahead of a UN meeting in May 2011.

This is where mobile data collection, and services liketxteagle, enter the picture: The Network is aiming for 14,000 face-to-face responses via paper surveys in 70 countries, and now, it is also aiming for 40,000 to 50,000 responses via SMS in 48 countries. The Network has completed a pilot phase to collect 10,000 SMS responses in 7 countries, and is now setting up to do the full survey.

For Gibson and his team, the question is how to increase reach and connect with a broader range of people. To travel and physically survey individuals is a very labor intensive process. The network has been interested in using SMS for some time, especially because web access is still frustrating in a lot of areas, Gibson said.

 How the Global Network for Disaster Preparedness Collects Data

Gibson describes the txteagle survey as a two-part process: recruitment and compensation. An initial invitation is sent out as a “blanket SMS” and assumes a 10 percent hit rate. So, in order for Gibson and the network to get 500 responses in a territory, 5000 invitations are randomly sent. The invitation asks people if they would like to participate in a survey connected with the UN about disaster reduction and explains that those who complete the survey will be compensated in airtime. It asks people to respond by texting yes or no.

In this case, the survey taker — the txteagle member — can elect to respond to the survey via SMS/USSD or online. Even is he/she chooses to respond via the web, the individual still receives airtime compensation.

Gibson said this particular survey is lengthy: it consists of 29 questions, 10 of which are demographic questions and 19 of which are more detailed indicators of progress in disaster reduction. The mobile data is then returned in a spreadsheet to Gibson and his team.

A face to face survey costs the network $125 USD per survey, which includes the cost of training, consultations, discussions, and other activities. An SMS/Web survey, on the other hand, costs the network $5 USD per survey through txteagle. “So there’s a clear cost benefit if one focuses purely on the data,” Gibson said. Nevertheless, the Global Network will continue face-to-face surveys as they are part of a social process which helps to train, inform, and mobilise people.

For each SMS survey, the $5 cost is paid to txteagle, and “from there it breaks down into SMS costs, payments to respondents, payments for access to network provider, and texteagle costs,” Gibson said. For the network, there is no other direct expenditure on the SMS survey, and all recruitment and compensation costs are contained within the $5 per survey cost. Eagle stressed that “there is no going rate,” and the per survey costs will ultimately depend on the region and the communication channel (USSD, Web, SMS, etc.)

 Incentives for Operators

Establishing partnerships with multiple mobile operators can be tricky. But with txteagle’s 220 partners, there is an incentive for the operators to get on board: a new revenue stream.

Average revenue per user is plummeting in emerging worlds, Eagle said. “The people who are buying phones in Kenya this year are making less than the people who were buying phones in Kenya last year, and that’s true in every emerging market.”

Txteagle provides the operators with a unique revenue stream. “Instead of focusing exclusively on extracting revenue from an individual making $3 a day, the operator can receive revenue from the companies and agencies — the clients — behind the mobile surveys.”

 The Technology Will Improve

But, working with the operators has also been a challenge. Networks can be unreliable, affecting compensation and survey delivery. “This means [members] don’t receive the survey, and therefore don’t earn any compensation,” Eagle sad. Also, exchange rates are in flux, changing the price of airtime denominations. As Eagle notes, “We’ve got something that is working right now, but there is always room for improvement,”

As for Gibson and his experience with txteagle, he is somewhat removed from the technical forefront and delivery — his team establishes the parameters and then receives data in return — but he has not heard of any disruptions or problems from surveyors.

 Lessons

In using txteagle, Gibson and his team not only try to increase the reach of the survey, but are also able to compare survey methods.

“We’re doing it on the basis that it’s partly increasing our reach. We’re doing it in countries where we don’t have our own network. But we’re also doing it in some countries we are working in already, so we can look at what happens when you ask people questions through two different methods,” Gibson said.

One interesting lesson from the pilot is a person’s chosen platform. Gibson said that 95 percent of respondents have elected to respond to the survey on the web, after being recruited via SMS. “They have been able to access a computer and the web much more successfully than we expected,” he said.  This may also increase response bias, however, indicating possibly that a more connected, wealthier and more educated cohort answers complez surveys.

What the network has also found is that the actual process of conducting face-to-face surveys in a community starts conversation and dialogue — “spin-offs” — that become a valuable part of the work. “There is a different question about the SMS approach and where you can take it next in terms of development,” Gibson said.

But what mobile data collection does do is change the landscape for information gathering, Gibson said. Responses now come directly from the people in the communities in the developing world. There is no middle man. “That very rarely happens. Nearly all the information about what people supposedly think, in fact is filtered through the people who can get that information,” Gibson said.

“Not only for our network, but for a lot of other projects that are concerned with humanitarian and development work, the opportunity to get the completely clean version of what people think about things is something than can be explored much more widely,” Gibson said.

Photos courtesy of Terry Gibson and the Global Network of Civil Society Organisations for Disaster Reduction. He can also be reached at terry.gibson@globalnetwork-dr.org.

Advertisements

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s