Magpi : a mobile tool for collecting data in developing countries

To know more about magpi: www.home.magpi.com

We met through Internet with Joel Selanikio, MD (Managing Director) of Magpi, who is based in Washington DC.

How did Magpi happen? (On this topic, best way to discover Magpi is to look the TED talk Joel gave in Austin: https://www.youtube.com/watch?v=Mb8x6vLcggc )

When the big data revolution happened, Joel was working on global health. In this field much of the needed data is lacking, for example on clinics in the developing word. And almost all of it is collected on paper, despite the many mobile phones available now in every country, even poor countries. As a result, for example, we do not know how many children were born in Bhutan – and without this information it is impossible to plan for how many vaccinations, or schools will be needed in the future.

Dr. Selanikio wondered about how to improve this process, and started using Palm Pilot as a data collection tool. He developed a new way of collecting data, and traveled the world to train people on how to use it. It was called EpiSurveyor. After a few years of doing that, Joel realized the trainees were using Hotmail, Linkedin or Flickr although nobody had trained them on how to use these IT services. The solution was to create a piece of software that would be user-friendly so that data collectors would not need to be expensively trained by Joel in person anymore (there is still a introduction video through: https://vimeo.com/channels/magpi).

How is Magpi used?

If WHO (World Health Organisation) wanted to know how many children had been vaccinated in a given place, the usual procedure was to send professional interviewers. These professional interviewers would type a form into word, print out several thousand copies, visit households, ask questions, and fill them manually, before entering them into Excel. With Magpi, they do basically the same thing, carrying around a phone or a tablet, and using the Magpi app on their phone. But they enter the data on their phone and through the app it gets directly uploaded to a website. This represents 75% of the usage of Magpi today: replicating the paper-based process, but more quickly and efficiently, and less expensively (often paper data collection costs three times as much as electronic data collection!).

More recently, a feature has been added, in which someone in the field sends the data to someone else (at the headquarters for instance). Instead of sending a professional interviewer, a community health worker for instance can collect data and send it to be analysed using his smartphone. This represents about 25% of the usage.

Another way of getting data through Magpi is to send text messages to all the relevant people, for instance pregnant women in an area, so that they will answer your questionnaire by SMS. Beforehand, you need to get their numbers, and create a “texting list”.

How many users does Magpi have? And who are they?

35 000 users have registered to the service and created forms. 70-80% of clients are UN agencies, governments or non-profits and about 20% are for-profits. These clients are operating in more than 170 countries around the world. Indeed, Joel underlines that he is a public health doctor, which kind of explains the strong orientation of clients towards the non-profit sector.

The goal is to be used as much as possible. Joel’s personal goal for Magpi was to allow global health organisations to collect better health information. But now, if you want to use Magpi to collect information about fashion, energy, it’s possible.

When asked how many people have collected and stored data on Magpi, Joel answered that he did not know since Magpi decided to purposely prevent themselves from seeing user data: the only thing that is shared by default is the structure of your form, so that any user can see the forms that any other user created, allowing users to build on each other’s work.

Are you a social business? “Talk the Talk, Walk the Walk”

Magpi is a business. Many people describe Magpi as a social business. For Joel, calling yourself a social business is only advertisement or PR. He wants his company to be judged on the results that it achieves.

When I suggested that being called a social business also builds trust from users, Joel said that as a technology company, clients do not need to trust Magpi, they need to try it. To him, the single most useful thing that has happened in SubSaharan Africa is the mobile phone, and that is just a business. If Magpi works, the clients would continue to use it.

What is your economic model?

The company is about 10 years old. For the first 5 years, we were supported only by philanthropy. But after 5 years, Joel thought that being funded by grants was a bad thing, because funding can prevent you from developing a product that really works.

So Magpi switched from grant funding to a commercial model, which is the freemium model, where most users use the free version. This is just like Skype, where they can have 75% of people not paying for the calls but 25% of people use Skype credit, or have a pro account and pay. The breakdown of Magpi is 99% who do not pay and 1% paying for added features.

5 years ago, Magpi had no revenue at all, and about a few hundred users. One year later, 1 000 users were on board and Magpi had some revenue, but not enough to cover our costs, 3 000 users last year and we broke even. This year we have 33 000 users, and Magpi is making a profit.

Since Joel was comparing Magpi with hotmail, I could only think of the business model of mail providers: advertising. Would Magpi consider using advertising as a new revenue source?

Obviously Magpi goal at minima is to bring enough money to cover the costs. So, one variable is revenue and the other is cost. The miracle is that with the Internet, the cost of information sharing is so low that new business models are possible. Freemium model is possible because Magpi has a really light cost-base: 12 persons are working at Magpi and they will try to keep it low. Joel says “we have engineered this to reach the poor users, the students, everyone”

Side note on the financing of ICT4D from Joel

Unfortunately, in ICT4D there is a common model for most start-ups or projects. They can get a grant, build a pilot piece of software, but when the grant stops, the software collapses, without any impact. Then, organizations need to apply for another grant.

This is not helping or developing but it is profitable. In fact, international development is organized around projects, in which programmers and consultants are charging for projects, and make some money.

What are the challenges of Magpi?

The main challenge for us is advertising. We wonder how to get people to know that we exist. We are not worried about competition since our main competition is non-consumption.

Readings suggested by Joel: the famous book by Pr Clayton Christianson The innovator’s dilemma (“The inventor of disruptive innovation”)

#interview #TIC