Project Diaspora
TMS Ruge

(A personal note: it’s been great meeting Teddy, who’s behind this project — five days of conferences makes for good friends quickly!)

We started with the reminder that the African diaspora is a highly educated and useful group of people!

As a starting point, looking at mobile banking. Many African banks may not accept the huge volumes that the Africa diaspora transfers. Solving the remittance problem is obviously a huge ICT issue. But regulator is even a bigger problem — 54 countries! Suggestion that some banks (mostly South African) are starting to gain footprint across the continent. Perhaps a better system is partnering with an existing provider — using volume to build incentive for providers to offer lower-cost providers. Finally, donating to NGOs rather than person-to-person reduce costs.

Big problem with NGOs: people don’t know what other people are doing — get lots of duplication of effort. This is a known but unsolved problem. Should Project Diaspora attempt to solve this problem? Shouldn’t duplicate effort towards preventing the duplication of effort. Suggestion: define a microformat, to allow aggregation of databases.

This led to discussion of whether diaspora should create new projects, or support existing projects. It’s hard to set up and manage something from another country! There are also already groups that manage relations between donors and NGOs, including accountability.

Challenges to diaspora using the existing organisations: too many projects to search through; and desire to be involved especially as being a diaspora member, rather than just another citizen of your host country. What about project information from government? Problem: a lot of people distrust the government.

Project Diaspora as a social network. www.mykenyanspace.net is an existing one for the Kenyan diaspora. A further comment on these — being part of an existing organisation is better for raising funds and being sustainability. One challenge is getting recent news from people actually IN the country concerned. nabble.com: provides opportunity for volunteers in projects, and has people on the ground — citizen media type activity. Focus HAS to be on providing value to the people in the social network — some ways are letting people from the same country get in touch where they are, and also provide sources of information from their original countries.

Difference with “normal” social networking is perhaps the need for “community co-ordinators” who extract key information and memory out of the forums, and organise the database and summary pages.

A very different suggestion: work on existing open source Java applications for chatting (eg. Praekelt foundation has one) / SMS replacement on phones, and make it available through Project Diaspora — this makes international SMS essentially free! Huge draw to the site too.

South African Government Communication and Information Service (GCIS)

Two representatives from the above government office are here, and looking for feedback on how best to use mobile technology for communication from the government. Great to see people working really hard on working out the best models.

Where have we been: GCIS descends loosely from propaganda wing of previous regime, but with different ideal: make government available to everyone. Very expensive currently (hundreds of millions of Rand), questions about whether it’s a useful service. Mobile devices is a new medium, but usage testing has NOT been very successful. Now looking for suggestions from us. Platform on trial was WAP-based.

Some initial criticisms and points: I discussed the difference between push and pull mobile technology, and how push is very disruptive, that needs to be immediately relevant. Someone else discussed that mobile should be a part of a larger spectrum of channels, and so needs to be considered as successful as part of the larger project.

Question of whether government should use details requisitioned from elsewhere, or build databases from more opt-in processes.

Some suggestions: tag messages onto “please call me”s, confirmation SMSs to social grant grantees with additional messages attached.

Discussion is really interesting! I’m so involved I’m not typing. Sorry.

Lots of suggestions around service delivery issues, but of course that’s a local government issue rather than central government, so that’s a problem. But excellent idea seems to be a lookup service by SMS, where one can at least find out who the appropriate people to contact with an issue are. Call to use open standards and open source systems, and open information — get the IT sector involved in spreading the information further.

Power of Mobile Group Communication – HIV/AIDS support groups
Anna Kydd – SHM

Project called Zumbido, in Mexico: support networks for those living with HIV/AIDS sufferers anonymously but intimately and conveniently.

Project development: lots of workshops, meetings with stakeholders. Pilot ran with 40 participants for three months. Groups of 10, mostly HIV positive, also family and health workers. All messages sent to everyone in the group, to build group cohesion.

Huge message volumes! Average 1000/week/group sustained, and messages remained relevant to HIV issues. Participants’ social networks expanded, felt supported, and perspectives to HIV/AIDS and the epidemic had changed. Also debate crossing lines of class, gender, sexuality, etc. Improved mental health, self management of treatment and other life issues.

Sustainability of model: expensive for messages! Socially, though, the groups did not require moderation or supervision. Good for empowering people to use phones, and especially good for working women who may not be able to attend support groups. Interestingly, though, many groups would NOT want to continue the project indefinitely.

I really think this model would work excellently over MXit, as this solves the price problem. Luckily, apparently Cell-Life are working on exactly this! Good to see.

Getting the numbers straight: Use of mobile phones by low-income youth
Tino Kreutzer

Incredible presentation — I’ve always said that South African cellphone users are very savvy. This really confirms it, even (and especially) at the bottom end of the income spectrum.

We all know about the huge growth in cellphone usage, but not a lot of data on what people are really doing with them. Industry report data insufficient for usage info, and cellphones are shared (which muddies data); household surveys are insufficient — and seem to be wrong (asking wrong questions)! More on this later.

This research project: quantitative and qualitative, cultural probes to get the questions right. Currently towards end of data collection, covering students in bottom 50% of income in Cape Town. Pilot study: 11th graders in extreme case (bad) high school in Samora Machel township, Cape Town. 100% have used cellphones, 97% use daily, 75% own (very few own a SIM card but NOT a phone — this is different to elsewhere in Africa). Everything else (including desktop computer) have around 30% EVER USED figure. Used for: roughly equally voice, SMS, “please call me”s. (As an aside: “please call me”s were offered to get people not to do missed calls, as missed calls are heavier on the network).

Most said hadn’t used internet, but about 83% had in the previous day used a service actually on the intern – instant messaging, news, weather, downloads — so people are not identifying this as “the internet.” About half of students were doing each of taking pictures, playing games (mostly by girls!), recording videos on phones. Lots of usage of Facebook — even amongst students who have never used a computer. Mobile internet messaging: MXit 29%, noknok 17%, meep 9%, 2go 5% (note: large MTN market share here — interesting trend this).

Average expenditure on airtime a week: R30 — about HALF of all expenses (obviously these are students living at home). Fair amount of this is for social standing value — callphone use is decreasing and is seen as a little embarrassing (phoning your girlfriend from one would be considered cheap).

Challenges: not yet a shared vocabulary, tricky to get questions right — eg. people are using other people’s phones to get around having an older phone themselves.

Full results: tinokreutzer.org/mobile

In the Elevator with Operators – Pitching New Ideas to Mobile Operators
Peter Verkat- Chief Marketing Officer MTN; Vuyani Jarana, Executive Director Regional Operations Vodacom; Moderator: from Global

Some lingo:
– A.R.P.U.: Average Revenue Per User/Unit (eg. UK $40/user/month, SA $15/user/month, elsewhere Africa as low as $5/user/month)
– Churn: subscribers lost per month (lower in developed countries with contracts)
– Acquisition cost: cost per new customer
– V.A.S.: Value Added Services (core: voice, SMS, some extent data)

Started with fairly lengthy discussion on expenditure and investment made by oeprators. Started with discussion on how volume is often the most important issue. Capacity is determined by peak demand, so operators try to drive volume to non-peak times. When building base stations, take into account financial, strategic as well as social concerns.

New services. Key issues: enhance revenue; provide strategic value / brand differentiation to operator; contribution to acquisition / retention. Then go to second tier detail: interface with open / existing standards; cost of implementation and maintenance (including, eg., the training of sales people); service activation costs over many users. For example, 3G networks: will lose money for a while, but long term value and retention.

Other interesting aspect of mobile networks: huge on-the-ground presence, including in very poor areas, through dealer / airtime sales presence. When looking at marketing to all areas, need to make technology simple to present, and focus on developing small businesses around the network.

On making pitches: first tier is obviously the above issues, but after that simplicity is a key factor, especially in projected uptake.

Nothing hugely unexpected so far. Unfortunately, I had to take a call now, so I missed the discussion part.

In the Elevator with Operators – Pitching New Ideas to Mobile Operators
Pieter Verkade, Chief Marketing Officer MTN; Vuyani Jarana, Executive Director Regional Operations Vodacom; Moderator: Jesse Moore

Some lingo:
– A.R.P.U.: Average Revenue Per User/Unit (eg. UK $40/user/month, SA $15/user/month, elsewhere Africa as low as $5/user/month)
– Churn: subscribers lost per month (lower in developed countries with contracts)
– Acquisition cost: cost per new customer
– V.A.S.: Value Added Services (core: voice, SMS, some extent data)

Started with fairly lengthy discussion on expenditure and investment made by oeprators. Started with discussion on how volume is often the most important issue. Capacity is determined by peak demand, so operators try to drive volume to non-peak times. When building base stations, take into account financial, strategic as well as social concerns.

New services. Key issues: enhance revenue; provide strategic value / brand differentiation to operator; contribution to acquisition / retention. Then go to second tier detail: interface with open / existing standards; cost of implementation and maintenance (including, eg., the training of sales people); service activation costs over many users. For example, 3G networks: will lose money for a while, but long term value and retention.

Other interesting aspect of mobile networks: huge on-the-ground presence, including in very poor areas, through dealer / airtime sales presence. When looking at marketing to all areas, need to make technology simple to present, and focus on developing small businesses around the network.

On making pitches: first tier is obviously the above issues, but after that simplicity is a key factor, especially in projected uptake.

Nothing hugely unexpected so far. Unfortunately, I had to take a call now, so I missed the discussion part.

Thought experiment: what would it be like not to have a bank account and live by cash for one month. But yet, applies to maybe five billion people — and this prevents them from being economic citizens. Estimated R12 billion under mattresses at any given time, in South Africa.

First speaker Brian Richardson, CEO of Wizzit, new model bank. Looks really awesome! Use around 2000 people on the ground rather than fancy branch offices, so aimed at the unbanked. Pay-as-you-go fees, no minimum balances. Backed by International Finance Corporation (IFC).

Next speaker: Alex Comninos from ICT Africa, The EDGE institute, looking at using data from a recent survey. Titled, “From unbanked to m-banked.” Major reason people cite for not having an account: actually not expense, but, “I don’t have a regular income.” Education failure? Current m-banking is mainly supplementary to existing banking accounts, and for larger transactions.

Airtime as currency models:
1) Airtime / cash equivalence: regulation a problem, eg., have to pay VAT on airtime purchases!
2) Treat airtime as cash: requires everyone accepting airtime as a cash substitute. (Or storekeepers sell phone use to “convert” airtime to cash).
3) Airtime can be converted back to cash: good for transactions, eg., remittances.

Lots of data available — we saw some briefly! Mostly airtime is NOT a currency yet. Zero transaction costs are vital — cash is cheap for people (though not, of course, for banks).

Next: Tonny Omwansa, lecturer in Kenya and a researcher. Starts again with idea that most transactions use cash. But 93% of people in Kenya know about cash transfer services, eg., M-PESA, plus others like Western Union. People exchanging airtime between cell networks to deal with M-PESA being only for Safaricom provider. Most businesses using M-PESA also have bank accounts — it’s a more useable system! M-PESA usage stats: peaks at end of months, and at times that school fees are due.

Some applications: mukuru.com (?), mamamikes.com — buy things remotely for people.

Lastly, Jonathan Donner, Microsoft Research India. Concerned with UI issues for users, especially illiterate ones. Many differences in different markets in how people use things and regulatory frameworks, but universally we need to look at links between design, adoption, and impact. Design approach needs to take into account existing systems for transferring value, as well as socially-embed transactions (giving to family vs. lending to friends). Determines size and frequency of transactions, handset usage patterns, etc. Are we replacing a wallet or a bus driver that takes cash to our friend far away?

These usage patterns determine what metrics we should use for impact. Ended with four “types” of transaction:
– P2P transfers (remittances, transfers)
– Payments (utility bills)
– Disbursements (payroll, pensions)
– Aggregations (fundraising, shared lending)
All of these will have VERY different impacts when moved to m-banking.

Panel discussion time: Kenya is regulated such that receiving money doesn’t require a banking license, which allows things like M-PESA more easily (they also don’t offer interest). This is very rare. Regulation, however, is also necessary.

Zero transaction costs plus no interest — is it a valid model to make money only on the rolling balances held? Brian mentions how for cellphone operators and other businesses, merely the reduction in customer churn makes the rest of their business more profitable.

“Bank” may not be the best model — pawn shops and bus drivers carrying cash around for people may be better models (even if it is a bank behind it).

Brian suggested that the best model is to focus on domestic transfers first, for regulatory reasons — which is counter to usual model of international remittances. Similar problem when looking at interoperability between different systems, even domestically. Money quote: “Mustn’t underestimate the difficulty of the cash-in cash-out question.” Domestic vs. international: international regulation forces local changes in systems for compliance, but nothing is entirely local anymore. Suggestion on interoperability: look at the early history of how this develop in banking long ago.

Effects on society of adding m-banking: Look first for amplification of existing dynamics, then change. Possible changes: women have more control; negative is makes people stay on phone, and spend too much if it’s easy; psychological empowerment.

Some suggestions on what the future will hold: operators, banks will understand families more; expanding applications to support personal finance as well as banking. There are many huge organisations with vested interest, so only collaborative models will work.

m-democracy: mobile technology in citizen participation
José Carlos Vaz

We’re only starting is this sphere, so we have only questions. Questions are: how can we use this technology in budgeting and accountability (these two are different areas of focus: deliberative vs. accountability applications), and what are the barriers and opportunities. Focus on Southern countries.

Currently, effective local government processes seldom use much ICT. Some success in parts of Brazil, where budget debate included cellphones. Mexico some financial accountability stuff.

Why so little? Cultural — people like face to face meetings. Economic — expensive.

Can m-democracy make e-democracy more effective? Speaker claims mobile apps easier to use (though it depends, I suppose, on what you’re used to). Definitely, and especially in the global South, people are more familiar with phones, and are much more likely, obviously, to actually have them.

Speaker argues that local government does not get enough focus from grant-makers and policy makers, obviously especially in the mobile space. If local government improves, the rest of our actions will be far more effective.

Actions: another exhortation to make networks and connections, to push information and case studies to a broader audience; as well as donors.

Discussion: I raised the point that often successful models of accountability have been peer-to-peer messaging, without government, rather than up-and-down between government and citizens. Some examples from Philippines raised, again mostly mobilisation of people rather than up-and-down. Important issue in adoption: need to use technology to amplify existing social groups, rather than targeted at individuals.

CommCare / OpenROSA
Helping community healthcare workers in Tanzania
Gayo Mhila

Community healthcare workers: visit houses, focus on collecting health data, talking about preventative health, etc. Suffer from lack of effective tools: CommCare is a mobile support tool. OpenROSA: Consortium for mobile data collection and decision support. JavaROSA is a part thereof, a phone-based Java app (uses XForms, open standards in background). Other OpenROSA apps: EpiSurveyor, … (As a side note, there are crowds of OpenROSA people here, and everyone seems to be in on the game. Looks very effective as a collaboration).

Features:
– checklists
– simple medical protocols
– day planning
– communication

Speaker stressed, and I think this is indeed absolutely vital: it’s not creating new systems, it’s replacing existing paper-based solutions with mobile technology — thus you can actually get buy-in. Development process: Focus group -> observe workers in action -> paper prototype -> build CommCare and test.

Future plans: system with scheduling and other additional components.

Challenges: infrastructure, both network and for servicing and charging the phones; also managing airtime and workers making personal calls.

Question asked: confidentiality of HIV/AIDS, etc. sufferers. Speaker passed on what community healthcare workers said: the clinics have developed relationships with NGOs, and patients are offered the opportunity to introduce themselves to the NGOs — so it’s an opt-in service in this case.

Other discussion: education is needed for the workers, but can be done. Authoring tools for NGOs to create their own forms not yet available. Government of Tanzania interested, but it is early days yet.