Telco Industry Research in ICTD: Telefónica R&D, mobiles and development
ICT4D research and private sector research in ICT4D
We have witnessed an evolution in ICT research and ICT4D research. During the 50s, there was ICT research without the ‘D’. It was mainly about governments using computers and measuring their impact. During the mid 90s, governments and NGOs began to use intensivelly ICTs to foster development; we saw the raise of telecentres, PCs and landlines and research around these topics. Lastly, in the mid 2005s, the private sector enters the arena of ICT4D research.
What are the roles of actors in ICT4D?:
- Governments: incentive packages to accelerate actors’ involvement; access to population at large and infrastructures.
- Private sector: R&D in hardware, software, services, infrastructures to innovate or adapt technology to new uses and users; access to customers base.
- NGOs & Intl. organizations, academia: access and understanding of local population and their needs.
- Intel PC Classmate: Intel developed a cheap laptop adapted to kids and which came with (not free) educational software also developed within the project. To provide connectivity, partnerships were established with public internet providers.
- Nokia Life Tools: C1 and C2 cheap cell phones with an adapted software that cna provide agriculture information, educational content, etc. Again, public-private partnerships are crucial to localize content, etc.
- Ineveneo analyses standard solutions in the market and does research on how to adapt them to developing countries.
- M-Pesa used the GSM mobile network to turn it into a mobile banking network.
Contributions so far:
- Hardware: adapted hardware lowering costs or setting up new specific features, etc.
- Software: adapting content in local language, new specific needs for specific users, etc.
- Content, services: new specific content and services that make full sense indeveloping countries, after analysing their needs, context, etc.
ICT4D research at Telefónica I+D
At Telefónica I+D, instead of developing new hardware or software, the focus is put on behaviour: as technology usage leaves a large trace of data behind, it is possible to perform quantitative analysis with the huge usage databases available. This quantitative analysis will come to complement many other qualitative researches that are often the ones taking place in developing countries.
In the case of Telefónica, 66% of their customers are mobile users in developing countries, thus their research will be a quantitative one and focused in mobile phones.
Telefónica stores data from each and every call, anonymized, encrypted and always with an opt-out option, so they can be used for research but very difficultly for other unfair purposes.
Data are mainly used for two main purposes:
- To improve the service, through usage analysis and pattern recognition.
- To provide policy recommendations, by combining data on mobile usage with micro- and macro-economic indicators.
There are, of course some limitations: the representativeness of the sample; the kind of usage (work, personal, etc.) of the mobile phone; the importance of plans or prices; the impact in data and pattern recognition of mobile phone sharing (though mobile sharing is not as usual in Latin America as it is in Africa); etc.
Gender and mobile phones project
Goal: to understand gender-related differences in mobile usage.
Data: behavioural variablesw (number of calls, duration, expenses), social variables (degree of the social network, weight of the contacts, frequency of communications), mobility variables (diameter of mobility, diameter of social network).
The characterization of the results showed that, in general, women (in comparison to men) make/get more out/in-coming calls, make the calls longer, expend more, and have a higher out-degree and weight of their social networks.
Causality tests proved to be less conclusive than characterizations, thus why there is a need to gather more data and define better algorithms.
Socio-economic indicators and cellphones
Goal: to understand the relationship between socio-economic indicators and the usage of mobile phones.
Own data are combined with national statistical institutes’ data.
It is possible, for instance, to know where a telephone is operated by asking the communications tower that handles the call. And this can be compared with geographical data that locates people and wealth indices. Now, we can test the relationships between wealth and telephone usage in a specific geographical area (location of usage is made through towers and not data from billing because only 10% of the users are on contract, being the rest of them users of pre-paid SIM cards).
Research question: does education level influence the SMS/voice/MMS behaviour? does socio-economic levels influence levels of usage or expense? etc.
By asking the communication towers we can tell where a phone call was made and, hence, how a calling person moved around while using their mobile phone (side note: 90% of people spent most of the time in just two places).
This can be combined and see where the social network of a person is located, what is the area of influence of a specific user, etc.
Research question: what is the impact of government epidemic alerts in the mobility of people? Can we trace through mobile phones whether people are more likely to stay at home if the government says that there is a high risk of contamination of H1N1 Flu?
The “areas of influence” were modelled during each of these three stages and changes of mobility patterns were looked for.
Results show that 80% of the population only reduced their mobility once in stage 2, but not during stage 1 (medical alert). This shows that the medical alert does not work, but that more interventional approaches (closing common infrastructures) does work, though it is also true that there is a side effect of increased mobility of people visiting other infrastructures (e.g. leisure ones).
Another research that related mobility patterns and urban planning showed that people spend much more time in (and move to) places that are along underground lines.
Pablo Arribas: what is the influence of education in mobile phone usage and the other way round? A: It is difficult to find causality with the data available (CDR, or call data records), so we should stay at the correlation or relationship levels.
Marije Geldof: how can you trust the data that comes from third parties (i.e. national statistics institutes)? A: Normally these are official data are validated at the international level, methodologies agreed, etc.
Ugo Vallauri: is data shared or available between companies? A: companies are on their way to share it, but not yet.
Christopher Foster: will these data be publicly available? A: protocols are being set up so that researchers can be visitors at Telefónica’s research centres.
Ismael Peña-López: if Telefónica reaches a quasi-monopolistic situation in a given country, could that influence users behaviours and thus “corrupt” the data set? A: yep, perfectly possible.
Fifth Annual ICT4D Postgraduate Symposium (2010)
If you need to cite this article in a formal way (i.e. for bibliographical purposes) I dare suggest:
Peña-López, I. (2010) “Fifth Annual ICT4D Postgraduate Symposium (IX). Vanessa Frías-Martinez: Telco Industry Research in ICTD: Telefónica R&D, mobiles and development” In ICTlogy,
#84, September 2010. Barcelona: ICTlogy.
Retrieved month dd, yyyy from http://ictlogy.net/review/?p=3486