Scale, FrontlineSMS and the local

What follows is the fourteenth in our series of FrontlineSMS guest posts. Here, Kelly Sponberg – a Project Manager at RANET – discusses the challenges of ‘local’ and ‘scale’, and the potential his organisation sees for FrontlineSMS in their work

“For about a decade now I have been fortunate enough to work on a small and niche-focused program called RANET (Radio And Internet for the Communication of Hydro-Meteorological Information for Rural Development). The program has a simply stated goal to make meteorological forecasts, warnings, and observations more readily available to rural and remote communities.  It does so through a variety of training, system development, and site deployment activities.

The technologies utilized by RANET have ranged from satellite broadcasts, to satellite telephony, to FM community radio, include HF e-mail networks, a variety of web based applications, and of course mobile phone messaging and data services. We recently began experimenting with and using FrontlineSMS to scratch a particular itch.  If you will bear with me, I’ll try to describe the challenge and problem FrontlineSMS uniquely addresses well.

RANETRANET began with the notion that rural communities are often most affected by and vulnerable to environmental changes and variability, yet the information products communities may find beneficial are not easily distributed outside of major cities in developing regions.  The quote that somewhat launched RANET came from an Algerian nomad who said when interviewed, “Just tell me where it has rained, and I’ll know what to do.”

A story about a nomad

Such a simple statement is loaded with insights and information.  Think of the challenge.  The nomad is constantly on the move in remote desert areas to shepherd his herd to food and water.  Under the best of circumstances, it is a difficult technical challenge to deliver information to this individual in a timely and sustainable manner.  Beyond the physical delivery of information, there are barriers related to language and perhaps literacy.  But moreover his statement counters assumptions about what information is valuable.  Most meteorological services try to improve forecast quality and generally the science behind weather products.  This hard work and dedication often leads to the conclusion that forecasts and newer products are the most valuable to an end user.  Indeed this is probably true for most end consumers of meteorological products and services.  In this application, however, the nomad wanted a simple observation of where it has rained as that is where there will be fresh water and new vegetation.  He cannot afford to follow a probabilistic forecast, no matter how accurate it might be.

The story of the nomad touches on the challenge of scale, which I suspect arises in all ICT4D programs.  Scale is the tension between macro and micro.  It is regional versus local.

What do I mean here?  If you abstract out why we do ICT4D projects, it comes down to solving ‘problems’ of information access and inequality, data management for efficiency, and letting individuals and communities speak in their own voice.  In the abstract, the development community, be they foreign or indigenous, wants to be able to replicate local successes across regions, countries, and continents where other people have similar needs.  Resources are simply too scarce to not strive for scalable solutions.

Scale vs. relevance

There are two major problems of scale here.  One is content, information, or the ‘byte’.  If you have a network that can distribute across a region, country, or even continent, the information distributed or shared often becomes less locally relevant and powerful the more widely distributed it is.  (The exception to this is of course sports scores.)  Certainly, technical or science based information does not change all that much.  Information on disease prevention is not going to change in substance from one locale to another, but it will necessarily need to transform how it is portrayed to fit within local cultural, religious, language, education/literacy, economic, and even political contexts.  To me the ‘what’ of information in ICT4D is perhaps the most challenging.

I work mostly on the ‘how’, which is simply the movement of information from point A to B in its most basic description.  Nonetheless, ‘how’ needs to be cognizant of the ‘what’, and as a result faces its own issues of scale.  Communication platforms that cover large geographic areas are often broadcast in nature, and therefore diminish the ability to target or carry information tailored to local needs. Of course broadcast systems are easier and more affordable to deploy than many networked systems, but with a broadcast you lose the ability to receive timely feedback or foster sharing / local production of information.   Many point-to-point or networked systems operating over large areas face regulatory challenges, are extremely expensive to operate, or require significant technical competence.  Community based systems, such as information centers and FM radio, require significant upfront investments and require considerable maintenance and training costs as well.   And these may not be connected such that local information and knowledge can benefit others.  When done right the results are clearly amazing, but the initial investment and time required to effectively establish such sites often prevents widespread deployment throughout a country; to say nothing across multiple countries.

All of this is to say its easy to find a communications solution that is sustainable and meets the needs of a small community or area.  Demonstrating success at this scale is easy.  Identifying something that suits local needs yet can be replicated elsewhere (often with an expectation of decreasing cost) is not so simple.

Image courtesy kiwanja Mobile Gallery

Enter the mobile

Mobile phone services offer a potentially interesting solution.  With rapid growth of networks in even the most remote of locations, there exists a standard and geographically expansive platform; even if operated by many different service providers.  Because of the point-to-point nature of the network, it manages to cater to local information needs and interests. In fact I would argue that in areas where the Internet has not penetrated, mobile phones change the expectation of how and what information is transmitted.  Even in comparison with a community FM station, a mobile device is inherently more local and simply intimate.  I believe this creates a new expectation for ever more tailored or individualized information.  And of course the basic economics of mobile, as well as the form factor, makes it ease to deploy.  Messages can be sent for mere cents, and as it is in the interest of commercial providers to make durable and easy to use devices, many deployment headaches are assuaged.

Clearly, mobile is not new.  There are hundreds of ICT4D projects out there utilizing mobile to collect and disseminate information.  But, there still remains a scale challenge I believe.  To process messages for collection of field data, reporting, or to distribute information beyond a small social circle or region of a country, you need some automation.  Creating these scripts and programs on a computer/server to interface with a mobile network is not always straight forward.  It requires expertise and funding to do so.  This unfortunately represents a barrier to scaling and replication.  If you examine many mobile messaging projects underway, many are either still pilots or very geographically limited.  At times it really is the content that limits scaling, but I also believe that as the technical basis for the local system is so highly tailored, it can’t be easily transferred without starting with a whole new reinvestment in setting up servers, programmer time, etc.

As a program working in multiple countries across Africa, Asia, Pacific, and recently Central America, RANET has been struggling with this issue for some time.  Do we help our country and community partners develop highly customized mobile messaging applications, but then be unable to transfer this effort to other countries?  Or, do we develop some generic data management and messaging interface that while feature reach is unwieldy or lacks the specific function needed in a local circumstance?  Frankly, we have experimented with both, and we have experience of success and failure with each approach.

FrontlineSMS and the local

In the last year I came across FrontlineSMS, which I believe represents an interesting genre of tool.  The desktop application prepackages most of the basic messaging features a small social group might want in order to exchange messages.  But the ease of use extends into more advanced applications.  The more RANET experimented with automation tasks and keywords in FrontlineSMS, the more amazed we frankly became.  Want to automatically collect field data and filter for keywords, users, etc.?  Not a problem. Want incoming messages to auto-respond?  Simple.  Have a database half way around the world you want to store incoming messages or use to feed an auto-response sent as SMS?  Easy.  Did I mention it is free?

Image courtesy Josh Nesbit, FrontlineSMS:Medic

Many reading this have probably experimented with or used FrontlineSMS, so I don’t pretend this is news.  But, I have also used other commercial applications that cost thousands of dollars and have half the features.  Some applications are feature rich and enterprise in scope, but these then require considerable technical expertise to use and maintain.  The FrontlineSMS team has done the hard job of creating an easy to use application that can be used to meet local community needs, but it also performs well for scaled applications that at the end of the day connect and replicate successful local implementations.

RANET has only begun to introduce FrontlineSMS into its country programs, but I already see the possibilities.  To help showcase some of the capabilities, as well as provide our community with baseline training, a few tutorials/discussions were added to our relatively new ‘Weaver‘ series.  The articles on FrontlineSMS are available in English, French, and soon Portuguese.  In the near future we plan to add some video tutorials and discussions, and after that hope to start sharing some of our experiences on how FrontlineSMS has been used for collection of data, ‘broadcasting’ weather information, and even allowing on-demand and automated information requests.  We are looking forward to utilizing this application, as well as learning how others might be using it in earth science and services applications”.

Kelly Sponberg
Project Manager, International Extension and Public Alert Systems (IEPAS) / RANET
Joint Office of Science Support (JOSS)
University Corporation for Atmospheric Research (UCAR)

98 thoughts on “Scale, FrontlineSMS and the local

  1. Pingback: FrontlineSMS o/
  2. Pingback: FrontlineSMS o/
  3. Pingback: FrontlineSMS o/
  4. Pingback: FrontlineSMS o/
  5. Pingback: FrontlineSMS o/
  6. Pingback: FrontlineSMS o/
  7. Pingback: FrontlineSMS o/
  8. Pingback: FrontlineSMS o/
  9. Pingback: FrontlineSMS o/
  10. Pingback: FrontlineSMS o/
  11. Pingback: FrontlineSMS o/
  12. Pingback: FrontlineSMS o/
  13. Pingback: FrontlineSMS o/
  14. Pingback: FrontlineSMS o/
  15. Pingback: FrontlineSMS o/
  16. Pingback: FrontlineSMS o/
  17. Pingback: FrontlineSMS o/
  18. Pingback: FrontlineSMS o/
  19. Sue Massey says:

    I found your blog on google and read a few of your other posts. I just added you to my Google News Reader. Keep up the good work. Look forward to reading more from you in the future.

  20. Pingback: changefeed
  21. George Reid says:

    Wonderfully written post, Kelly (thanks for sharing, Kenny). Brings to light some of the tensions between the drive (obsession?) of large scale with the human need for local relevance and ownership. Shame most of the ICTD sector seems to be headed in the opposite direction

  22. Pingback: Ivan Boothe
  23. Pingback: rootwork
  24. kiwanja says:

    @George – Totally agree. This is one of the best guest posts we’ve had thinking more about the concepts of what we’re trying to do with the software, rather than the actual uses themselves. I couldn’t have said any of this better myself, so a huge thanks to Kelly for putting this all together! 🙂

  25. Klaas van Boeckel says:

    Is the issue with scale or broadcast? I wonder. A map can have different scales each with its roles and functions. If scaling (up) information or content means broadcasting then I follow the argument made in this excellent blog. But if scaling (up) means more nodes (devices, FrontlineSMS instances, members on an online platform, etc.) in a network, then it does not quite apply as the added nodes would still be local while attached to the global. Some try to solve this by referring to scaling out rather than scaling up.

  26. Linda (@meowtree) says:

    After some excellent presentations/discussions last week at an internal ICT4D/Mobiles in Development workshop, something has been nagging at me, something I need to figure out. We heard from those with experience working at different ends of the scale — community health workers+district level hospital vs large country-wide NGO/government surveys and program evaluation data gathering — and how mobiles are being used in both instances. There’s something there about approaches that I want to think through but haven’t sorted out yet or had an opportunity to discuss (maybe I missed the discussion somewhere!?). I think this post is a piece of the puzzle, so thank you!

  27. Pingback: joshnesbit
  28. Pingback: Josh Nesbit
  29. Pingback: paconmiller
  30. kiwanja says:

    @Klaas – In the FrontlineSMS world, we talk of horizontal scaling, rather that scaling up, which does then mean you have the local ownership and control at all levels. Most people I meet in the mobile space talk of scale meaning a large, centralised system which runs everything, an approach I don’t subscribe to at all. I prefer to see a world of locally owned – and locally run – distributed systems

    @Linda – You probably know more about what we’re trying to do with FrontlineSMS than many, so I’m glad if Kelly’s post helped you think through some of the challenges a little more! 🙂

  31. Pingback: Ken Banks
  32. Pingback: kiwanja
  33. Kelly says:

    @all: Thanks for the kind comments and insightful feedback. I should rewrite this to incorporate the new thoughts. Heh.

    @Klaas: Point well taken on scaling ‘out’ / horizontally. From my experience and observation, as there is growth in the number of nodes / sites / users for a particular program, it becomes less flexible and more centralized due to donor or other management pressures. Clearly this isn’t a foregone conclusion, but I believe it’s a pretty common arc for many ICT4D projects. If I had to guess why, the heterogeneity of content becomes difficult to manage, so only if the program is focused on providing a tool or application rather than a message with content (relying on local voice and knowledge), do I think it is easy to avoid the pressures of centralization. Of course as centralization occurs over a large geographic area, the growth slows and the program becomes less relevant.

    It is funny to think that over a decade ago much of ICT4D funding was not as subject focused as it is now. I remember projects that were focused on developing the tools that could serve a number of areas such as health, education, agricultural, etc. Totally my perception here, but I think the pendulum swung to the point where only if a technical project focuses on a specific user group and topic would it be competitive for resources. Clearly there are exceptions and this is not absolute, but I have noticed the need in my own projects to emphasize a specific subject focus. I think while there is a lot of merit to this, it undermines concepts of networks and how networks grow. Perhaps the pendulum will swing back?

  34. kiwanja says:

    @Kelly – Really interested in your comment:

    “It is funny to think that over a decade ago much of ICT4D funding was not as subject focused as it is now. I remember projects that were focused on developing the tools that could serve a number of areas such as health, education, agricultural, etc”

    When FrontlineSMS was conceived, it was certainly trying to be a general-purpose tool, and not one to solve a specific problem in a specific place. I think this has been its strength. What’s been interesting from my perspective is to now watch it start to make great in-roads in certain sectors (agriculture, health and so on). The way this has played out seems to fit somewhere between the two approaches you highlight

  35. Pingback: Rohan Dixit
  36. Pingback: rd108
  37. Rohan says:

    Great article, thanks for posting. I was struck by the ‘scaling’ discussion too. I think I agree partially with Ken in that all systems need to be local and distributed- this is the most organic, most resilient to damage, and most contextually relevant network model. But perhaps there is room for some larger scale (region/country/continent) information sharing as well. For example, imagine a mobile health application which receives SMS messages about the number and type of sick patients in local villages. Most local texts are stored on the laptop/PC at the healthcare clinic running the mobile app, and these are then forwarded to the clinic’s ambulatory vehicles or medics for transport and treatment. However, all texts marked with a certain word (e.g. ‘bird flu’) could be automatically pushed to a remote server, which can be accessed globally. Now you have the ability to monitor the spread of disease at a global level, but with local “resolution”. I wonder what the problems with this model are- lack of internet connectivity probably being a huge one- but it seems that we should be finding ways to make data sharing easier. (BTW, Ken graciously pointed out that FrontlineSMS makes this very simple with the External Commands. (see

    The mix of local and global communication has really powerful effects, in fact there is even a word for this type of network (the “small-world” network, in complex systems jargon).

  38. Pingback: Heather LaGarde
  39. Pingback: heatherlagarde
  40. Pingback: NicolaAvery
  41. Pingback: Nicola Avery
  42. Rohan says:

    @Kelly Great points, all. For now, we may talk primarily about SMS services, but certainly high-speed wireless internet, coupled with cheaper “smartphone-lites”, will eventually cause another paradigm shift. I believe there are governments already rolling out these relatively cheap alternatives to broadband wire technology.

    This is exciting stuff, but of course doesn’t reflect the ground reality in many places. That’s why I think constructing modular components of a system is very important here: using FrontlineSMS to locally manage SMS, while building a “traditional” internet server back-end for the eventual migration to full-IP, real-internet capability. That way the database and services can be easily reconfigured to new on-the-ground realities, or so one would hope. Thank you again for a great article and prescient feedback– it’s much appreciated! Cheers.

Comments are closed.