Question 2 (opens 27 Nov.) What critical challenges persist in our field, and what is needed to overcome these challenges...

I'll start by saying I'm not sure what our 'field' actually is!
 
We use terms like 'ICT4Ag' or 'mAg' or 'eAg' -- but that could lead us to focus too much on the ICT rather than the domain of human activity that is most important, removing barriers to poverty alleviation, food security, employment stimulation, etc. After all, we don't focus nearly so much on 'ICT4Entertainment' or 'ICT4Sport' -- apart from device and apps industry members. 

If there was a single critical challenge I'd point to in discussing how we can use ICT most effectively as part of tackling the problem of choice, I'd point to the issue of business models. While it's always dangerous to make any sweeping generalisations, I suggest the underlying issues that affect how and under what conditions we currently try to address the problems of agriculture by using ICT tools include these four:

1. No-one wants to pay for Extension:  farmers need knowledge, farmers want knowledge -- and everyone pretty much wants someone else to pay for its creation, dissemination and basic application.

2. Islands of technology don't make a solution:  I don't think we are being 'open' enough in developing frameworks for initiatives to find common cause, useful interfaces or integration points, etc so that good ideas start to get "joined up" into something more useful to the people we want to support.  

3. Linking to national systems is critically important but seldom done:  in the parallel world of Health services, one East Africa country recently called a moratorium on any new mHealth projects (the country had over 30 of them scattered across the landscape): few if any were 'at scale', and the Ministry could not absorb the constant requests for information, collaboration, evaluation, etc.
I suggest mAgriculture is not much different.

4. If we're not (really) measuring it, we're not really managing for results: the commercial sector is becoming more and more intent on defining and agreeing some common metrics for sustainability in key value chains, enabled by convening organisations like Sustainable Food Lab (as one example - http://www.sustainablefoodlab.org/); yet a common refrain from donors, value chain partners and technical partners is that 'ICT4Ag' projects suffer from a plethora of under-defined or incompatible metrics, making comparison and covergence difficult.   

I would very much like to hear your thoughts on how we can develop a set of business model principles, standards, 'interfaces' and metrics that can allow a better ecosystem of innovations and competing ideas to flourish while actually getting on with delivering value at scale.

-- John

John,

very great post, and very well summarized. Allthe points you are making are indeed essentials.
I would like to react on the first one, about content. I started to mention it in a previous post, but I'm not entirely convinced by the need of content as such. 
In the few places I worked, I've experienced that people (agri stakeholders) are vrey ready to pay for extension services if it helps them resolving their issues. This seems obvious, but not that easy. It means that people are far more interested in getting answer to their questions, rather than getting general tips. This is very general, and it is usually relatively hard to give answer to people in a Q&A mode, but in my experience, there is a demand for such service, provided that a trust link exists between the provider of services and the farmers.
Then starting from this, I believe that building content could also be bottom-up instead of top-down: you can start without content, but experts, and based on the requests, the knowledge base grows, requirign less expert advice over time. Moreover, the knowledge is then very localized, and is also a gold source for tracking evolution based on the type of questions. It is the model that e.g. lifeline india (http://lifelines-india.net/agriculture/ ) has put up. 
I would be happy to get opinions on such approaches?

steph

Steph,

I really like your description of building content based on user demand.  You are absolutely correct that we all prefer to have answers to our specific questions rather perusing regularly scheduled tips that may or may not be relevant to us.  And I particularly enjoy your description of how user queries result in a knowledge base that requires progressively less expert input over time.  Our e-Afghan Ag website (eafghanag.ucdavis.edu) strives to be just such a knowledge bank of demand driven, user-relevant information.

Stephane and Megan and others, 

OK, we agree that demand driven content is a good thing. But where is the content coming from? And do we have examples of processes/systems that show the potential to reach a large number of people at managable cost?

There are quite a few cases of organizations (public and private) that struggle with this. What are we missing?

Hi Michael,

There are three key points in your post:
-content
-scalability 
-sustainability

About content, all my points is that, in my experience in e.g. India or Mali, there are "experts" (would it be NGOs or research institution or innovative farmers) that are able to help. It is not imho possible to extract the essence of this and "build content".  Instead, those experts are great to answer farmers questions, and thus capturing the questions and answer is how you build a knowledge (aka FAQ). SO it is not really "demand-driven content' but closer to user generated content. As an illustration, I've worked with Lifeline India I mentioned in my previous post, and in that case after 4years of operation, 85% of the requests have already an answer in the knowledge base. 

About scalability and sustainability this is clearly a major issue. See a post I wrote a 2 years ago about it (http://www.webfoundation.org/2011/10/review-of-the-new-vodafoneoxfamaccenture-report-on-mobile-for-agriculture/ ).  There are a few examples particularly in India about similar initiatives that are successfull in terms of scale, being able to reach a huge number of people. However, the financial sustainability is problematic with current set of technologies (traditional call centers) as explain in the link above. But there promising options that need to be explored further. I'm convinced that recent progress in e.g. voice technologies made such approaches promising in terms of sustainability, as soon as it will be possible to enable people to find the answer they need in the knowledge base. 
The beauty of such approahces are two folds:
1- the initial investment can be very low: you just have to setup a service and involve a group of experts that can be remunerated per request, but you don't have to make major initial investment to source content. Moreover, this is a very scalable option that could fit an ngo that is supporting a small community, and that would have to invest only a few k to setup such a system. 
2- as mentioned in few posts, you can imagine that such a Q&A service is not restricted to one topic only, but could be kind of generic one-stop-shop option where you can ask health, legal or agri questions. That are answered by different team of experts.

steph

Steph,

I really like the concept that you've presented here.  Could you expand a bit on the nature of the role of the experts in the projects in your experience?  You mention they are renumerated per request.  How are they accomodated as the knowledge base of FAQs build and their input becomes less necessary?  What is the process for accessing the knowledge base (user access, or is there an intermediary that receives and responds to queries?)? Does the expert add to the knowledge base or someone else as new queries come in?  How does the system ensure that the expert responses are of quality?  etc.

Regarding a small point in your post, I've found the COST of voice services to be a consistent barrier in using IVR rather than SMS.  Confounded by the fact that many users are challenged by SMS due to literacy limitations.  What is your experience with this?

Hi Meg,

In my experience, experts, because it is not their primary job, and it is also usually their own job to disseminate their expertise, are not really impacted by the decrease of demands. In the two cases i experienced, in fact experts were more happy to have less answers to provide.
About, accessing knowledge base: in the best world, users should be able to access directly the knowledge base. However, this is still a major technology research topic (how to guide effectively a user without ict experience to find the right information, either through text or voice, without discouraging him). In our experience, we extracted faqs (most requested information) per type of users and geographic areas and the system was proposing this upfront. For all other questions, the user ask his/her question. Then a knowledge worker asynchronously search the knowledge base and either associate the answer or send the question to expert (it does the association expert/question based on tpics/language). The expert then answers and the connection is made. When the user call back he does not know if the answer was retrieve from the knowledge base or answer by the expert, he just retrieve the answer to his question. In our case, teh expert adds directly to the knowledge base, and there is no quality check over expert input.
The expert answers using his prefered communicaition channel: it could be email, web, ivr or sms (even if sms is usually inadequate to capture the answer). He receives an alert when a new question has come for him (through again the channel he prefers).
(you might be interested in a piece of research on this: http://public.webfoundation.org/2012/02/VBAT_Pilot_Report_Public.docx )

Now about voice, here again you may be interested in a more deep answer at http://stephb.org/2013/03/demystifying-voice-technologies-for-development/ 
To be honnest, in my experience in about 8 countries, the cost of IVR has never been a challenge.
Well this needs to refined. What is essential is the return on investment. If the service behind has a direct value for the user, he will place the call. 2 very specific examples: the helpline i was mentioning, where people  were paying for the call plus a service fee. A second example is a business matching service  same model (phone cost+service fees), and no problem, if there is a direct revenue. in the case of business matching, we realized that our model (collecting offer and disseminating them by radio) was very successfull for some products and very unsuccessful for some others because of different challenges in the value chain. so obviously you see those who see a direct benefit calling,a nd the other not calling.
In another domain, i worked with radio stations setting up an ivr system for personnal communique. Broadcasting personal messages is a huge business for community radio, that can go up ot 20 or 30% of their revenue. we worked with 5 radio in mali. 4 failures 1 success. here again it was only the business model: people could call and let a message, but in hte case of 4 failures, the radio will not broadcast them, because they need to be paid for that, and just doing the travel to pay, was not worth it for most people, and for those close to the stations it was cheaper to go to the station. So nobody was calling and the cost of call was mentioned, while it was the service behind not carefully designed. In the case of the success, the radio was already structured with 50 local correspondents in different villages. So people were visiting the correspondent paying them the cost of the broadcast and then they were calling the system to drop the message. Here total success: not only people can have access to a service they didn't have before (or it was far less easier) but there was a great side effect: now it is their own voice that is broadcasted, and being heard on the radio was a great thing. In such case again, the perceived gain compared to the costs of the call was perceived as a great deal by users.
In very few words, i've experienced lots of success and at least the same amount of failure with IVR, and the cost is sometimes identified as the failure factor, while it is not the case, and it is the overall busienss model or the value of the service that is problematic.
(but there are also lots of other possible factors for IVR to fail miserably, see the post I mentioned above)

Steph

John, you've summarized this topic beautifully.  Below some comments that I hope will spark continued discussion on how we could develop business models that support the scenario you describe.

Regarding metrics: ICT devices are very well suited to collecting user data.  Perhaps the users themselves could be engaged in measuring results.  One possibility could be giving permission for the device to collect information regarding how the content is accessed (most visited topics, which functions employed, etc etc).  Another possibility would be to provide user with opportunity to evaluate the intervention and offer feedback thru the device.  Does anyone have other thoughts on how the device itself could help collect meaningful data?

Regarding islands of technology, this ties back into some of our discussion on Question 1 and seems more than anything to be rooted in the approach and "open-mindedness" of the implementors.  As I suggested in Question 1 discussion, the best solutions are repurposing what is already there.  Does anyone have any examples of a project that "connected" underserved stakeholders to existing services rather than attempting to create an isolated island of service?

Regarding extension: Answering this question may well revolutionize farming in developed and developing nations alike.  How can creation and dissemination of extension knowledge be made marketable?  In developed nations it seems the best response has been that it is not marketable, but it is necessary and it is needed by everyone.  Hence, taxpayers and national foundations support such endeavors.  So, what is the alternative in nations where governmental funding does not adequately support a functioning extension network?  Can this issue be efficiently resolved non-governmentally?

Regarding national systems: oftentimes critical mass is what makes a project become part of the "national system" ex: Facebook, Twitter.  So what's the difference between a project that "goes viral" and one that never catches on?  Metrics may help answer this, but one suggestion I have is that the most successful projects enable multi-way discussion among peers and social groups that would otherwise be unable to connect ( thus leading to knowledge sharing) rather than simply spitting out prescribed bits of information to an isolated user.  Think Wikipedia versus Websters.  As Steph explains, the presence of "experts" and a variety of other types of stakeholders increases the potential quantity and quality of information to be shared.  But the point is--an open, undefined communication channel often holds the greatest potential for impactful information exchange.  Does anyone have other concrete examples along this line?

Hi Megan, all

I just wanted to react on one small element of your post.
you wrote:
"Does anyone have any examples of a project that "connected" underserved stakeholders to existing services rather than attempting to create an isolated island of service?"

From my experience, this is touching a very sensitive point related to trust. There is a big difference between information and knowledge. I mean here that you can send informatin to people, but then it is up to them to decide whether this information is knowledge that helps them or not. In my experience, this is where trust is essential. Farmers would e.g. change their palnting technics only if they trust the person telling them that this will be effective, and also if they can witness it. Just to say that just focusing on conencting people to existing services is not necessarily the way to success. I've experienced this on a particular project related to business matching (people posting sell/buy request): setting up a business matching service is technically doable, but i realized the hard way that it is not only a question of connecting buyers and sellers, it is about ensuring that both buyers and sellers are trusting the system in terms quality, reliability of information etc. This aspect was the most important part, and it is very similar to the ype of relationship farmers have with their cooperatives.
So from my perspective the key question is how to have global services but keep/establish a trust link?

steph 

Steph, you hit an incredibly important point here, and I believe another of the major obstacles that remains to be overcome in our field.  Your comment got me thinking about a) what makes me as a consumer trust a service provider and b) how those characteristics can be transferred into a setting where transparency, customer service, TRUST, etc.  have not traditionally been part of the business relationship model.  

Consumers generally trust systems that have a series of checks that consistently result in quality.  The nature of those "checks" will depend entirely on the nature of the project itself, but in terms of using ICT, one of the inherent qualities of devices is that they are ruthlessly objective.  This can be both an advantage and a disadvantage in terms of trust depending on the nature of the service offered, and therefore something the implementor should remain very concious of in designing the system.

On the "advantage" end of that, a large part of my own trust--and I presume others from similar business cultures--in the service provider comes from the fact that so much of the exchange of services occurs via computers.  Thus, even if there is a "rotten egg" employing the system, the computer system disallows any nonstandard practice.  Such device-controlled interactions can help build the trust of the consumer. E.g. mobile banking.  

On the other end, especially in more traditional cultures, face-to-face contact and "seeing is believing", as you mention, are absolutely key to adoption of new practices.  Thus in these cases the device as representative of the service is actually an obstacle to building trust in the system.  In such instances I would suggest that the ICT service is better targeted at an "expert" or "representative" in the community  (such as a gov't ag agent), rather than the end benefactor (such as a farmer).  E.g. ag extension services.