Rain-fed agriculture provides the bulk of humankind’s food supply. Its potential to increase productivity is greater than that of industrial-scale farms. For this purpose, however, smallholder farmers must change the way they manage land, water and crops.
Most rain-fed farms are operated by marginal farmers who barely manage to make a living. They lack information and are sceptical about new approaches. Agricultural extension efforts hardly reach them, so they do not get expert advice. Often they do not even get relevant information from peers. High illiteracy rates further hamper their ability to learn about how they might boost the productivity of their farms.
Much of the information that can help farmers is specific to where they live ("does mango variety X grow as well in our region as it does 500 kilometres farther north?"). Such contextual information is not readily available from scholarly sources. It matters, moreover, that smallholder farmers tend to shy away from risks. They tend to prefer minimising losses to maximising gains. Most rely on traditions, trust peers and are sceptical of external experts.
Thanks to the proliferation of mobile phones and sinking prices for telecom services, information and communication technology (ITC) is beginning to make a difference in the rural areas of developing countries, however. Increasingly, even illiterate people use ITC applications. It would make sense to use such opportunities for spreading knowledge about best agricultural practices.
To find out to what extent digitised social networking can serve this purpose, we designed and implemented "Coli". It is a multi-modal social network that is based on smart phones and deals with farm issues in the Indian state of Karnataka.
For three years, we conducted a small field trial in a village for ten months. Fifteen farmers took part. Seven of them were basically illiterate, seven read the region’s language Kannada and only two read English. Their livelihoods were diverse (day labour, farming, animal keeping, trading). All of them, however, lived in households in which at least one mobile phone was already in use. The village is the home of 200 households and a small non-governmental organisation called CKT. CKT is involved in various grassroots activities and is led by an agricultural researcher who owns a farm in the village. Our project benefited from his and his organisation’s support. The project was funded by the Swiss Development Council and run by researchers from the École Polytechnique Fédérale de Lausanne and the Indian Institute of Sciences in Bangalore. Four designers and researchers spent a total of six months on site for field work, tests and design sessions, tackling various interaction ideas. Focus-group discussions and individual interviews helped us to understand the farmers’ needs and tackle the problems they faced.
The goal was to spread information about what the farmers’ peers were doing in terms of sowing (varieties, amounts and treatments used), fertilising (types and amounts) and irrigation. Incidents of pests and diseases featured too, and so did pest and disease control. We wanted the to share information about what approaches they were taking and what results they were achieving in terms of both yield and monetary revenue. The ITC application included information on all related matters. It also informed about current market prices at the local level and at the nearest processing hub. Other items were historical yield data and a one-week weather forecast. All information was conveyed in a mix of text, numbers, images and icons along with audio messages.
By touching an item on the screen for a second or two ("long tap"), illiterate users could start an audio programme that read the information to them. Audio announcements and symbols such as question marks made navigation easy. Audio error messages, moreover, indicated when users failed to enter all information as required. The trial devices were low-end Android smart phones with the ITC apps. The phones were distributed by CKT. Initially, two farmers were taught how the system works, and they explained everything to the other participants. These two leaders later became human access points (HAP) whom the others turned to when they ran into problems or had questions.
The leaders taught the other participants how to operate and charge the smart phones. No mobile internet existed at the project site. All information the farmers entered was sent by text message to a server, which in turn forwarded that information to the apps on the other smart phones.
The participants became quite fond of their new smart phones. They used them daily for calls and for entertainment purposes such as listening to music, watching videos or taking pictures. Often, their batteries were down before the end of the day.
The more active participants used the agriculture application every 10 days on average. Less active participants only used it every 40 days. The farmers most frequently consulted the weather forecasts, but they also showed interest in peer actions, advice and market prices. All farmers made use of the audio read-outs to check numbers and ensure they had understood everything correctly.
None of our participants said they worried about others learning about their farming decisions and possibly copying them. Nonetheless, half of them showed little motivation to enter information. The farmer least interested in the application said he wondered why anybody might care about what he was doing. He argued that everybody was farming in whatever way they thought was right and had been doing so for years. He obviously saw no scope for advancement through learning.
The farmers who took a more active stance tended to be the richer ones. All farmers, however, found it beneficial to get first-hand information from peers, concerning, for example, the effectiveness of applying a certain pesticide. Such locally generated knowledge was preferred to generic expert advice ("against pest Y spray X"). However, only few farmers understood right away that information about what peers are doing can serve as a kind of advice. It was necessary to make them aware of this simple fact.
The ITC system was originally designed to give expert advice on some issues, but this component was not used. It really only applied to wet years, and unfortunately the weather was unfavourable. It became evident nonetheless the farmers – and especially the poorer ones – tended to want comprehensive, step-by-step advice of the kind offered by vendors of agricultural inputs (seed, fertiliser, pesticide).
"Coli should tell me what to sow, when, how much of which fertiliser to put", one farmer said. The farmer who used the application the most wanted more explanations: "When I report a problem, Coli should tell me why it happened and what I should do."
The farmers did not find it difficult to use the app. Most problems they had were not due to illiteracy but to technical incompetence. For instance, they were not familiar with operating a touch screen and found it difficult to distinguish short taps from long taps. Even these technical problems were minor, however. No participant complained about difficulties when using the devices for phone calls or entertainment.
The SDC, our donor agency, was interested in creating an application to boost the productivity of small farms. Our trial proved that there is a real potential for achieving that. In its current inception, however, our app did not provide enough incentives for farmers to enter data and seriously consider their peers’ actions.
Moreover, it became clear that the app appealed more to the richer and more pro-active farmers. This probably has less to do with the application itself than with poor famers’ general risk-averseness. Their main concern is to prevent harm, not to improve their livelihoods.
Accordingly, poorer farmers’ experience did not appeal to the more advanced ones. "I’m not interested in what most of the farmers do, they are too conservative," one participant told us, "I would not learn anything new."
Future work should test approaches in which not all farmers are given equal attention. The focus should be on the more successful and ambitious ones. Moreover, it would make sense to more ensure the accuracy of the information. About 30 % of the data entered in our field trial was duplicate or bogus – it was the result of farmers testing the device or explaining it to one another. Any future system should therefore put more emphasis on verifying data.
The digital revolution is well under way. Villagers in India do not find it difficult to use mobile devices for phone calls and entertainment. Technically, it is no problem to provide them with agricultural extension information. The challenge is neither digital technology nor illiteracy – it is to acquire the appropriate knowledge and to interest farmers in productivity-oriented peer debate.
Hendrik Knoche is an assistant professor at Aalborg University’s Department of Architecture, Design & Media Technology in Denmark, and previously worked as a post-doctoral researcher at the École Polytechnique Fédérale de Lausanne in Switzerland.
HS Jamadagni is a professor at IISc Bangalore’s Department of Electronic Systems Engineering in India.
PR Sheshagiri Rao is a researcher and farmer in CKPura, Pavagada in India.