On Friday, we had the startling realization that fishermen do not use mobile phones to compare market prices of fish. Besides cultural and legal limitations to their ability to sell in different markets, as Tony Roberts points out, fishermen don’t see neo-liberal wealth maximization as their ultimate goal in life.
But what about farmers? Could they be using mobile phones to learn about current market prices, and thereby maximize incomes? Well, they might be, but they are probably not using SMS to learn about market prices.
Join us for ICTforAg to explore the many ways farmers are using mobile phones to improve their lives.
The Case Against SMS Market Prices for Farmers
In “Why Don’t Farmers Use Cell Phones to Access Market Prices?” researchers Susan Wyche and Charles Steinfield of the Department of Media and Information at Michigan State University, looked at the use (and disuse) of mFarm by farmers in rural Kenya.
Despite the hundreds of thousands of donor darling dollars showered on mFarm, Wyche and Stienfield found seven simple and obvious reasons why farmers do not use SMS text messages for market price information.
1. Mobile Phones Are Social Tools Not Information Delivery Platforms
Farmers primarily perceived mobile phones as devices that can support verbal communication among their friends and family, not a business tool. This reference even extended to the most popular input keys: the green call and red hang-up buttons on a feature phone.
The response in Kenya is mirrored by similar responses across the continent that consistently show that mobile phone usage is primarily for maintaining individual social networks, which are vital to survival in extreme poverty contexts.
2. Text Messaging Isn’t Easy
mFarm, like many SMS-based services, requires the farmer to start and modify the service using text messages. While texting can be difficult in English, its much harder still in local languages like Swahili, Bukusu, and Luo, which have long words that can be spelled multiple ways.
Farmers would need to understand how to input letters, spaces, and symbols, and switch between upper and lower case letters, all on a tiny keyboard that could have T9 predictive text. Oh and do all this correctly in Swahili or English.
3. No Confirmation That a Text Was Received
When farmers call someone, and speak to them directly, they know their message was received. However, if they send a text message, they never know if the intended recipient got the text until they respond.
This makes text messaging an unreliable communications channel. Farmers don’t know if their text went through or not, or if it was read or not, or if the recipient just didn’t know how to use a text message, or felt it was too difficult to respond. With all that uncertainly, it’s just easier to call.
4. Using Mobile Phones is Expensive
We may not appreciate how hard a poor person works to manage their money. When farmers are making a few hundred Kenyan shillings a day, every shilling counts. In this context, sending a text message feels expensive, especially when they don’t know if the person received the text.
Even voice calls can get expensive fast, but at least the farmer knows if the other party received their news, and what their response might be. That can matter, and easily be more cost-effective than a text message.
5. Reading a Tiny Screen is Hard
Farmers typically have poor eyesight, which is a consequence of age, health, and using fuel-based lightning vs. electrical lighting in the home. With this reduced vision, farmers found the small screen of their feature phone handsets too inadequate to read even short SMS text announcements.
The earlier language literacy issue compounds the eyesight problem to make small feature phone screens pretty much unusable for many farmers in rural Kenya, even if they can read the letters when they squint.
6. Harder Still With a Poor Quality Handset
Farmers typically received their handset from a friend, or its was bought many years ago and is now a relic of the past. Either way, phones are held together with rubber bands or string, making it very difficult for farmers to use their phones.
That’s before screen parallax and straight-up broken or cracked screens that obscure the handset screen completely, or keyboards so old that letters and numbers are worn away.
7. Or Without Any Battery Life
The researchers found that many handsets were turned off to preserve their charge, as access to electricity is still very low in rural Kenya. Some participants even talked about charging their phones up to 10Km away in a charging kiosk.
These power issues make it extremely difficult to send farmers real-time price information, and combined with the other 6 reasons, serve as a disincentive to using SMS text messages to find market prices for agricultural products.
Could it be that other technology tools are better than SMS text for disseminating market price data? Like voice calls, or FM radio? Or maybe even old-school printed paper communication?
Good to see you boldly questioning the merits of certain widely-promoted SMS projects in the ICT4D sector that often have very little evidence of impact but continue to garner enthusiastic donor support. Another recent study on a program called Mobile for Water (M4W) in Uganda reached very similar conclusions: people don’t want to spend their time crafting some coded text message when they could just call someone instead.
http://opendocs.ids.ac.uk/opendocs/bitstream/handle/123456789/11544/WaterAid%20Policy%20Brief_Online%20FINAL.pdf?sequence=1
As with the fishermen story, something that never seems to be asked at the beginning of these projects is “what are the biggest challenges facing farmers in rural Kenya?” If such an open-ended question was asked, I doubt many of them would respond with, “I don’t have enough information on which market has the best prices for my crops.” I bet they would list things like lack of water, lack of fertilizer and farm equipment, and lack of access to capital; problems that don’t have the easy solutions some donors seem to be looking for these days. Although it will disappoint economists, farmers might not be seeking to maximize their utility because they have bigger things to worry about!
Great find, John. I really like this lesson learned:
M4W’s reliance on SMS turned out to significantly hinder uptake by local water users and user committees. Local people preferred a direct phone call to their hand-pump mechanic to report a breakdown over an SMS, because it was easier to explain the problem in detail, agree a convenient time and set out a programme of action for repair.
I researched this back in 2008 with Save the Children, and found another reason that may be most strongest of all – most farmers do not sell their goods directly to market. They sell through middlemen. Farmer after farmer told me that it wouldn’t matter if they knew the prices at various markets. When the middlemen came with their trucks, they were given a take it or leave it price. Knowing how much the goods were sold for only made them feel like they were being even more ripped off. And for the farmers that produced enough quantity that they could serve as a middleman, they were intentionally cut out. I spent time with a woman who grew over 60,000 pineapples a year, properly sorted and graded them, and said that she could easily rent multiple trucks to bring the goods to the capital where they were worth 10x as much. However, when she tried this, the retailers refused to even talk to her, calling her a poor farmer and that the middlemen were more reliable. So… knowledge doesn’t always equal power if the power structures are seriously titled against them.
Sean, this is the very reason I always suspect the market price liberation theory. Typically, agricultural producers, be they farmers, fishermen, dairy or meat producers, are part of an ecosystem that is decades, if not generations in the making. Merely adding in new price information doesn’t change the power asymmetry, it just showcases how wide the asymmetry is.
I don’t have any experience with M-Farm, but I don’t buy the premise behind some of these. Mobile phones are __both__ social tools and information delivery platforms in Africa.
#2 & 3 ignore that SMS is widely used in many communities. Nokia has had T9 in Swahili for a decade. People in rural Africa seem to text as naturally as Americans. I’ve seen little picture dictionaries sold in rural Kenyan markets to help preliterate people use numbers as simple codes for the pictures so they didn’t have to spell complicated things.
#4 (price) is something that every rural African is hyper-conscious about. If the SMS app is actually adding to their income, they are very savvy about using it; if not, they’ll quickly abandon it.
#5, 6, & 7 are all true, but easy to overstate. Africans deal with these problems every day. I’ve seen plenty who only turn on the phone when they expect to use it. Again, though, if the SMS service is actually helping their income, they’ll use it; if not, they’ll abandon it.
My experience is that if Africans aren’t using something, it’s usually because it was designed by westerners without their input and ignoring actual field conditions. If it really is helping them, they’ll use it. Better yet, if they’re involved from the beginning, it’ll be hugely effective.
And that’s the greater issue, isn’t it Stephen, that tools like mFarm and others can sound cool when pitched in an app contest, but don’t really work in the real world.
If flipped, this seems to me like a great start to a requirements list, assuming the needs and use cases are first anchored with real world evidence. (That could be a need for price or any other information need.)
While I intuitively agree with some of the above, I suggest that we need both honest, rigorous scientific research to determine to what extent mobile applications are used and to what extent they are useful, and careful reading of the evidence and conclusions drawn from the facts.
The article that Wayan paraphrases – Wyche/Steinfield – is neither rigorous nor scientific. It is replete with assumptions drawn from desk research that includes questionable sources which are, in some cases, themselves the result of literature reviews. Further, the authors attribute findings to the referred sources that are at odds with the conclusions of the papers cited.
For instance, the claim that “farmers typically had poor eyesight, a common consequence of living in off-grid settings
where fuel-based lighting is the norm” is attributed to Kittle. (Kittle, J. (2008). Reading in close–‐up settings will damage your vision—A myopic understanding? http://healthpsych.psy.vanderbilt.edu/2008/ReadingVision.htm, Vanerbilt (sic) University, Department of Psychology 2008))
The Kittle article does not result from any field research, but is a literature review. A reading of the source reveals that there is no mention of farmers in the article, nor of fuel-based lighting. In fact, the article concludes that there is insufficient evidence that reading in low-light causes myopia to warrant even a recommendation against the practice. Yet, Wyche/Steinfield cite this as if it were scientifically proven that farmers have poor eyesight and that the cause of such is lack of bright lighting.
Similarly, Wyche/Steinfield cite Burke, et al., as the source for their claim “poor lightning conditions, and rural farmers’
inability to afford glasses, (are) reasons for widespread limited visual acuity among this population,” when in fact Burke, et al. is concerned with presbyopia, a condition that occurs as a part of normal aging and is not considered to be an eye disease.
Unfortunately, the attribution of findings of the referred sources that are at odds with the true conclusions of the research cited is also in Wayan’s piece. Regardless of my opinion on the quality of their paper, in fact Wyche/Steinfield do not present their research as a condemnation of the use of mobile. Rather, they conclude that software developers and development practitioners should be be encouraged “to adopt an ecological perspective when creating mobile applications for sub-Saharan Africa’s rural” and urging the development of “innovative educational interventions.”
The answer is not “old-school printed paper communication,” but better use of better research including examination of how and why people use the technology they have, how to provide people with better technology resources, and how to educate and train them to get the most out of the tools they have available.
@Andrew Sideman. Questions about (without, to be frank, reading Wyche/Steinfeld). What about the substantial body of research that shows impact of mobile phones and SMS in terms of use and in terms of increased incomes (Martin/Abbott, “Development calling”; the substantial research in Tanzania of Simon Bachelor, I think dating from 2007, etc.)? And what are the factors that influence use – language group? educational attainment? geography / logistics (in terms of access to markets?). I don’t doubt that there are instances where mobile phones fail to provide benefit to farmers (e.g. per @Sean Blaschke the need to use brokers), but there’s a reasonable body of work that supports the intervention. Best not to rush to judgement.
I would like for you to define these terms “rigorous” and “scientific.” The community of ICTD researchers and practitioners needs a much broader education on empiricism and research methods. Valid and necessary critiques of work by economists (like Rob Jensen) will not come in the form of RCTs, bogusly defined as the “gold standard” of research (and RCTs are absolutely contentious and controversial among development economists).
If you are trying to figure out broadly the barriers to design of interfaces for poor, low income rural farmers, you need to use exactly the methods Wyche and Steinfield have used, up close conversations with farmers, one on one. Jensen’s study of fishermen in Kerala was rigorous in one respect, and absolutely lax in another (his work was very much done at a distance). An RCT could be carried out on the variables Wyche and Steinfield identify, but that will never happen if the foundational research, inductive (not deductive), doesn’t take place first.
Hi Jenna, i strongly agree with your critique of RCT fetishism. They’re generally unlikely, impractical and unaffordable. I tend to think they’re also unnecessary (but perhaps this very conversation suggests otherwise).
But isn’t this really a big-problem? Don’t we want to know if mFish or mFarm or m4W(hatever) has impact on incomes or on some other indicator? Over time, so that the “noise” of weather and other factors is controlled. I dunno the indicator, actually: Perhaps it’s mPesa transactions by users and non-users.
Two other notes: Trying to figure out interface design, as you rightly suggest, MUST rely on conversations with potential users (and on rapid prototyping to be sure that they have something to talk about), but figuring out the impact of designs once fielded requires a different approach. No? AND: Just to be clear, “rigorous” (=stiff?) and “scientific” (=performed by a scientist?) are Mr Sideman’s criteria, they are not mine. My only concern, as usual, is to be sure there’s no baby in that bathwater.
Well put, in my own experience these m-agric solutions have served developers and NGOs looking for ”impact” more than they have saved farmers from rogue traders and monopolistic buyers. They probably work best were the farmer has looked for a suitable solution as opposed to some program supporting market linkages for 10,000 farmers.
Love this discussion, as it mirrors a number of our findings at VOTO mobile.
In our work with rural communities – and with farmers in particular – we’ve found that while mobile phones can be very effective information delivery tools (contrary to finding #1 and building on Stephen’s comment), but that the particular channel in question – SMS – may be a less effective mobile channel than voice calls, and particularly interactive voice response (IVR).
In fact, when we were launching a new agriculture project in Ghana we asked farmers if they’d be interested in receiving daily information (on things like weather, market pricing, etc) and they said – but 95% preferred to receive the information via IVR vs. via SMS, which makes sense given findings #3,4,5,6. Engaging in two-way SMS conversation is inconvenient to say the least, but in our experience this is less a factor of “mobile isn’t the right tool” than “SMS isn’t right channel”. In other contexts in rural sub-Saharan Africa, we find a 10X engagement rate when we use IVR vs. SMS.
So while I agree with the findings of this report – and appreciate the sector digging deep into what works in reality and what doesn’t, I continue to believe that mobile interventions can provide access to timely, relevant information to rural farmers (and other populations disconnected from ther internet).
Thanks for starting an important conversation, Wayan! If others would like to continue it offline, please don’t hesitate to reach out – I’d love to chat! (rebecca@votomobile.org)
This is a great discussion, and frankly well overdue. Although no longer with the Esoko team, I spent ten years with them delivering market prices to farmers.. and we spent millions of dollars. But a few clarifications because I feel this often gets to be a religious or binary debate, and I think it’s much more nuanced than that. First of all, I never found a huge willingness from donors to support market prices… we spent millions of dollars of private funds (and much contributed by myself) because we believed there was a commercial opportunity. Most donors and M4D practitioners we approached on market prices were, and remain, highly sceptical. I would also add that all the points made in the initial post have some truth to them… phones are primarily for talking, they are about connecting people by voice, literacy is a huge issue, phone sizes and languages are difficult. So in general, after years of trying, I’m generally quite cynical about an SMS price delivery project. But to be honest, I’m not aware of any such projects or businesses where that alone is the intended intervention (mFarm is more about facilitating linkages to markets and finding buyers). We found early on, by actually engaging with and talking to the farmers, that price information alone was not sufficient. I would suggest that farmers would wish for financial services, weather information, market linkages/prices and agricultural advice in that order. So we, like all the others I’m aware of, do not ONLY offer market prices, but offer it as one feature of a bundle of products that clients or intermediaries choose is the most relevant for themselves or their clients. The other critical learning is that it’s not about SMS vs. voice vs. radio. vs. IVR. The truth is that you are engaging with a very diverse market, and they all may have different preferences of accessing this information. Trading partners may be comfortable with smartphone data-driven displays. Rural communities generally ‘get’ the market prices on SMS, but light up when you demonstrate how they can call in and speak to someone in their own local language using our Call Centre. And when challenging farmers on whether ‘we’ can tell them ‘anything’ over a ‘160 character message’ that most of them ‘with respect’ can’t read… they would respond… “but there’s always someone in the house, or the community that can read it for me, that the agric work we’re doing now is different to what our parents know, and that we can save/share/show the information when and how we need to”. Anyway, it’s complex. When we started, I really imagined that market prices were such a simple concept, and given almost 98% farmers stated clearly they wanted them, and that we suddenly had a novel (personalized) delivery channel, that it would only be a year before thousands of farmers would be subscribed. The honest truth is that numbers of subscribers remain shockingly low, even when the service is offered for free. So it begs the question, is it really useful? Again this isn’t a yes or no answer. For some absolutely. And I have many personal experiences of farmers who are so grateful for the information. But it’s nowhere near the numbers we had anticipated. But they do talk about how the message allows them to negotiate a better price. And although i agree infrastructure and access to markets is important… I think that the power dynamic embedded in trade is more important than anything. If you pitch a trader to a farmer… they’ll bully the farmers into the lowest price (mostly). But I’ve seen farmers who can’t even read the SMS messages, telling me how they show it to the trader… It’s the fact that the farmer now has evidence (not the information), that changes the power balance somehow. We’ve had farmers call in to a radio show how after getting market prices, the entire relationship of mistrust with the traders had evolved into one that was more constructive, even where the traders were now loaning money to the farmers. In a fascinating study by NYU over 2 years, they did an exhaustive and expensive RCT, which demonstrated a 9% revenue improvement for the farmers. More interestingly, they showed that the price information was really changing the behaviour of the buyers, and less the farmers. This was evidenced by the spill over effect they refer to, where the control group started to get the better prices, but when asked, still demonstrated a lack of knowledge of the current market prices. We’ve even had traders pushing out market prices to their farmers because they’re sick and tired of the farmers being convinced the market prices are actually higher than they were! My conclusion is that of course knowing your markets will give you an edge. So in general, I would agree with all of you… there are significant challenges. Clearly some people benefit (and they can benefit a lot) but the others may not be as driven as we think they are by our own assumptions about economic behaviour. That’s where I think the anthropology is missing. Actually, in general, I’m shocked that over the ten years that we were doing this, not a single donor actually came in with any substantive research to understand what we were doing, and the impact. Only CIRAD and NYU bothered to try, and we’ve always been seen as the biggest market price service on the continent. I would also note that unless you invest in enough markets (we do 46 in Ghana) and invest in enough local community relationship building and training, it will probably fail.. and that’s just really expensive. So are we even measuring well deployed services? Finally, I would just re-iterate that this isn’t a binary debate… market prices do form an important component of any intervention and different farmers will value and use them differently. Certainly, not as many as we had all assumed. We’ve graduated early on such that only two of our ten countries actually deliver market prices that we collect, and we’ve learned to offer a much broader range of services (weather, tips, planting, disease, prices, promotions) over a series of channels (sms, android, voice, tv, radio, print, call centres). The fact that you have to do all this to create a viable service is a huge barrier to anyone, and why most don’t take on the entire package of services, and why it’s much harder to conceive of and imagine and measure the kind of mixed services and channels that are offered. One more thought… I think there’s one other big transition that we’re all seeing, and that is it’s not so much what you can push or deliver into these communities, as it is how you can enable these communities (and the businesses that serve them) to report back data (planting, acreage, production, diseases, stock etc.) which is where much of the focus is. So not only do you need a multitude of services, over a multitude of channels, but you also have to enable a two way type of communication. All the comments about farmers not interacting with phones is true from our observation, so that has its own set of challenges. My personal feeling is that we will see a category of local ‘mediators’ emerging as enough services emerge (financial, health, agricultural, social etc.) and at that point, with data, tablets and commissions, we’ll see real local activation of these communities using digital services… a bit like the old letter writers or typists who sat outside government offices or railway stations.. new technology has always presented huge challenges and has generally disappointed the dreamers. My advice is to take stock of what we have, be a little more patient, understand the real costs, and all take an anthropology lesson. But don’t throw out the prices yet! Bundle it and let the clients choose their content and channel mix.
BTW, the NYU study can be found at:
https://drive.google.com/file/d/0BwpF_MREiy8jSUp3Z3BiQXFucUE/view?usp=sharing
And the abstract is here:
We conducted a randomized field experiment to determine the impact of providing rural farmers with commodity price information delivered via text messages on their mobile phones. Using a novel index of inter-village communication networks, we show that the intervention:
(1) led to a sustained positive increase of about 9% in the prices received by treatment group farmers, and (2) had substantial indirect benefits on the prices received by certain control group farmers. We discuss a novel mechanism of bargaining spillovers which can explain the rise of such positive externalities, even in the absence of information sharing between the treatment and the control groups. Accounting for spillovers is crucial because otherwise the longer-run estimates would be biased and one could erroneously conclude that the intervention had no long-run benefit for farmers. The direct return on investment of the service exceeds 200%, a result that underscores the huge potential of ICT interventions in emerging markets.