Kenyan farmers try out SoilCards

By Dr Tempest van Schaik, 26 February 2018

Visiting Kenya was a landmark event in SoilCards development. It’s the first time that our prototype was in the hands of customers. Luckily it was a hit with the farmers we spoke to, with one even saying “we need these yesterday!”.

SoilCards: soil testing for every farmer

The work of small-scale farmers is vital, producing ~70% of global food supply. But these are some of the poorest people on earth, with over 1 billion living below the poverty line.

SoilCards are the devices that we’re developing to make soil-nutrient analysis easy, quick, and affordable to smallholder farmers who don’t have access to soil-testing laboratories. SoilCards are made entirely of paper, so they’re ultra cheap and portable. They’re inspired by the home pregnancy test, so they’re easy to use and read. They use microfluidic technology which we’ve translated from medicine to agriculture, but so far they’ve been a proof-of-concept device which has been developing in the lab.

SoilCards

The soil to be tested is added inside the paper SoilCard (right), where it is analysed using microfluidics technology. Spots appear on the display, which represent the levels of the soil nutrients N, P, K, and pH. These test results are compared with a reference card (left) for a specific crop that a farmer wishes to grow. This allows the farmer to see precisely which nutrients to add to the soil.

Why go to Kenya?

Up until this trip, we had been designing and building SoilCards in London, far away from our intended users (smallholder farmers in developing countries). We had tested ideas with UK farmers, but African farming is very different. Thanks to a grant, we were finally able to travel abroad to understand the context that SoilCards will be used in, and get feedback directly from our users. We chose Kenya partly because of existing links between our partners, the UK’s National Institute of Agricultural Botany (NIAB) and the Kenyan Agriculture and Livestock Research Organisztion (KALRO). There are also three important characteristics that helped us choose Kenya:

  1. Kenya has loads of farmers. About 75% of working Kenyans make their living by farming. Compare that to South Africa where it’s around 10%, and the UK where it’s around 1.5%.
  2. Kenya is fairly well-off, with the largest GDP in East and Central Africa, and it has had some good agricultural programs. We therefore expected that some farmers would already be aware of the benefits of fertiliser and we could focus on the benefits of soil testing. It would be far more difficult to promote soil nutrient testing in an environment where people are not even familiar with soil nutrients.
  3. Kenyans are tech savvy. They’re famous for pioneering M-Pesa, which sets a precedent for successful adoption of simple, powerful tech.

What we set out to learn

Having never been to Kenya, we had a long list of questions for our trip, including:

  • What is life in Kenya like, what do rural and urban areas look like? We needed to get a basic understanding of the context SoilCards would be introduced into.
  • Who are our customers and users? Will SoilCards be used directly by farmers themselves? Our mission is for SoilCards to be a self-sustaining social enterprise (rather than relying on grants), so we need to find out who the key payers are.
  • What do stakeholders think of the product and how would they change it? It was critical to get feedback on ergonomics and product design from the users themselves.
  • What are general attitudes to soil health and testing? Are farmers open to the concept?
  • What channels could we use to get SoilCards to our customers?

The research team in Kenya

Travelling from the UK was myself from Science Practice leading the project, and Eric Ober from NIAB, who has a wealth of knowledge about UK agriculture and soil health. Our hosts from KALRO in Kenya, whom we spent all 8 consecutive days with, were Joseph Mriti, Esther Muriuki and John Mwangi. Joseph, Esther and John are agricultural scientists based at one of KALRO’s Nairobi offices. Not only are they experts in soil science and farming intervention studies, but crucially they are experts on Kenyan farming in particular.

SoilCards team

The team in Kenya, in the Great Rift Valley. From left to right: Tempest, Eric, Joseph, John and Esther.

Who we interviewed in Kenya

In Kenya we spoke to stakeholders from across the spectrum of soil testing:

  • 28 smallholder farmers from 3 different regions (Nairobi, Kikuyu, Naivasha)
  • 26 government and research institute soil scientists and agricultural researchers
  • 5 extension officers from Ministry of Agriculture
  • 2 extension officers from private agrochem companies
  • 5 owners of agrovets (the stores where agri products are sold)
  • 3 startups who pay for soil testing
  • 1 fertiliser company
  • 1 farm manager of one of the largest farms in Kenya

In Kenya, almost all farming is done by smallholders. According to the FAO, the average income in these households is $1.40 per person per day. The smallholder farmers we spoke to ranged in age and income level. While most had small plots of one acre or less and were struggling to make a living, some had larger farms of a few acres and had other people working on their land. Many farmers were self-taught or were taught how to farm by their community, rather than ever receiving any formal training or advice. Most farmers sell their produce at their local market. Their main challenge is low/fluctuating market price for their crops, which makes it hard to cover the cost of farm inputs like fertiliser.

Our approach: prototypes and contextual interviews

Prototypes

There was lots of uncertainty around the timing of the trip, which was delayed by the anxiously anticipated 2017 general election in Kenya. However as soon as we got the go-ahead for the trip, we worked tirelessly in the lab at Imperial College with Dr Agnes Chi Leng Leong to make a whole box of prototype SoilCards by hand:

We brought along prototypes that closely mimic how SoilCards will work in future, to demonstrate to different stakeholders. The demo involves placing some soil into the SoilCard and adding water, and after a few minutes spots on the display begin to appear. When the blue bar at the end changes to pink, the test is ready to read. The number of spots is compared to the reference card. After demonstrating it once for them, the stakeholders were given a chance to try it for themselves.

Contextual interviews

We insisted on doing farmer interviews on their farms, rather than meeting at say, a school, which is a normal community convening point. This was to give us a chance to observe the surroundings and see exactly what conditions SoilCards will be used in: for example, the brightness of sunlight, availability of shade and shelter or places to store possessions, and quality of roads and transport options (boda boda!). There were two occasions when we asked for water for the demo (water is required to use SoilCards) and it was scooped up in a banana leaf. Without being there in person, there would have been no way of knowing that a banana leaf is the handiest vessel for a farmer to carry water in. For the same reason, we met soil scientists at their laboratories, fertiliser manufacturers at their factory, and shop owners in their stores.

I was honest about being an engineer who doesn’t know much about farming, and my reason for being there was to learn from farmers. We told farmers that we are building the product for them, so we wanted to hear from them how to make it better. It’s hard to know for sure whether our interviewees were just being polite, but we tried to avoid yes/no questions framed in a positive way, like “Do you like SoilCards” by asking instead “What (if anything) would change life on the farm if you had SoilCards?”.

Going with the flow

We had originally planned to interview farmers one at a time or in small groups. However bigger groups gathered and we didn’t want to turn enthusiastic people away, so we just went along with it. We had spent days writing detailed questionnaires, but in the end, having a chat seemed much more natural. I tried to maintain an open, honest, relaxed and collected composure, even when I stepped into a muddy irrigation channel, disturbed a nest of soldier ants, and had a little girl hugging my calf during a demonstration.

Answers to our questions

The response to SoilCards was overwhelmingly positive. Farmers were intrigued by the devices and were delighted when the spots appeared. They could immediately see the value of knowing nutrient levels before buying fertiliser. After one demonstration they spontaneously started explaining SoilCards to other farmers in the group, and almost every farmer wanted to buy or keep the SoilCards (of course I reiterated that these were just prototypes, not a finished product).

• Attitude to soil health: The farmers we spoke to all understood the importance of soil health and regularly applied fertiliser (organic and mineral), and crop rotation was fairly common practice too. They knew that soil testing was important but said that it was too expensive, and delivering the soil sample required too much effort. To understand soil health, farmers observe how well crops are growing, but this is a costly and lengthy experiment. They tend to apply the same fertiliser that they always have, adjusting it intuitively based on growth. Because SoilCards make soil nutrition visible, farmers instantly recognised that they would take the guesswork out of estimating soil health, and would help them avoid spending money on fertilisers they don’t need.

• Product design & usability: The farmers we spoke to found the spot display easy to understand, and we noted small language modifications like referring to acidity rather than pH. At the end of all the interviews, we had a workshop to decide the easiest way to communicate the actions farmers should take after a test: what fertiliser to buy. We were consistently told that the card should be bigger which was not surprising; engineers in the lab tend to want to make things as small as possible, and farmers hands are larger than ours. One piece of very honest feedback we got was that the dark pink spots looked like drops of blood, an off-putting connotation which we will try to fix.

• Connectivity: Most farmers we spoke to had a mobile phone and access to somebody with a camera phone, but few had smartphones. The M-Pesa payment system is widely used. They were keen to photograph the test result and send it if they could get a detailed interpretation back.

• Distribution: Different stakeholders had different opinions about how best to get SoilCards to farmers. Some farmers favour a free market approach and would prefer to buy SoilCards directly from the agrovet stores where they buy fertiliser. Some agricultural researchers felt that SoilCards would deliver most impact if distributed within a farmer training program. We will probably explore a variety of distribution channels and see which ones have the most impact.

Unexpected findings

Discovering new customers: We discovered two new customers on our trip. Both have an interest in increasing farmer productivity and already pay for soil testing:

  1. Contract farming companies: These are large companies who provide smallholder farmers with seed and farm inputs, and buy the produce from farmers at the end of the season at a pre-negotiated price. These companies obviously have an interest in farmer productivity and they promote soil testing. I actually learned about this because I stepped on one of their branded seed bags while walking around a farm, so these unexpected discoveries are another reason to do contextual research.
  2. Businesses who provide loans to smallholder farmers: some businesses provide their agricultural products on credit. They have an interest in farmers being able to pay back loans. They pay for soil testing and pass these costs onto farmers, so they would welcome cheaper, faster testing.

SoilCards can facilitate dialogue: We discovered a new use for SoilCards, which is to allow a more meaningful discussion between farmers and anyone who wants to increase farmer productivity. Advice about managing soil nutrients is much more convincing, engaging and persuasive if farmers can see results of their actions. We spoke to one soil scientist and adviser who goes to check up on farmers. She said that she would like to include SoilCards in discussions with farmers because it would make soil testing much more tangible to them. She said that she sets off each day with a backpack for her equipment and lunch, and the best way to get to farms is on a Boda Boda motorcycle taxi. She has to be very selective about what she carries in her bag, but because SoilCards are so small and light she could easily bring them on her visits. We learnt that it is crucial to calibrate the SoilCards display spots to Kenyan lab standards so that the spot display can be interpreted by both farmers and scientists, which facilitates a seamless discussion.

Conclusions

Feedback on our approach

We were told by our hosts that our approach was unusual in that we brought early-stage technology to demonstrate, rather than a finished product, which is the norm. They said that this worked well because it meant that farmers and other stakeholders are involved in the design and development process. This is much more likely to result in a product that farmers will use.

What’s next for SoilCards

We are currently looking for funding to continue developing SoilCards. Our trip has demonstrated demand for SoilCards, so we need to go back to the lab to refine our prototype and turn it into a fully functioning product. Ultimately our goal is for SoilCards to be a social enterprise which can sustain itself through sales of an ultra-affordable product. Please get in touch if you know of any funding opportunities!

In the longer term, we see huge potential for farmers to photograph their test result on their phone and get more detailed recommendations back by SMS. This could include information about what specific fertiliser products are currently available at the shops. The discrete spot display will be trivial to count and digitise, whereas colour hue from a traditional chemical test is notoriously difficult to determine using computer vision. Test results could also help build up a detailed map of soil health which could be very useful for government and anyone wanting to monitor and improve farm productivity. Lastly, there is no reason why SoilCards couldn’t work beyond Kenya, and we hope it will make an impact on farmers in other parts of the world too.

Thank yous

We are very grateful to our hosts and research partners at KALRO who did a fantastic job making everything run smoothly. There was an incredible amount of preparation leading up to the trip, and they worked with us for 8 straight days without a break. They were enthusiastic and hospitable and flexible when we wanted to do things differently. We could not have done it without them.

We are also extremely thankful for all the time that each and every stakeholder interviewee gave to us, and for their invaluable, honest feedback.

Looking for a good problem?

We are a close team of designers and researchers who are passionate about tackling ambitious and important problems. If you’re looking to grow your impact, we’d love to hear from you!

Get in touch

Privacy Policy