The application of big data to agribusiness in Africa could be transformational in unlocking access to financing for an industry that has traditionally been classed by banks as too risky. But how practical a solution is it really? Shannon Manders reports.


“Everybody throws the phrase ‘big data’ around like it’s the name of someone they know.” Omri van Zyl, executive director of South African agricultural industry association Agri SA, believes there’s “no such thing” as big data in agriculture in Africa.

The South African, whose credentials include being president of the World Farmers’ Association audit board and having authored studies on information and communications technology in African agriculture for the World Bank, thinks it’s important to keep the conversation on the topic “realistic”.

The problem, he says, and many other experts agree, is that Africa’s agribusiness sector lacks any kind of meaningful data – a necessary starting point to transition to big data.

According to the Food and Agricultural Organisation of the United Nations, as much as 80% of the farmland in Sub-Saharan Africa is managed by smallholder farmers, working on up to 10 hectares of land. With the majority of these farmers part of the informal economy (they may not be registered, tend to be excluded from aspects of labour legislation, lack social protection and have limited records) there’s simply no technology at play to collect data in the first place – or to feed it back to any kind of platform or system.

Certainly, there are pockets of data collection, storage, analysis, and even some forecasting, on the continent – the majority of which is taking place in South Africa’s private sector. But there are very few initiatives that can be classified as big data, the exact definition of which is constantly in flux, but which is widely understood to be the accumulation of large data sets that allow for predictive analyses and reveal patterns, trends and associations.

Nevertheless, progress is under way. Over the past few years, a handful of multinational organisations, mostly traders such as Olam and Cargill, have been running big data-type projects with African farmers and co-operatives using various tech platforms.

The majority of these projects aim to improve yield, traceability and financial access: the Olam Farmer Information System (OFIS), for one, was launched three years ago as a tool for collecting farm data from Olam’s network of smallholder farmers based in remote corners of the world.

The platform sends real-time data and images on crop progress to the company’s global team of traders and sales personnel, enabling them to better understand an upcoming crop profile.

“Via an app which works offline, we can GPS map and survey individual farms and harness the data to give farmers and our customers insight into the production of crops like cashew and cocoa,” Simon Brayn-Smith, director of OFIS, explains to GTR. “We’ve recently added new functionality to the platform including a comprehensive farm to fork traceability tool that allows customers to understand the exact provenance of the raw materials going into their products, as well as innovative new features like an integrated mobile money solution. This allows farmers to be paid for their crops straight into a digital wallet and build up credit histories over time.”

Olam recently announced that 100,000 smallholder farmers are now registered to OFIS across Africa, Asia and South America.

Farming technology has also caught the attention of entrepreneurs, who have been churning out solutions for small farms at costs that African farmers can afford. Many of these initiatives harness the explosion in the mobile phone market (Quartz last year reported that lower prices for handsets has doubled the use of smartphones among Africans over the past two years, meaning an additional 226 million people now have access to the internet).

In Kenya, M-Farm is a mobile phone app that allows farmers to send text messages requesting information about crop prices. The platform also connects farmers with food suppliers. FarmDrive, another Kenya startup, collects and aggregates datasets from multiple sources, in Kenya and around the world, to build credit scores for smallholder farmers in Africa, thereby facilitating their ability to borrow from financial institutions.

Other solutions have focused on precision farming, the modern-day farming technique that uses information technology, such as aerial images from satellites or drones, weather forecasts and soil sensors, to survey a particular farm on a regular basis. The practice creates an optimum environment for crops and soil to receive what they need – and when. It offers higher productivity, but also saves time, reduces wastage and creates greater environmental sustainability.

Farms making use of these automated systems are scattered across the continent. Zenvus, a Nigerian startup, brands itself as an “intelligent solution for farms” and uses proprietary electronic sensors to collect soil data such as moisture, nutrients, pH levels and so on, which it sends to a cloud server. Algorithms in the server then analyse the data and advise farmers on best practices. The system also deploys special spectral cameras to build crop vegetative health indices, which helps detect drought stress, pest and diseases. Zenvus then uses the aggregated data to build platforms for lending, insurance and trading for farmers.

In Kenya, UjuziKilimo uses innovative sensory technology to carry out soil testing, allowing farmers to understand which crops will grow best on their land, what inputs they should be using and how much irrigation is required. The company operates in real time, meaning that a farmer can have an extensive soil report and associated advice within three minutes of the test being carried out.


Africa’s best-in-class

Few precision farming and big data initiatives on the continent can match the more comprehensive work being undertaken by some of South Africa’s commercial farmers and agricultural co-operatives.

Precision farming is a relatively common practice in this more sophisticated market – and one being used solely as a means of differentiation for commercial reasons.

GWK, a farmer-owned agribusiness operating in South Africa’s Northern Cape, began its precision farming unit in 2006 and started with the concept of data mining two years ago.

“It gives you the competitive edge. You get the best planting data, practices and cultivars,” Dup Haarhoff, head of agricultural services at GWK, tells GTR. The yields of the 86-odd farmers taking part in GWK’s precision farming projects are more than 10% higher than those of the average farmer.

Precision farming for GWK means making use of the very latest technology, including spot and sentinel satellite images, chemistry and physics cards, leaf analyses and a bespoke platform on which all the data is stored to support their farmers in making the right choices and getting the best results.

“One of our successes is the fact that we are ‘growing’ with our farmers as we develop custom-made solutions,” says Haarhoff.

Equally important to GWK’s holistic approach is its team of agronomists (agricultural scientists), which interpret the data to the benefit of the farmers.

“All the imagery and data is of no use if it’s not implemented on-farm,” he explains. “Big data can be very useful if it is used by agronomists with experience. Our secret is that our agronomists visit our farms and make observations on-farm. That links to the technology: you need that link – you need someone on-farm to help the farmer to better his crop.”

It is this agronomical experience to interpret the data that many believe is one of the missing links in the application of big data practices elsewhere on the continent.

It’s just one of the hurdles keeping international companies from taking their lucrative big data projects to Africa – although interest in doing so is growing.

Daniel Redo, head of agriculture research at Thomson Reuters, leads a team of researchers who use weather data and satellite imagery, amongst other tools, to forecast crop production. While their current focus is on specific countries in the Americas, Asia and Europe (and specifically in corn, soybeans, wheat and rapeseed), Redo tells GTR that there are a few African crops on the company’s radar to potentially cover in the future. West African cocoa, South African corn and North African wheat are of particular interest to the company’s clients, which include financial and risk managers, large traders, governments and food processors.

While Thomson Reuters already has many of the required foundational elements, such as weather and satellite imagery coverage, in place in West Africa, what it’s currently lacking is on-the-ground expertise. “From an internal logistics standpoint, you need well-trained analysts in order to look over all that data,” he explains to GTR.

Other requirements include the need to link up with local contacts that can facilitate access to – and engagement with – smallholder farmers to enable a good understanding of what’s going on.

But there are bigger challenges. For many crops in many African countries, the market needs more clarity on what is currently being produced – as well as historical output.

“If you look around the world, you have the US Department of Agriculture, Statistics Canada, Companhia Nacional de Abastecimento (Conab) in Brazil: all of these governments employ people to look at their crops and provide estimates on an annual basis – which then set the benchmark for what the market tacks to,” says Redo, whose job it is to get ahead of those forecasts and inform market participants prior to those government reports.

But, for markets like cocoa – as well as other crops in Africa – he explains, those benchmarks just do not exist.

Another impediment for Africa, from a production standpoint, is that investment into commercial agriculture, in terms of better quality feeds, the input of fertilisers, herbicides, pesticides and machinery, for example, has not been as forthcoming as it has been elsewhere in the world.

“If you go to places like Russia and Ukraine – and other countries – in the last 20 years yields have increased significantly, simply because of the investment into better inputs, seeds and technology,” Redo argues.


Closing the financing gap

It is these problems – and more – that inform Van Zyl at Agri SA’s notion that big data is “not going to change the game” when it comes to facilitating better access to financing for the start of the supply chain.

“The problem in Africa is that you don’t have big farms and the farmers don’t own the land,” he states, matter-of-factly.

The first difficulty for these farmers when it comes to securing loans, he says, is sourcing insurance cover.

Van Zyl explains: “If the farm is not mechanised, it cannot reach the economies of scale to farm profitably. And just on that basis, insurance companies wouldn’t be interested – because the risk is too high. What would you do if the farmer only has four or five acres and plants the wrong kind of maize?”

The reality is that smallholder farmers have to contend with far greater constraints. “You need to make the units bigger, provide ownerships, create access to capital to buy the seeds and plough the land: without these things, the rest of it is not going to happen,” he says.

And even if they did have access to huge swathes of accurate, reliable and timely data about Africa’s small-scale commercial farmers, mainstream wholesale banks will remain reluctant to get involved in financing them directly.

“Very few commercial banks have the infrastructure to bank thousands of small-scale farmers on a couple of hectares each in Africa,” says Zhann Meyer, head of agricultural commodities in Nedbank’s global commodity finance team.

He describes big data as a helpful tool to enable traders and agricultural co-operatives to mitigate their risks (and make them, as the obligor, more appealing to the banks), but not something that is going to change the way the model works.

“The compliance tsunami that banks are currently experiencing, makes it logistically challenging for a bank and a small-scale farmer to engage on a bilateral basis,” Meyer explains. The know your customer (KYC) process, for one, would be very difficult to carry out. “Small-scale farming invariably involves seasonal movement of farmers to other arable areas to let the land rest – and this alone creates massive hurdles in terms of loan administration and recovery.”

This is the primary reason why micro-financiers and regional aggregators, such as traders and some secondary processors – and not banks – are involved in financing these farmers: because they have the infrastructure and the on-the-ground presence to monitor the loans.

Nevertheless, Meyer says that there’s been a lot of engagement – particularly from the insurance companies – to try and establish workable models involving the banks by setting data points and subsequent long-term information collection. “An immense amount of data on rainfall patterns, average daily temperatures and sunlight days are then used as a basis for weather derivatives or index-based insurance policies or products,” he says.

“We looked at these policies in some detail, but it is challenging to make it work on a practical level because it’s not farmer specific.” Such policies don’t address the moral risk of non-performance by the insured farmers, Meyer explains: “When it comes to the setting of premiums, if one farmer doesn’t deliver and has a claim, it’s going to impact the premium of the farmer that does deliver above-average yield in a below-average rainfall year. An unfair policy is not sustainable – and one can only engage on these opportunities if the intention is to make it work in the long term.”


The next step

Bank financing aside, local and global corporations continue to press on with innovation in digital technology to improve the efficiency of food production in Africa.

Smallholder farmers provide up to 80% of the continent’s food supply, and as such are critical to maintaining food security. With some international donors turning inwards (evidenced in President Donald Trump’s decision to significantly axe the budget for the State Department and the US agency for international development), unlocking farmers’, investors’ and entrepreneurs’ potential in this sector will be key to its success.

Government involvement – and investment – in driving innovation will be crucial.

“If one considers food security and sustainability, it should primarily be the role of governments to make sure data points are specific, accurate and wide-spread enough to give a normal bell-curve for the average rainfall for a region, for example,” says Meyer at Nedbank.

If government funds are invested in making that kind of data available, then it is conceivable that some sort of financial product could be developed in a partnership to allow small-scale farmers to access the funds they need for inputs, such as seed, fertilisers and insecticides.

“You could then design a payment mechanism that disburses inputs in a staggered manner to a co-operative specific to a region or community as the land is prepared and the crop cultivation cycle progresses. Every farmer could get inputs in proportion to the land area he commits to cultivate, and then get rewarded for production in line with the yield and harvest in the co-operative,” he says.

In the case of crop damage or failure, he explains, claims could then be paid to the co-operative and managed as a collective – regardless of whether single farmers produce above or below the average “because the insurance product is designed to pay on the basis of average rainfall and heat units on the right time in that area”.

Redo at Thomson Reuters believes that governments could harness the widespread use of mobile technology to produce the benchmarks that the sector is lacking. This could allow farmers to contribute information on what they’ve planted and when – and how much they’re harvesting.

“If enough people are contributing then you’ve got a very rich and accurate data set to better inform the market and the governments,” he says.

But when it comes to reciprocal data flows, which are already being implemented by various agri players in the private sector, Redo advises that farmers being asked to help governments build their data sets will need to be properly incentivised.

“It has to be made clear how that information is getting used – we all have inherent scepticisms about giving away personal information,” he says.

Trust is a big one, and that’s going to take a long time to develop.

Although money is clearly the best incentive, it’s also the most difficult to offer. But this is where African farmers’ lending needs for purchasing inputs could ultimately be met. Redo suggests some sort of loan programme or discounts on seeds, herbicides and pesticides. “The farmers have to see a benefit,” he says.

It may be that Africa’s agribusiness sector needs a change of mindset as much as it needs digital innovation to secure its future.