WHILE artificial intelligence (AI) is being touted as a solution to many data-related challenges, it cannot generate agricultural market intelligence without human involvement. The most important step in building an agricultural platform is generating and processing data into information and knowledge.
Sources of that data are farmers, traders, vendors, transporters, consumers, formal institutions, and many other actors.
The two main processes of collecting data from these sources are human and digital.
But to what extent can digital data collection alone be relied on without human beings interacting with human sources of data? There is a need for human interaction as part of building relationships and confidence with human sources of data so that they open up and are willing to share data. This is what eMKambo has been doing for more than a decade.
Interpretation brings more value to data
When data is collected, technology may process it into visuals like graphs, but technology won’t be able to interpret the meaning and relevance of the analysed data. Processing price data into visuals like crop calendars requires human interpretation. What is the meaning of this and that graph? It is through interpretation that data become valuable to farmers, traders and other users.
Among the key roles of the human market agents in building an agricultural marketing platform are raising awareness, educating actors about how the market works, demystifying market operations through persistent engagement and building confidence.
A lot of explanation cannot be done by digital technology alone. For instance, if technology sends price information to a farmer, the farmer wants to know why today’s prices are lower than yesterday’s prices. Technology cannot answer such questions.
- Environmental justice: Think big, start small
- In Full: Fifteenth post-cabinet press briefing, June 07, 2022
- 33% drop in Byo diarrhoea cases
- Health talk: Mandatory wearing of masks can now be scrapped
Keep Reading
These questions can only be answered by people physically present in spaces where prices are continuously changing.
This becomes the role of market agents physically interacting with farmers, traders, vendors, transporters and consumers in the market to provide explanations of why prices are what they are and answer other emerging questions that cannot be answered by digital technology or artificial intelligence (AI). If a looming shortage of tomatoes is going to trigger imports from a neighbouring country, AI cannot predict such developments.
Keeping knowledge fluid and up to date
To be useful, digital technology needs to be fed with information. There must be someone collecting information from the source and inserting in digital solutions. For instance, somebody has to feed data like prices, volumes/quantities, varieties, and sources.
That data will enable those trading on the platform to see ground-level information, especially from mass markets where farmers, traders and vendors are interacting. What comes first is understanding the existence of physical markets before onboarding farmers and traders who have been trading physically for decades onto a new digital trading platform.
The other important role of human market agents is to keep market knowledge fluid and up to date. Just like physical markets where updates happen every minute in the form of price changes and some actors moving out of the market, a digital market platform should also stay fluid.
This can happen through market agents providing early morning, mid-day and close of day market updates. Such updates are critical in educating and showing everyone that market performance is not a static affair but changes in response to several factors.
Early morning updates can show that morning prices are not stable because each commodity will be searching for its price. Mid-day updates can show how each commodity has found its price for the market day. End-of-day updates can show how the market is responding to what has happened during the day.
For example, if most cabbages have been bought, the remaining small stock may attract a much higher price. Or if the stock is still huge, farmers and traders may slash prices to finish the stock and prepare to bring new stock tomorrow. Some farmers may want to go back to the farm and not come back to the market tomorrow. So, they sell at a giveaway price at the close of the market day. Others may not want to carry over today’s stock to tomorrow and pay again for trading space to local authorities.
Brokering relationships and building trust
Unless they have met before, farmers and buyers cannot trust digital trading to the extent of concluding deals online without face-to-face-interaction. To what extent can a farmer send his/her commodity to the buyer and be paid later? Can the buyer pay first and then receive the commodity later?
Which comes first, trust or trading? This is where market agents play a key role in brokering relationships, building trust and processing orders to facilitate transactions.
As brokers, market agents make the farmer and buyer comfortable to trade with each other by enhancing trust and relationships.
There are also issues around quality control and product specifications. Who will ensure quality and specifications of different buyers are met before commodities are sourced and delivered? Market agents have to be responsible for this important task, which cannot be outsourced to AI. Usually, the farmer/seller is so passionate about his/her commodity that s/he may find meeting the specifications of different buyers exhausting and costly. It is in such cases that market agents can assist farmers by handling quality control and grading issues.
This can be done by following how mass markets source commodities from farmers and conduct grading, as well as breaking bulk. When commodities are sorted by the market, agents assume responsibility for onboarding tomatoes and other commodities from the market, not from each farmer. The market agents can then enable the digital platform to display big urban markets like Mbare in Harare through digital technology so that traders in Bulawayo, Masvingo, Mutare and other markets can see what is happening in Mbare without going there physically. Institutional buyers like boarding schools and processors can also see what is happening in Mbare online from wherever they are.
Closing knowledge gaps and information asymmetry
Agricultural sectors in most African countries lack consolidated systems for collecting, analysing, interpreting and sharing market information and knowledge. High levels of fragmentation continue to undermine the whole agricultural sector and food systems. Many digital platforms are too narrow to be a full marketplace. Their focus is on trading only, yet trading cannot happen without data, knowledge and information about those who should trade.
Buyers, traders and consumers want to know about quantities, varieties of commodities, sources and shelf life, among other parameters. Some prefer potatoes from Nyanga, others want potatoes from red soils, while others want potatoes from sandy soils. Some buyers want tomatoes from Macheke, while others want watermelons from Chiredzi.
Other consumers want commodities from smallholder farmers because these do not use a lot of chemicals. All these unique selling propositions for different commodities are critical components of market intelligence that should be on the market platform.
A narrow definition of a marketplace will not attract other actors such as NGOs, development organisations like WFP, academics, financial institutions and policy-makers to a market platform. Data, information and knowledge will attract many actors and bring several income streams.
The Industry & Commerce ministry will be interested in a platform that tracks market performance in real time. The Health ministry and WHO will be interested in a platform that captures what happens at the intersection of food systems and public health through mass markets. In fact, data and knowledge will attract more users. Other actors just want data before making decisions about what to grow. Trading does not attract all actors.
But information for decision-making, like a crop calendar, attracts many new farmers. Before planting a new crop, some farmers may want to know its market performance.
On the production side, market agents can collect data on surplus commodities ready for the market. Such early warnings will show when farmers are ready to supply particular commodities. When analysed, the data will reveal shortages and surpluses in the next few months so that pre-emptive action can be taken by decision-makers.
Farmers have their own structures for peer-to-peer knowledge exchange, where the lead farmer is the main convenor. The lead farmers can be some of the local aggregators who assist in consolidating different kinds of commodities from diverse farmers within their groups.
For traceability, the local aggregators can keep a register of farmers from whom commodities are collected and assist in educating farmers that going to the market individually is not cost-effective. Traceability starts from farmers through local aggregators and market traders. The information flow for trading comes through market agents working with local aggregators, traders and vendors. The other information flow moves from the vendor, the trader and the local aggregator. Below the vendor are users/consumers. Market agents can consolidate all this contextual information. AI cannot go to such levels of detail.
