We got together for the Hack4Hummanity Hackathon during the World Humanitarian Summit in Istambul. We worked with Bernard Kowatsch, Andrea Amparore and Hila Cohen from the World Food Programme (WFP). Together we outlined key challenges faced by their organization and small producers, focusing on problems caused by asymmetry of information.

CropCart solves the asymmetry of information between small scale food producers and local traders in developing countries. This asymmetry is particularly prominent among refugees who turn to farming in foreign environments, and are therefore at the mercy of experienced traders who exploit their lack of knowledge of the market prices, patterns and organization. Nonetheless, the solution can be implemented in many other contexts.

Problem Statements

1. Producers have no access to updated market prices, and therefore do not know what is a “fair price” for their produce. Traders often offer a very lower price to keep most profit margins for themselves. Producers also lack information regarding the quality of their produce, and therefore do not know what 'quality bracket' they fall in.

2. Current data collected by local authorities takes 3-4 days to be gathered and distributed, resulting on producers having to deal with outdated and geographically inaccurate data. The data is often misleading since some governments don't use statistically proven methods or rely on biased sources.

3. No analytics are performed on the few data from local farmers being collected by authorities. This does not allow the WFP to define a better strategy for tackling their food issues.

4. Producers often travel long distances to sell their products at markets. When they get there, traders know they have traveled and offer a lower price than expected or than announced at the radio. Producers have to accept though, otherwise they will have to return empty handed and their product might rot.

5. Traders have to travel to many locations to meet producers and buy their produce. This results in a lack of efficiency in many of their processes.

Prototype 'Hacked'

Trader App: This allows traders (buyers), like Sam, to publicize a request for commodities from local farmers, like Johnson. Sam sets the produce, amount, prince range, quality standard, location and time requirements he is looking for the exchange. This request will then be seen by Johnson, who will then send a request to sell to Sam. Sam will then see that he has a request on the App, and he can decide to reject it, or accept it and place a deposit for the purchase. This deposit prevents Sam from taking advantage of Johnson's travel. When the final exchange takes place, Sam will be able to enter the final price and conclude the transaction. Now Sam can better manage his purchases, keep track of his commitments and be more efficient.

Producer SMS Interface Local farmers, like Johnson, will use this interface to interact with our system. Johnson will, in natural language, type a message with the details of what he wants to sell. The message will be processed using Watson Dialog and our system will make sure the local farmer has provided all the information needed for an exchange process. After the data in confirmed, Johnson will see a list of nearby traders willing to buy his produce. The list will include price range information and the average market price for comparison. Johnson will also get a text message with information about the quality of produce he is trying to sell. Now Johnson has a tool to access information about the market price, local traders, and product quality bands. Refugees willing to get into farming will now face one less barrier of entry.

World Food Programme Dashboard All the data inputted into the Trader App and the Producer SMS Interface will be stored in a WFP database. Using dashboarding software and IBM Watson Analytics, the WFP will be able to get valuable insights. For example, how far people travel to sell, who are the most important traders for each product and how respectful are traders being of initial agreements. This will allow them to do actions like identifying food routes that need support, contacting important traders and remove repeatedly negligent traders from the system.

What's next for CropCart

- Develop the prototype into a Minimum Viable Product (MVP) and deploy it on IBM Bluemix.

- Find a mobile payments partner, similar to M-Pesa: to process payments between producers and traders using phones.

- Continue the discussion with the World Food Programme to bring in their expertise and resources, if they wish to contribute further.

- Field test the MVP in developing countries to test growth model, adaptability and scalability.

- Gather feedback from users and get them engaged with the App model to generate buzz.

Business Growth

- Users will be able to use the service free between 2-4 months, depending on the specific harvesting trends.

- After that users will have to pay as small monthly service fee ($1-2) to keep using the system in full and get access to all the analytics brought to them by Watson.

- If users that don't pay the service fee, functionalities will be limited. For example, producers might only be able to see a maximum of 3 traders nearby, as opposed to a the full list.

- Users would hopefully now see the benefit of $1/month and subscribe to the service after 4 months of usage.

- Once usage picks up more data and insights get generated by the system, there will be potential to further monetize this by selling data insights to third parties.

- Our service needs to first grow on the trader side, and then on the producer side. We expect both to grow mostly by word of mouth once we hold workshops with key people in certain geographies about the benefits.

Try it out on your mobile

TraderApp –

Share this project: