Every week hundreds of Kenyan farmers turn to online platforms such as Facebook to try and find a buyer for their produce. We’ve been mining, cleaning and analysing this data to see patterns in the way people buy and sell online. By analysing this digital behaviour we can start to see potential future trends in agriculture tech. Here’s what we’ve found.
What types of farmers are selling online?
We tagged each online selling post by type of produce (e.g. tomatoes) and calculated the proportion of posts by produce type. We then compared this to the share of farmers by crop in Kenya found in country level reports. While the country level data is only approximate, you can see there is a strong skew towards fruit and veg when it comes to digital selling patterns. Unlike crops such as maize, fruit & veg cannot be easily stored so farmers often have a very short window in which to sell their produce. And unlike crops such as dairy or tea, the fruit & veg value chain is made up of lots of small, individual buyers which makes finding them a challenge. We think this explains why so many fruit & veg farmers are going online to find selling options.
Diving deeper into the online data can give us splits of the fruit and veg sold by item. Traditionally good data on these sorts of patterns is hard to capture, interestingly the rise of social media and digitally active farmers is changing this.
How likely am I to find a buyer through Facebook?
To answer this question we looked at the most popular crop to sell online – onions. We pulled every onion selling post in the largest farming Facebook group in Kenya (approx 40,000 members) between October 2015 and April 2016. We then analysed each one to see how many people commented on each post expressing an interest in buying. We found that 60% of posts had at least one interested response. Of those that received a response, the number of responders fluctuated between 1 and 9.
We were interested to understand why some posts received no responses whereas others were receiving many. We wondered whether the price people quoted in their selling post was affecting levels of interest so we took all posts that quoted a price and plotted the price against the number of responders.
The lack of correlation seen in this chart shows that price doesn’t appear to be driving success in online selling. While some factors such as seasonality may play a role, we think the fluctuation in number of responses is primarily due to the structure of Facebook Group feeds. Facebook Group feeds are structured chronologically by the date of the last comment. This means if someone comments on a post then the post gets pushed back to the top of the feed, meaning it’s more likely to be seen and commented on by someone else. This virtuous cycle perhaps explains why one post could have got the attention of as many as 9 interested buyers.
What does this mean for the future of online selling?
While Facebook Groups are currently the most popular way to sell produce online, analysis shows that these groups are not well designed for a marketplace. With up to 200 posts a day on some of the largest groups, information gets lost in the noise. Many posts receive no response at all whereas others receive lots of interest purely because they keep getting bumped to the top of the feed. The demand shown by farmers for selling online together with the failure for Facebook to provide a functioning marketplace shows that the time is ripe for a better digital solution to market access in agriculture.