2,800 crowdsourcing and crowdfunding sites
Can farmers in developing nations use crowdsourcing to improve crops yields and harvest quality?
That’s the question that Jacob van Etten has recently been putting to the test in northern India. We wrote about van Etten, who works at the CGIAR Research Program on Climate Change, Agriculture and Food Security (CCAFS), last week, and recently got in touch with him over Skype to find out more about his project and future plans. Check out the interview below, and let us know what you think of his initiative in the comments.
Anton Root, Crowdsourcing.org: I’ll dive right in – how did you first come up with this idea?
Jacob van Etten, CCAFS: I’ve been working on crop improvement, on and off, for ten years now. It is important to farmers is crop improvement, [but] it’s expensive and requires a lot of institution-building at a local level to organize groups. I was looking for a way to do it cheaper, but still in a very robust way, so NGOs and farmer associations could do it themselves. That’s the main thought behind [the initiative]. Crop improvement has developed over the last ten years or so, but it has not reached tens of thousands or hundreds of thousands of farms. It always remained in small groups.
And what drew you specifically to crowdsourcing?
With crowdsourcing, you start as a crowd, you don’t start with groups. You start with unorganized people who then organize themselves. You create a community, you don’t take for granted that it exists from the beginning. People can engage individually – that’s what brought me to crowdsourcing. I also worked with Vizzuality, which is a company that has done some very interesting crowdsourcing work on weather data. That really inspired me to think this [method] through for crop improvement and farmers in poor countries.
How reliable has the data been thus far?
We don’t know yet. (laughs) We just started in India, and that’s one of the main things we want to find out. On the other hand, there have been other projects that have done this, especially the World Agroforestry Centre. They found that [the data] is quite reliable.
Speaking of reliability, how do you plan to measure that?
We use statistical methods to find out how well we can estimate the different parameters. There’s a formula that can estimate the parameters, as well as other statistical methods.
Obviously the project is all about crop improvement, but are you worried at all that an initiative like this would promote monoculture – farmers planting the same kind of crop?
No, because it’s producing recommendations that are really tailored to each farmer’s conditions. We take into account what each farmer prefers, but also some of the main characteristics – is the farm market-oriented, or not, et cetera. Depending on that, we can also give very specific recommendations back to the farmer. In the end, it’s about diversification. We also always take a broad set of [crop] varieties to give farmers. Each farmer tests three [varieties], but the overall set is bigger. We still need to find out how big it can be. For us, more varieties is better because [we can better tailor the seeds] to the different conditions the farms have.
And how many varieties did you use in India?
In India, we used ten for the crowdsourcing experiment. We have some other experiments where we used eighteen.
A big part of this project is being able to scale up – can you talk about this and what your plans are to implement this among more farmers?
One of the things we try to do is have farmers themselves recruit other farmers, who want to join next year. Basically, this is done by sharing mobile telephone numbers and addresses. On the other hand, we are also looking into the costs of the whole exercise – what is the cost per household, the fixed costs and the variable costs. Based on that analysis, we can then propose it to other types of institutions as an approach to implement on a large scale.
In India, there are different types of initiatives, like direct cash transfers, that are looking to scale, so we are also looking at how to implement this at that kind of scale. To do that, we need some evidence about costs, some kind of business model, and cost structure, and then we can think about the different models we can use in the implementation.
When is the project in India finishing up?
In April, they are harvesting, and after that [we’ll be doing] data analysis. Probably in July or August we’ll have the data to analyze, so by the end of the year we should have the results. The same thing for Honduras, where we’re also doing [this project]. We’re starting in May, and should have data in July or August. By the end of the year, we’ll report on both experiments, jointly.
I read you’re also launching a similar project in East Africa – is that still the plan?
That is true, but it will probably be later. It will start in August or so, and we’ll have the data ready in 2014. The idea is to get a comparative perspective – different crops, different places, different organizational models. We’re really at an experimental stage now. Later on, we’ll sit down and think about possible business models and so on, and see how we can scale up.
How do you actually get the seeds to the farmers?
We’re working with national agriculture researchers and NGOs, as well as local, grassroots organizations. In India, we basically worked with one NGO called IFFCO [Indian Farmers Fertilizer Cooperative], which has connections to local organizations like women’s groups, village cooperatives, et cetera. That’s how it works. In Honduras, what’s quite interesting is that you have local farmer innovation groups. The interesting philosophy of those groups is that they’re quite open to others, so they do their work, but anything they plant is also for use by other community members. One thing we’ll try to do is to also encourage them to invite non-members and to give seeds to people who are not, formally, members of these organizations.
For now, I think it’s unavoidable for us to work with organizations that have some presence on the ground, but the idea is to be really inclusive. That is one of the possible advantages that a crowdsourcing approach could have over a more group-based approach, because it’s individualized and you can include more people. You don’t have to be a formal member of anything. That is the community-forming aspect of crowdsourcing – the starting point is a crowd and you get some sort of community.
Obviously you don’t have the results yet, but is there anything you learned from the project in India that you’re doing differently in Honduras and will do differently in East Africa?
No, I think the essential bits of the approach worked quite well in India. We’ve generated a lot of enthusiasm around the project. We’re working with a different crop in Honduras, so some things we are discussing are plot sizes – the difference is more in the details of the approach. There’s no big lesson we’ve learned in India other than that people like it, it was practical to do and easy to organize. It actually simplified previous approaches, where you had big plots and people had to find the actual land to plant all the varieties together.
Maybe we’ll place more emphasis on explaining to the local scientists how the approach works, and that there are statistics behind it to back it up, so it’s not only about diffusion. It’s also important to have the scientists in the area itself participating, to build up the capacity for people to do it, or to analyze data, at least. That’s something we’ll emphasize a little bit more in Honduras.