In the last period of time I have been talking with lots of different people about crowd-sourcing. In between them academics, practitioners, people working in big organizations like the World Bank and the UN and people working in small NGOs.
In the current project I am working on, PakReport, we are using the Ushahidi platform to map issues and needs related to the Floods in the country and with the use of Crowdflower, we are crowdsourcing the mapping and categorization tasks by allowing volunteers from all over the world to connect to the website from wherever they are.
The most common request I received from the people I talk to, apart from the verification issue, which I already discussed here, is to show some success stories.
I always find this question interesting, and until some time ago, I was always struggling to be able to find out how I could prove to people that what I was doing was successful. Of course, to understand if it is, I need to understand what success means to the people posing the question. The common answer is that success in using crowdsourcing for emergency setting, is to be able to prove that your information have been used by humanitarian workers to save lives and to provide help to affected communities.
Now, if this is what success is in using crowdsourcing, my answer is always the same: I don’t know. At this point people get a bit upset and disappointed, and start asking me why I do what I do if there is no guarantee that my job is in any way successful.
To answer to this question I have to make several points, which all depends form the fact that, first, success is not necessarily measurable using those parameters (number of people saved, number of organizations using your information), second, that the effects of crowdsourcing are not necessarily related to the humanitarian community and to the specific one-to-one relationship created by need-response.
First. When there is an emergency the most romantic catalyzing factor is the idea to be able to help. We badly want to save lives: looking at the images on TV about poor children starving or looking around with scared eyes is what moves us to donate money, to engage in all sort of philanthropic activities, to participate in fund raising, and of course, in wanting to participate in projects like PakReport. We badly want to help. This is all very good, but this is not the all picture.
The advantages of having information about what is going on in the field are not necessarily strictly related to the short-term issue “I give you food and so you live”. In an emergency setting, having information available means to give to stakeholders the possibility to understand the big picture, what is going on where, what are the most affected areas, what are the urgent needs, and in this way being able to design meaningful projects. This entire dynamic is not straight-forward and it is not depending on one information being able to activate one response, it is about the all system being able to give the big picture. This outcome is not measurable in terms of number of lives saved, but in effectiveness of response as a whole.
Also in this context, we may never be able to know who is using our platforms, because not necessarily the organizations using it are going to tell us, and not necessarily we can find one report that activate one response. Sometime is the all platform that activate a response, sometimes the response is activated by many factors and our platform is just one of them.
Second. Information in emergency is not something that only humanitarians need. They are not at all the only actors in scene: the affected communities are the main actors, and after them and before the humanitarians, are the people living in the country but not affected by the emergency. Looking at the Russian deployment of Ushahidi or the Snowmageddon one in Washington, one important fact emerged: before the humanitarians the people living in the country and not affected by the emergency are in the position to help, and they can do more effectively and faster than others.
In addition to that affected communities needs information, more than humanitarians again: they need to know where the IDP camps are for example, where the medical camps are, where they can go to have clothes, what are the roads blocked and if they can come back to their villages. The use of crowdsourcing in emergency settings is not at all related to “how many organization use my platform” but, what is the impact on the people in need, also in terms of how can we give to them useful information. I agree there is still the problem to understand how to get the information mapped on the Ushahidi platform to them, but the existence of the Alert system allows people with a mobile phone to receive those information immediately, even if they cannot see the platform or go on Internet. Is this going to save their life? Probably. Will I be able to prove it now? No. But after the emergency I can do a study and see if it worked or not, if people used it or not, how and when. This what the evaluation study about the Ushahidi Haiti Project is doing right now, for example.
Third. Crowdsourcing is not only about collecting information, is also about creating a circle and feeding information in the system. This means that we loose track of where the information goes from the platform. Let’s make an example. In the PakReport instance we are twitting the actionable reports. Now, those twitts get retwitted. And one of those twitt can get picked up by a newspaper and inserted into an article. Then someone else can write a blog about it, and someone else more will add the message to the one of the Facebook groups about the Foods. If any action is taken on that need, it would be almost impossible for me, as manager of the platform, to know where the information was taken from. The organization responding to it may have taken it from the Facebook page, or from Twitter, or from the platform directly, or from the newspaper article. The point here is: who cares? As soon as I keep alimenting the circle of information I am increasing the possibility that someone at one point will see it and act on it. This is the point.
People keep thinking that there is one main measurable effect of crowdsourcing project using Ushahidi and other tools in emergency settings. I argue that the effects are multiple, some of them are easy and measurable and some others require a very deep study that can only be done after the emergency, not during it. Of course, there is a need to have an evaluation system for those types of projects, but there is also a needs to stop oversimplifying the picture by thinking that it is useful only what give me numbers of people saved, or numbers of reports, or numbers of organizations using my project. This is not what this is all about. Crowdsourcing information in emergency settings is about a much broader landscape, and until we focus on counting, we will not be able to truly evaluate the effectiveness of a project.