Could this be applied to the broader context of a humanitarian response, where there are a large number of different actors, working with little coordination, and certainly no central controller directing the tasks? Are you saying that we shouldn’t bother, just pack up and go home?

Basically, yes. Unless we can get our act together in future, I can’t see a strong moral or practical reason for most of what currently passes for the humanitarian response to continue. There will always be a need for essential lifesaving activities, but if we as a community claim to be doing something more than that, then we need to either put up or shut up.

To use the example of a village sending an SMS to a NGO warehouse to make an inquiry about a distribution: I don’t think that any organization would be happy about opening direct communication between beneficiaries and their warehouse (I can imagine the warehouse being overwhelmed with direct requests for items).

This is exactly what I mean by putting up or shutting up. I agree that we’d be deluged by requests from beneficiaries if we did this, but if we don’t do it, we can’t claim to really be “empowering” them. I think we can have one thing or the other, but we can’t have both.

However I think that NGOs should be more open to SMS contact from communities… Organizations could have focal points to get the information from the warehouse (or appropriate sector), and reply to the community.

Agreed – I think that this is an area where sms-based communication could play a huge role in the future. You wouldn’t make it the backbone of your logistics, but it would certainly strengthen the usual organisational processes and the links to the community that are so weak at present.

The main example you cite here is the search for Steve Fossett. Of course, this was not a disaster (if only all disasters received this level of attention).

And that raises two more questions for me about distributed approaches. Firstly, scalability. This was a relatively simple, focused task, yet we see how difficult it was to manage and co-ordinate those efforts, the relatively weak delivery and the problems with ground truthing the results. Imagine if it was a large scale disaster with multiple factors at play – how well would it work?

Secondly, sustainability. As we’ve discussed, Mikel, there are always problems with keeping peoples’ interest once an event leaves the news cycle. Do we really want to set up systems where there’s no accountability for delivering results if we can’t rely on their consistent performance?

I should stress, I am not saying that distributed approaches have no place in the system; I think that they do, and I think that we’re getting there. I would worry, however, if we made such approaches part of our mission-critical work…

I think one of the reason some people spent so much time and felt eventually helpless is that they didn’t see the result of their effort.

I couldn’t agree more. One of the first lessons I learned in the field was that if the people that you are asking to contribute don’t get feedback quickly, they rapidly stop contributing – and you don’t get a second chance.

1. Training. The costs of managing training across a large number of people in different locations and organisations is high. The question is, does the cost bring in enough benefits to make it worth doing? I struggle to think of tasks that wouldn’t require fairly substantial training – any thoughts? – but if there are then this might have potential.

2. This interests me more, because it’s more about empowering individuals and communities. The biggest barriers here (for me, at least) are to do with the reporting itself. To some extent we already have this system, in emergency service numbers, so all that this would do is shift that system to another platform for extra redundancy in the immediate response.

In the recovery phase, I don’t think that this would make any difference. In a large scale disaster, nobody is going to provide assistance on the basis of an SMS, phone call, email or letter – any organisation with an ounce of professionalism should provide assistance on the basis of an on-site needs assessment*. Yes, SMS can provide an additional platform for such requests, but would it remove the need for that on-site activity?

I do think there are interesting possibilities for SMS reporting problems in the system – for example, if the trucks don’t turn up on the designated distribution date, the village can SMS the NGO warehouse to find out what’s going on – a delay, an accident, an error. That would put the beneficiaries in a much better position in terms of their right to information – but might overwhelm already harassed warehouse officials!

* but don’t get me started on needs assessment, since this is an area in which our record isn’t particularly good either.

In another case, the Katrina effort, there was a similar information impedance mismatch as I would call it. Nearly everybody in the U.S. had somebody they knew in the affected area and very little information was coming out in the first few days. The big question is, while information is available to first responders, how do we make this data available to the rest of the country ? This data has different meaning to different people and it is somehow unfair to the rest of the population to restrict what should be made available. This is one of the reason we looked into a cheap way to do maps with Geocam:
http://sei.tamu.edu/geocam/overview.htm

One of the other ways the mechanical turk effort should be improved is to give feedback to people about what has been done and what hasn’t. What has been confirmed to be bad and what hasn’t. I think one of the reason some people spent so much time and felt eventually helpless is that they didn’t see the result of their effort. If for no other reason, you spent 250 hours on the search you should have learned some type of skill. If you do not have a way to find out if your results have been acted on or declined then you might as well as have done something else because you learned nothing in the process.

Please also note that during the search for Fossett, some planes that had been lost for a long time were found. It is by no means a bad result. Some type of closure has been attained by some families:
http://nuit-blanche.blogspot.com/2007/09/search-and-rescue-new-directions.html

Another item is not mentioned in your analysis: while it is good to gather data and have competent people making sense of it, it is extremely difficult to share this data, not because of lack of capable ears, but rather because there is little effort spent in methods for sharing this data. In the effort for finding Jim Gray, it became obvious that even current sophisticated tools developed by the Coast Guards did not convey the full picture of what had been done and what had not to the people involved in the search.

http://nuit-blanche.blogspot.com/search/label/jim%20gray

Igor.

The main example you cite here is the search for Steve Fossett. Of course, this was not a disaster (if only all disasters received this level of attention). But the response is in the model of the resources we could expect in an actual disaster response. I’m not entirely sure why people were phoning and emailing — this should have been handled and managed within Mechanical Turk itself, and if people are over eager to report results, then the instructions and volunteer management should be more clear about what’s helpful and what’s not.

If some training is required, then I’d say it’s fair to put in place some virtual training before people are unleashed on the problem (without too many barriers of course). Assuming volunteers were effectively managed, and I think they could be, the question becomes is this pool of volunteers useful to the response? You seem to agree yes. So how can we structure Recovery 2.0 reponses?

The other aspect of the distributed response is requests for aid from people within the emergency. Can they SMS their needs and situation, to help guide responders? Definitely people directly affected (or their family/friends) should be able to communicate their needs .. so how can that be managed? I think this becomes even more vital beyond the initial response phase, and into the recovery phase, when the world’s attention has moved on.