This article is the second installment in Orlando's series on virtual operations support teams (VOSTs). The first installment, Lessons Learned From The Social Media Tabletop Exercise, is available here.
Command & Control
Emergency response is fundamentally about solving a coordination problem — how to best coordinate groups of individuals to respond to a disaster? This is not unlike the coordination problem that businesses must solve in finding the best way to organize workers to produce a product or service, and the business experience says something about how emergency management solves the problem.
The coordination problem first cropped up in business with the Industrial Revolution. Prior to the Industrial Revolution, individual artisans made products start to finish. Afterwards, businesses needed to create a system that best integrated the efforts of different workers towards a common goal.
Business solved the problem by adopting the only large group coordination model of the time — the military. The military operates on a hierarchical command and control structure with decisions made from a single person at the top, which then filter down in steadily increasing layers until they reach the bottom. Business copied this model by organizing itself into layers of reports to supervisors and directions.
Emergency management also adopted this model, probably because most emergency managers come from business, government, or the military — all of which use this model.
But this is not the only way to coordinate group behavior. Markets, for instance, allow coordination to emerge from the individual actions of participants. Crowds walking down sidewalks in Manhattan at noon also coordinate their behavior without and command and control structure.
While the command and control structure has advantages, it also suffers from numerous shortcomings. First and foremost, it requires quite a bit of administrative overhead in terms of paid managers who are watching those below them. This overhead caps the number of participants that can be involved in the effort. Given the high cost of both paying the workers and all the managers above them, the number of workers has to be as small as possible and each worker must make an individually large contribution to the effort to be worth the cost. The big problem this creates is that it leaves on the table the contributions of large groups of people who can make individually small contributions that add up in the end.
For instance, I was born and raised in Wisconsin — at the top of the tornado belt. The local authorities would prepare for a storm by placing tornado spotters, normally police officers, at strategic locations around the area.
At the time, this was the best way to gather information about tornadoes. But there was no way to collect and act on the vast majority of observations. Thousands of citizens were on the road or at their homes or businesses during these storms and could identify a tornado if one crossed their path.
Think of the observations as following a classic power law distribution:
The number of tornado observations reported by spotters and the general public.
The horizontal axis represents individual people who are observing the weather, ordered according to the number of observations they are making. The vertical axis represents the number of observations per person. Those making the most observations are on the left, with those making the fewest observations on the right.
Notice how the curve is not linear. A few people on the left have a large number of observations. These are the professional tornado spotters. But there are a very large number of people who each have a small number of observations — the ordinary public such as those driving home from work or sitting near a window in their office. They are not watching for a tornado, but they will know one if they see it. This area in red represents the bulk of the power curve, sometimes called the “long tail.”
Because the city chose to manage the problem with the command and control institution of professional tornado spotters, it incurred high costs of hiring and managing that group, and thus could only afford a few. This means that they were only gathering observations at the far left of the graph. In other words, they were leaving the vast majority of observations out of the picture. With most observations in the “long tail” region, any particular tornado was most likely to be first observed by someone in this group, rather than a professional tornado spotter.
A second problem is that the command and control model is slow. Information needs to be passed up the line for decisions and orders need to be passed back down. The result is that the public often decides to respond to events itself rather than wait for official action. As an excellent study into the Queensland Floods reported, “The self-organizing community responses to such events ... bypass or leapfrog ... most organizational or administrative hurdles.”
A coordinated community response during the Queensland floods.
Self-Coordinated Disaster Response
It’s important to understand that the command and control model was created thousands of years ago, and partly in response to the lack of mass communication. Two thousand years ago a commander of 10,000 men could not communicate orders to everyone at once, or gather information from everyone at once. So the commander gave the orders verbally to ten sub-commanders, who then passed it to their sub-commanders, and so on. Information also travelled up from the bottom this way.
Social media has radically transformed communication. Now anyone can communicate to a group at almost no cost. More importantly, the group can communicate directly to each other at almost no cost. There are no longer the barriers to communication that drove the command and control model, and as a result new forms of coordination have emerged.
Facebook is at the center of much of this coordination. Much of the East Coast breathed a sigh of relief when Hurricane Irene missed major cities, but it caused heavy flooding in Vermont, and even cut off over a dozen cities from outside help. The Facebook page Vermont Floods 2011 went up almost immediately, and was used by citizens to self-coordinate disaster response.
For instance, people used it to volunteer to help others, and communicate where help was needed:
These self-organizing efforts create networks of people that are helping one another without any central organization, and they have proven remarkably effective in getting resources to where they are needed. Much like markets, the choices of individuals communicating with one another often succeed where central command cannot.
Social media has also produced an entirely new form of organization in the “structured network.” This is a network of people with a common goal that are given a structure in which to realize it, but are still making individual choices and contributions.
Google was one of the first businesses to adopt the structured network approach. It is well known that Google gives engineers one day a week to work on anything they want. They take this time to experiment and talk with others about their work. Google has a corporate policy that nobody is ever more than 100 feet from free food because they want to encourage the informal water-cooler discussions that lead to innovation. If an engineer has an interesting project others might chose to join in, and once the project looks like it has potential it is passed up to the top management who dedicates resources to it. Most of Google’s successful products have come from this model. This contrasts with the top-down model used by organizations such as General Motors, where the CEO comes up with an idea for a car and then assigns 500 engineers to build it. Guess which company is doing better.
This model is now being applied to emergency response. After Hurricane Katrina there was a great need for people to find information on friends and loved ones who had been displaced by the storm. A group of volunteer programs created “PeopleFinder,” a website with two simple buttons: “I’m Looking for Someone” and “I have Information about Someone.” This website was used to connect people in the affected areas with those looking for information about them. The website is built on a free, open source platform that can be used for any disaster. It eventually had over 55,000 entries.
As another good example, within a day of the Haitian Earthquake, a “Crisis Camp” was called in Washington, D.C. The camp had no set agenda; people showed up with computers in hand and simply mapped out ways they could help. Within a week over 40 other crisis camps appeared in cities around the world, and today the movement has been formalized into a response system for all new crises — The Crisis Commons.
Two of the systems created are particularly impressive. Within three days of the earthquake, an iPhone app was developed that translated Creole into English and back again. Then hundreds of volunteers hand-transcribed phrases into the system so that English speaking rescuers could speak with the victims to learn their needs or provide directions.
Crisis camps also created a software program called Open Street Map, which took the Google map of the affected area and, using post-disaster satellite images, allowed volunteers to code on the map every building that was damaged and how badly, which streets were blocked, where hospitals had been set up and other critical information for rescuers. Thousands of people hand-coded the information, which was converted into a format to display on GPS units. Rescuers on the street used the information and reported that it saved lives.
Open Street Map was created within days. Consider how long an organization like FEMA would have taken to create it. Organizations are simply not designed to move quickly, nor harness the contributions of thousands of people to aid in response.
Open Street Map allows volunteers to populate a Google Map with information important to residents and responders.
The goal of emergency managers is to transform the unstructured networks of Facebook, Twitter and the like into the structured networks of Open Street Map, PeopleFinder, and the crowdsourced disaster mapping systems. People want to help out in a disaster, and if they cannot fly to Haiti, they can still do an hour or two of computer work from home. Emergency managers need to learn how to provide a means for doing so by creating systems that structure contributions into an aid to response. Without this, emergency managers will find themselves sitting on the sidelines of citizen response to disasters.
Orlando will conduct a one-day pre-conference workshop on Creating & Running A Virtual Operations Support Team (VOST) at the 11th Annual Continuity Insights Management Conference, April 22-24, 2013 at the Sheraton San Diego Hotel & Marina. See http://www.cimanagementconference.com/pre-post-conference-workshops for more information.