|
|
|||||
|
|||||
| Path: Main Street > Resources/Library > Research Articles > Feature Article |
Social networks and knowledge managementBy Gillian Kerr, RealWorld Systems
The information in this article is current as of April 6, 2002.
Knowledge management is a fairly new term that describes the efforts of organizations to use information more effectively. Here are some examples of problems that knowledge management tries to address:
- How can we prevent the loss of "organizational memory" when a staff person leaves?
- How can we orient and train new staff so that they can learn the important information more quickly, including where to find the information they don't know?
- What information would make us more productive or successful, and how can we track it and make it accessible? Then how do we convince staff to access it? How can we ensure that it's updated regularly and kept accurate?
- How can we be sure that our services are really helping our clients? Are we using the most effective service models? Are we managing our programs effectively?
- How can we create information systems that give us what we need, as inexpensively as possible?
Corporations and governments are spending millions of dollars on these issues, because the effective use of knowledge is absolutely central to almost everything we do. As usual, most of these initiatives have mixed success, but we're beginning to get some idea about what works and what doesn't work. (One popular web site on this topic is Brint.com).
Throughout my career in clinical psychology, public policy, allocations and information technology, I've been trying to understand one core issue: How can we use information to change behaviour? For example, how can we communicate the risks of AIDS in a way that will encourage people to engage in safe sex? How can funders improve the impact of what they fund? And how can we change public policy on the basis of solid published research? For years I've seen irrational behaviour from organizations and individuals who know, or should know, that what they are doing is ineffective or wrong. But somehow the information doesn't get through to them, or it doesn't influence them. I've been involved in doing evaluations that don't change program design, even when the results clearly show that change is needed. Why would organizations pay for evaluations that they don't intend on using?
As a result of these observations, RealWorld Systems has developed pragmatic approaches to knowledge management that include the following advice to our clients:
- Only kill what you can eat. In other words, only collect information that you are willing and able to use in your decision-making. Otherwise you're wasting time and money. Information is expensive to collect and to analyse, especially if it is of decent quality. Make sure it's worth it.
- Before you collect information, figure out exactly what you're going to do with it. It's good to begin with a thought experiment. Pretend that you have collected the information you want, then track it (mentally) through the organization. Who would summarize the results? Who would translate them into recommendations for action, and what kind of actions would they be? Would the recommendations go to the Board, or to senior staff? What would they do with it? What would be the most effective way of presenting the information? Would they believe the results? If the answer is that nothing would change, you're just saved yourself a lot of grief and saved your organization hundreds or thousands of dollars. Drop the project immediately. And make sure that your agency understands the reason (if it's politically possible to be that blunt), because an organization that is unable to respond to feedback is an organization that is incapable of learning.
- If possible, test the information pipes before turning on the full flow. Use a cheap and quick method to collect some relevant and significant information, and then see whether the organization can respond to it. For example, if a funder says that it wants to improve the impact of its funding, I often begin by asking staff to identify a couple of high performing agencies and a couple of poor performing agencies. They can all do it instantly. Then I ask them why they are not de-funding the poor performers and giving more money to the good performers. That immediately identifies key barriers in the information-to-action process that should be dealt with instead of collecting meaningless or politically unusable data. We can deal with the real issues (e.g. the funder needs to develop better skills in managing political pressures or improved record-keeping) before inflicting additional data collection demands on overworked agencies.
- Before programming a web site or intranet, develop a prototype (possibly in PowerPoint or Word) that allows the users to see exactly what they will get, screen by screen. That will enable them to clarify their real requirements before writing a single line of code. In some cases, clients can play around with a free or cheap web service that is close to what they want, so that they can experiment with how their organization would really use it. In some cases, they realize that they need an entirely different approach.
All of these tactics are useful in identifying the areas of greatest impact for information systems that will work in the context of the individual organization. However, we're now experimenting with specific ways of building more effective organizations through the analysis of social networks. (See this introduction to social network analysis.)
It's becoming clear that social networks are at the core of information flow. Most people learn through their social connections, both formal (at work) and informal (family and friends). We pay more attention to some people (called "hubs" or "authorities" in social network analysis) than others. Most of us are too busy to constantly scan the literature or question all of our activities, so we depend on credible sources to tell us what's important. For knowledge that is complex and requires judgement (like policy development or counseling), most of us learn by asking trusted sources for advice, not by reading articles. Even when we read articles, we pay more attention to the ones published in credible journals or written by authors that we respect. All of this has to do with trust, credibility, and our limited amount of attention.
This is obvious to most people. Of course we pay more attention to people we respect, and of course we "grade" information according to its source. So how can organizations take advantage of this natural human tendency to make itself more effective?
One of the most powerful answers is by identifying the experts in organizations or communities, and making it easier for their colleagues to take advantage of their expertise. Locating the experts might involve questionnaires asking 'Who do you go to when you need help on immigration issues?' or, 'Who do you call when you need advice on how to use computers?' It's possible to create a profile of experts and their networks, and then to encourage information flows that use these natural networks.
One of the fascinating elements of social network analysis is that the structure of the World Wide Web works exactly the same way. Like social networks, some web sites are hubs or authorities, in that other sites link to them because they have high credibility. Search engines like Google use mathematical algorithms to identify the most credible sites, based on the number of sites that link to them. We can use the same approaches to identify leaders in social networks.
In the human services, we talk a lot about the importance of social networks, communities, and information sharing. The new mathematical discipline of social network analysis may give us the tools to understand how to build and engage communities more effectively, as well as how to create information systems that actually change the behaviour of organizations.
************
Gillian Kerr, Ph.D., C.Psych.
President, RealWorld Systems
gkerr at realworldsystems.netRead my weblog at http://blog.realworldsystems.net
|
|||
