Monday, March 10, 2008

A Wiki Is Like A Room And Other Lessons Learned From 15 Wiki-based, Open Source, Intelligence Analysis Projects (Part 1 -- Introduction)

Imagine that your bosses have just assigned you to be the team leader for a large scale analytic project:

“There isn’t any extra money,” they say, “for travel and stuff.”

That doesn’t matter so much, though, as long as you have bright people, right?

“Well,” they say, “they’re very bright and you have lots of them…but they’re...young.”

No problem. That's an advantage. Young means energy and few preconceived notions. Maybe we can use some special databases or other info to fill in some of the gaps?

“No, sorry, no extra money. Didn’t we already say that? You can only use open sources and your people aren’t yours full time either; they have other jobs that they need to do. And by the way…”

Yes, you say.

“This project is incredibly important.”

Right. So what can you give me?

“We can give you a wiki.”

A wiki? What's that?

"Well, a wiki is like a room..."

In this latest foray into what I like to call "experimental scholarship" I intend to discuss the lessons learned from using a web-based collaborative tool, commonly referred to as a “wiki”, to create custom intelligence products for decisionmakers in national security, law enforcement and business. The data in this series of blog postings will be the same data I am using in my paper and presentation on lessons learned concerning wiki-based analysis at the upcoming International Studies Association Convention in San Francisco.

While I consider the conclusions in this series of posts tentative, almost exploratory, in nature, they are based on a considerable body of evidence. Over the last year, students in my classes at Mercyhurst College or students working for me on funded research projects through Mercyhurst’s Center For Intelligence Research, Analysis, and Training (CIRAT) have used wikis to produce 15 large scale estimative products for real world decisionmakers (or intelligence professionals who support real-world decisionmakers).

Covering topics such as “The Impact Of Chronic And Infectious Diseases On US National Interests” or “The Role Of Non-State Actors In Sub Saharan Africa”, collectively, these 15 projects have generated over 6000 pages of finished analysis and have involved 97 analysts (many who worked on multiple projects). These students spent, conservatively, 21,000 analyst-hours in total on these projects. In addition to insights gathered from observing and supervising the projects themselves, 63 of the analysts who worked on the projects also took surveys designed to get feedback regarding their reactions to producing wiki-based analytic projects.

The analysts who participated in these projects were mostly seniors and graduate students who have experienced working with conventional methods for developing analytic products either through their applied coursework in Mercyhurst’s Intelligence Studies programs or through jobs and internships within the business, law enforcement or national security intelligence communities. Despite their experience with traditional methods of producing intelligence analysis (or perhaps because of it), these analysts came to overwhelmingly prefer, for a variety of reasons I will discuss throughout this article, to use wikis to produce intelligence.

Quantity does not equal quality, however, which is why I also asked the decisionmakers who commissioned the analysis to take surveys designed to capture their reactions to and the relative strengths and weaknesses of wiki-formatted analysis. These decisionmakers, as I collectively refer to them, are all senior leaders in their respective fields or real-world intelligence analysts supporting senior leaders. In only one case, where the decisionmaker was an alumnus, was there more than a passing relationship to the college. They ran the gamut from elected representatives to very experienced intelligence professionals to Chief Executive Officers. While I will discuss the results of their surveys later in this article, all indicated that they were satisfied or very satisfied with the quality of the products they received and all indicated that they would be willing to receive products in a wiki format again (with the majority expressing an outright preference for the wiki format).

Tomorrow -- What Is A Wiki And Why Is It So Different


markcblair said...

The first casualty of all combat, whether in the boardroom or the sandbox, is the plan. Analytical products are beginning to adapt from products that contain an answer (noun) to a question, to products that provide informational frameworks that answer (verb) the questions, which spring from the chaos. The evolution of this type of fluid, adaptive intelligence requires lots of IT tools that manage information, crunch numbers/produce probabilities, and connect analysts to each other and the decision makers. The wiki format is the most basic foundation of this type of structure.

The wiki is becoming the Aristotelian premise of modern intelligence analysis. Plato wrote down the dialogues of Socrates, much of which were arguments over whether justice was better than injustice. Aristotle (Plato’s student) stated that if one did not accept that justice was better than injustice he had nothing to say or discuss with them. The wiki may not endure forever; however, it is highly likely that the functional role of the wiki has securely seated itself in the practice of present and future intelligence analysis.

1 said...

As a research analyst, I depend upon open-source information. Wiki, in addition to a Nexis account, and basic knowledge of boolean search terms can result in an accurate assessment. Obviously field research is vital too, depending on the subject. I look forward to reading your series of posts on Wikis.

Keep up the good work.

Kristan J. Wheaton said...

Jeff, Mark,

Thanks for the comments!

Mark, not sure if wikis are of quite the Aristotelian level of import but agree that they are clearly a useful tool in producing top quality intel analysis.

Jeff, using a wiki with analysis can really help get the value out of multiple sources. I hope to talk about this some in later posts.


Anonymous said...

So, where is part II? Certain opensource wiki's lend themselves to be more robust, tools than others. How they scale comes to be important, but possibly the most important factor is search capabilities. True you can link pages, however how do you solve the issue of redundant information, or information that is not properly categorized, or formatted. While wiki's are fantastic, they are also limited. How they are implemented should be carefully considered. I think it works well as a sort of repository for analyst notes, observations, summaries, etc. However to review more factual information, there are better products. One of which is already being used by the intel community. Unfortunately it is pretty expensive, and probably isn't being used to it's full capacity.

I've been looking into producing a more comprehensive data/intel tool (I'm a software engineer, veteran, and technology manager for a security related non-profit.) Contact me if you're interested in some form of collaboration ( Yes, this is a real email address--> ).