Sad but True: Intelligence Failures & Bad Information Systems
“Those who do not learn from history are doomed to repeat it.”
Preface: on Monday, we wrote Human Error (versus Systemic Failure), which supplemented our longer post from Sunday: Intelligence Failures and Bad Information System Design. Much of that ‘Human Error’ post was devoted to mentioning that within organizations, most failures, including human failures, are systemic failures. You can’t blame it on your subordinates!
In the Sunday post, we hypothesized and speculated that bad information system design could be the cause of the recent intelligence failures. We based those suppositions on our knowledge of information systems, common design flaws, and the dysfunctional nature of the federal bureaucracy but with no real or specific knowledge of the circumstances. We don’t work for the government, and we’re too lazy and too busy to investigate on our own, but we figured our hunches were correct (and were willing to stake our meager reputation on them).
So, in the Monday post, we used L. Gordon Crovitz’s column, Intelligence Is a Terrible Thing to Waste, which appeared in that day’s edition of The Wall Street Journal, to provide some anecdotal evidence to support our conjectures of the overly-centralized and overly-rigid nature of the systems.
We closed Monday’s post with: “Sad, but true.”
Unfortunately, an article in Friday’s edition of The Wall Street Journal, Years of Spotty Data-Sharing on Suspects, provides additional evidence to support much of the criticism that we levied on Sunday (based upon our speculation).
We write “unfortunately,” because this is one of those cases where we hate to be right, but read it (the article) and weep. Here are several items mentioned in the article and our comments.
President Obama ordered agencies to bolster information technology.
- It’s unlikely that the failures are about technology or inadequate budgets. Note, using open-source web apps, our database-driven site and e-mail costs less than $150 per year to operate. It is a state-of-the-art publishing system that could be easily used by departments and agencies to post (and categorize) qualitative information and leads. Those categories could include substantiated versus unsubstantiated claims.
- More likely it’s about system design. We’re not under-estimating the volume of data for some agencies, but we are questioning the need to centralize its storage and management. More on this below.
A previous integration attempt, appropriately called the Information Integration Program, failed.
- Is anyone surprised by that result?
- We suspect it is overly-rigid and centralized.
- We also suspect that if such an integration attempt were to ever succeed, it would be immediately obsolete–most likely because some such agency upgraded one of its databases, and it is no longer integrable.
Supposedly, another integration attempt won’t be complete for two years.
- Remember: the last attempt failed. So, why believe the two-year deadline?
- It likely involves many industrious and very hard-working consultants spinning around on the little hamster wheels and sweating profusely, but with no real chance of success. It would be a Greek Tragedy if it weren’t an American one.
- There are needs for large systems, but we suspect far fewer than presumed.
- The issue isn’t how to accumulate all information and data, it is how to access information as efficiently as possible. So, why should a middleman aggregate it when individual agencies could publish it and searchers (with proper clearance) could immediately find it.
Emphasis on connecting e-mail systems
- Please see our post, Inexpensive but Valuable Web-based MIS, especially the section, ‘E-mail as the Central Nervous System.’ No need to repeat the argument here, but e-mail is an inefficient management information system. Better and inexpensive substitutes exist.
- Communication should be about be about publishing facts, speculations, and opinions, and letting others search those posts or reports (and/or receive feeds of future ones).
- E-mail is archaic for these purposes. We ask, dear reader: do you know any one of our several e-mail addresses? Unless or are a friend or acquaintance, no, you don’t. Yet you can read our current and past speculations and be automatically informed of future ones.
- Why shouldn’t intelligence analysts, within their own communities, have the same capacities that you, dear reader, have throughout the worldwide community that is the web? Provided you live in a free, uncensored society, you have the capability at little or no cost. You can search for items of interest and read and evaluate them based upon your knowledge and perspective. You can think we’re a fool or not, but you can make that assessment yourself for your particular problem or need. Why shouldn’t analysts be able to do the same on their intranet?
National Intelligence Library permits searches of finished reports
- That’s good, but it’s not enough.
- How much subjective and unsubstantiated and unverified data are eliminated from those finished reports? Again, that’s the stuff of new leads and threat identifications.
- How long does it take for such reports to be “finished” and available for general consumption?
- If agencies or work groups had their own (secure, intranet) publishing platforms, why bother consolidating? Let potential users, with the right clearances, surf. Another way to ask: why bother consolidating when the consolidator cannot necessarily anticipate the needs of users? Also, each blog on the web has its own system of permissions for access to private and password-protected information. Has anyone investigated whether a central clearinghouse is more efficient than maintaining access to data at local levels. We don’t have many subscribers, but we know when we have new ones, and can grant various levels of permissions to them.
Problems searching unprocessed information, especially clearances
- See Sunday or Monday’s post.
- Regarding who has access to which databases, security clearances are a major issue for a variety of good reasons, but distinctions should be made between data about citizens and foreigners, and there is no reason to endow foreigners with our rights; so, information about foreigners should be more easily accessed.
- Obviously necessary, clearly a constraint. In fact, by definition, they are constraints on sharing.
- We don’t have an answer to this issue, but we do have questions: Is clearance a status-symbol? Should lower level investigators and analysts have greater access? What are the costs and benefits of greater access? How could leaks compromise various investigations? Obviously, records of visits, queries, etc, can be kept (just like we have at our site and most other web publishers have).
Terrorist Identities Datamart Environment
- We ask: who, other than a government (or corporate) bureaucrat (or parasitic consultant), could like that name? Seriously? Is it that crucial to create a word from the acronym?
- Does the mangling of English imply anything about the construction of the system? We wonder.
- It’s clearly a centralized system, and based upon the Crovitz column we mentioned above, it seems very difficult to get on the list. We suspect it is harder to get off of the list.
What would we do?
For certain standardized monitoring and detection systems, there is a clear need for large databases. These are similar to record-keeping systems for transactions and events, i.e., not much different than, say, keeping track of checking account transactions or purchases and returns at WalMart. In a world-wide endeavor like terrorist detection and monitoring, such systems need to be search-able web applications (on a private intranet). That very much reduces the need for consolidation into one ginormous database.
In fact, the web is nothing if not one, large, searchable database (made of millions of small ones). However, the consolidation and aggregation is inherent and organic, rather than commanded or centrally-planned. In fact, modern sites are database-driven, and a visit to a page is the call to an (actual) database. Every time a Google search is performed, the web surfer is running a query, and has access to some sites but not other, password-protected ones.
Moreover, the search engines have developed algorithms to present the results in particular ways, and they are incredibly good at it. (At least on those searches where we rank high.) That is where time and effort should be devoted–not in attempting to physically consolidate disparate databases.
In that respect, let the disparities grow so that each agency can best serve its own mission, yet produce and publish intranet-accessible reports and notes.
We’d imagine that many of investigations are ad hoc and involve a bit of serendipity. We would imagine that with slightly different missions, the agencies have slightly different data and information requirements and emphases and traditions and cultures. So, why try to centrally consolidate (and therefore homogenize) the unique systems that may have evolved for specific and good reasons.
However, small, idiosyncratic systems that comprise a security intranet, can be index-able and search-able–just like the web.
So, we say harness the power of existing web applications and technology to protect our nation. Allow investigators and analysts be entrepreneurial publishers of their idiosyncratic views, facts, and suppositions. (All private and all secure on an intranet.)
Let investigators and analysts publish their reports and speculations for themselves and other agencies, join forums, and converse with their colleagues–even anonymously. (We reiterate: all published securely and privately on a huge intranet, of course.) Let them use their intellects and training to behave entrepreneurially, not bureaucratically.)
Use central resources to develop search algorithms and security clearance/permissions applications that operate seamlessly in a secure environment. Integrate intelligently, not by consolidation, by query. User management and permissions are immensely important, but millions of sites have solved such problems. With a bit of guidance and in time, we think the government can, too.
Information: it’s like the economy (and wealth) stupid. Try to centralize it, and you’ll kill it and destroy the incentives to produce more. In that respect, see The Wall Street Journal’s Review & Outlook, ‘A Failure to Connect the Dots’, for more corroborating evidence and perspective.
We’ll likely edit this post in the morning. (We did, and will likely do so again.)