Criticism of the Intelligence Cycle

The Intelligence Cycle presents intelligence production as a sequence of distinct stages. A requirement is established, information is collected and processed, analysts produce an assessment, and the finished intelligence is distributed to the customer.

Its simplicity makes it useful for explaining the main functions involved. However, several intelligence practitioners and academics have argued that it provides a poor account of how those functions interact during real intelligence work.

In his book “Intelligence Analysis: A Target-Centric Approach”, Robert M. Clark described the cycle as:

“…an antisocial series of steps. It separates collectors from processors from analysts and too often results in “throwing information over the wall” to become the next person’s task”.

Former CIA officer Arthur Hulnick reached a similar conclusion:

“It is really not a very good description of the ways in which the intelligence process works.”

Mark Lowenthal also warned that the familiar CIA diagram “misrepresents some aspects and misses many others.”

These criticisms concentrate on three connected weaknesses: the cycle presents intelligence work as linear, separates teams that need to collaborate, and fails to take into consideration of how often requirements and assessments change.

In a literal interpretation of the intelligence cycle, each individual stage does not begin until the previous one has been completed. Collection appears to follow direction, processing follows collection, and dissemination follows processing.

In reality, intelligence work doesn’t proceed in such an orderly sequence. New reporting can expose a gap in the original requirement, raise further questions or undermine an early assessment. Processing may reveal a name, location or relationship that needs immediate investigation. The customer may alter the requirement as the situation develops.

These events can send the process back to an earlier task at any point. They may also require several tasks to take place at the same time. Analysts can examine existing information while collectors pursue new leads. Processing teams can prepare material as it arrives. Preliminary judgements can be shared while research continues.

During his testimony to the 2005 congressional hearing on “Using Open-Source Information Effectively”, former intelligence official Eliot Jardines described the linear cycle as “more a manifestation of the bureaucratic structure of the intelligence community” than an account of the intelligence exploitation process.

Essentially, he was arguing that while stages of the intelligence cycle may correspond with different teams or departments, these organisational divisions should not determine the order in which the work actually occurs.

Imagine a chef in a busy restaurant cooking one dish at a time. The chef waits until a dish has been prepared, cooked and served before beginning the next one. Service would be extremely slow. Customers at the same table would receive their meals at different times, while ingredients prepared earlier might go off before they are used.

A literal application of the Intelligence Cycle would produce similar delays. Collectors forced to wait for direction to be finalised before beginning their work. Processing teams remaining idle until collection formally ended. Analysts receiving a large quantity of information at once, with limited time to examine it. By the time the finished assessment reached the customer, events or priorities might have moved on, rendering it obsolete.

In reality, busy kitchens operate through concurrent activity. Several dishes are prepared at once, and work begins as soon as the necessary ingredients and equipment are available. Chefs continually tell one another what is ready, what is delayed and what they need next. If a customer changes an order, the kitchen adjusts while work on the remaining dishes continues.

An effective intelligence team operates in much the same way. Its members complete whatever work the available information allows and expand their findings as further reporting arrives. Each team understands what the others are doing and can redirect its effort when a new gap or priority emerges.

Clark argued that the Intelligence Cycle can encourage teams to “throw information over the wall” once they have completed their assigned stage. Responsibility for the material passes to the next team, while the people who collected or processed it have little further involvement.

This can deprive analysts of important context. For example, a collector may know why a source used a particular phrase, how the information was obtained and whether further questioning is possible. They may also have noticed technical limitations or inconsistencies that are not apparent in the reporting itself.. If those details are absent from the handover, the analyst may assess the information without fully understanding its strengths and weaknesses.

Separation also restricts feedback. For example, analysis may reveal that a source is highly unreliable, but, without this feedback, a collector may continue to use this source. Poor communication can also lead analysts to submit formal requests for information when a brief conversation would have resolved the issue.

Regular contact reduces these problems. Analysts can explain which gaps have the greatest effect on their judgements. Collectors can advise what further information is obtainable and how reliable existing sources are. Processing teams can direct attention towards the material most likely to affect the assessment.

Collaboration also makes challenging theories and findings easier. Early assumptions can be tested by people with different expertise before they become embedded in the product. Contradictory information can be examined as soon as it appears, and duplicated effort can be identified before it consumes further time.

The drawbacks of sticking to a rigid process become particularly clear during fast-moving events. Customers may need an initial judgement within hours and further updates as the situation develops. Waiting for every possible source to be collected and processed could leave the customer without intelligence at the point when it is most useful.

An effective way to deal with this is to produce a series of assessments rather than waiting till the process is finished to deliver a single, comprehensive one:

Each product should reflect the best available understanding at the time of publication. New reporting may confirm an earlier judgement, increase or reduce confidence, or require the assessment to be revised. This continuing process depends on communication between the customer, collectors, processors and analysts.

None of the above should diminish the usefulness of the Intelligence Cycle nor the value of having specialist roles within the intelligence process.

The Intelligence Cycle is a fantastic model for teaching the intelligence process. It identifies the core tasks involved in producing intelligence and gives new practitioners a simple way to remember them. It is also useful for ensuring you have covered all bases when designing an intelligence capability or reviewing whether an intelligence tasking has been handled comprehensively.

It should be regarded only as a checklist of tasks that need to be completed to produce intelligence - and a loose steer on the order in which to do so - rather than a definitive blueprint on how to structure an agency’s workflow.

As for the value of specialist roles, collection, processing, analysis and dissemination all require different skills, tools, and forms of experience. However, the boundaries between those roles should never prevent the movement of questions, feedback and information.

A concept that provides a far more realistic overview of how intelligence works, and one that is far more suitable for structuring an agency’s workflow, check out Clark’s Target-Centric Approach. This places the intelligence target at the centre of a shared body of knowledge to which collectors, processors, analysts and customers can all contribute. Participants can see what is known, question the available information and identify gaps as the work develops.

This approach addresses the main criticisms of the Intelligence Cycle: intelligence production depends on repeated exchanges between people performing different functions. Any workflow that restricts those exchanges will slow the process and weaken the resulting assessment.