Skip to content

Structured Analytic Techniques for Intelligence Analysis

Book link

Types of analytic techniques

  • Expert judgement - An expert looks at the data and writes down his conclusion
  • Structured analysis - Analyst goes through a pre-defined process to rigorously reach conclusion from data
  • Quantitative methods, by generating data - Using Bayesian inference, modeling, simulation etc using expert-created data
  • Quantitative methods, given data - Using Bayesian inference, modeling, simulation etc on given data

Expert analysis doesn't work because humans are biased and have limited memory. That's why a structured method for analyzing intelligence is useful.

8 types of structured analytic techniques

  • decomposition and visualization
  • idea generation
  • scenarios and indicators
  • assessment of cause/effect
  • challenge analysis
  • conflict management
  • hypothesis generation and testing
  • decision support

Selecting analytic techniques

Common mistakes

  • choosing techniques that fit the evidence
  • avoiding techniques that take too long
  • only using the techniques you know

Core techniques

  • structured brainstorming -- systematically enumerate variables, stakeholders, hypotheses, etc
  • cross-impact matrix -- write down the relationship between each pair of variables
  • key assumptions check -- write down and question the assumptions you’re making
  • indicators -- write down things to monitor (things that affect our hypothesis)
  • analysis of competing hypotheses -- list all hypotheses, refute one hypothesis at a time, believe the hypothesis with the least evidence against it

Choose a technique based on your goal

Get started

  • Decomposition and visualization
  • Idea generation

Make sense of data

  • Decomposition and visualization
  • Idea generation

Explain the most likely outcome of a situation

  • Hypothesis generation and testing
  • Assess cause/effect
  • Challenge analysis

Monitor situation for changes

  • Scenarios and indicators
  • Challenge analysis

Generate and test hypotheses

  • Hypothesis generation and testing
  • Key assumptions check

Assess possibility of deception

  • Deception detection
  • Assess cause/effect

Decomposition and visualization

  • Mind maps
  • Network analysis
  • Tables

Checklists

  • Leverage checklists for quality assurance and to help you get started with an analytical task

Example: writing a report for a client

  • Who is the key person this report is for
  • Who else will read the report
  • What question does the customer want answered
  • How will the customer use this information
  • What format(s) will convey the information most effectively

We may not know how to answer a question, like "is Iran buying missiles from China?".

But we can ask related questions that generate leads we can follow to answer the original question.

  • Why would Iran buy missiles from China? Why not another country?
  • Is there a partnership between China and Iran?
  • What kinds of missiles would China sell Iran?

Create a timeline

  • Don’t assume earlier events caused later events!
  • Creating a timeline can help identify time periods we need to learn more about
  • Events that happen close together may be related

Ranking techniques

Say I wanted to rank leads A, B, C in order of priority. I could:

  • assign weights to different factors, like importance of the system, how recent the lead is, etc. then compute a weighted sum
  • have each person in a group rank the leads. and then combine the individual rankings somehow
  • compare each lead against every other lead

Gantt charts

  • useful for describing project schedules