Making Intelligence Smarter: Analysis of Competing Hypotheses
A portion of competitive intelligence work is simply gathering source information. However that information’s value depends in part on how well analysts can develop defensible conclusions from hypotheses that are often at odds with each other. Borrowing from public sector intelligence practice, the analysis of competing hypotheses (ACH) framework tests different interpretations by plotting hypotheses against evidence using a two-dimensional matrix.
Populating the Horizontal Axis: Hypotheses
The first step of the ACH approach is to gather hypotheses about a potential future event, business situation, or competitor move, and create a column in the matrix for each hypothesis. Potential entries for this horizontal axis could come from sources such as brainstorming sessions, blog-based commentary, social media, interviews with subject matter experts, and many other common intelligence sources.
There is often a hierarchical structure to the hypotheses. “Windows 8 will rapidly achieve large market share” could be a top-level hypothesis, with sub-hypotheses such as “small businesses will be slow to adopt Windows 8” and “Windows 8 will accelerate upgrades from Windows XP.”.
Populating the Vertical Axis: Evidence
Companies weighing their responses to events must consider multiple hypotheses. Following the example above, a software company deciding what resources to apply to supporting Windows 8 will want to have an understanding of the likely adoption of the new OS within its customer base. To support that understanding, pieces of evidence are populated along the horizontal axis of the matrix, so that the impact of each can be considered against the hypotheses.
Pieces of evidence to be used for ACH can come from any source used in competitive intelligence research, including open source intelligence (such as job postings or resumes from LinkedIn), first-party interviews, and industry events. General marketing materials such as feature lists or price points are also valuable.
Gleaning Insight from Intersection Points within the Matrix
Once the matrix is populated, analysis should proceed in a structured manner.. Each intersection in the matrix is first identified as “consistent,” inconsistent,” or “not applicable,” based on the support (or lack of support) by a specific piece of evidence for a given hypothesis. The next step is to tally a score for each hypothesis, based on the number of pieces of evidence that support it, minus the number that contradict it. Potential conclusions can then be ranked on that basis; note that, in some cases, a single piece of inconsistent evidence can disqualify a hypothesis from further consideration.
Using ACH can provide the ability to empirically establish the validity of multiple hypotheses, where there would otherwise be a tangled, contradictory mass of opinions.
By Sean Campbell
By Scott Swigart