Not too long ago I came across an article by Roger Martin, a very influential strategy writer, where he discussed the inherent problems with the commonly used SWOT analysis framework. You can guess his attitude to SWOT by the title of his publication – It’s Time to Toss SWOT Analysis into the Ashbin of Strategy History | by Roger Martin | Medium. He points out that the SWOT analysis, which stands for Strengths, Weaknesses, Opportunities, and Threats, is flawed due to its vague and inconsistent nature.
SWOT analysis encounters a major issue because the same factors can be perceived as strengths or weaknesses, opportunities or threats, depending on the context and objectives. This inconsistency makes the analysis essentially impossible to execute effectively. In his view, SWOT analysis is essentially an example of misusing data mining in a hope that if you torture the data long enough, you might get something. However, true analysis starts with an explicit hypothesis that guides the data collection, analysis, and standard of proof required. SWOT, on the other hand, is simply a useless collection of vast amounts of data without a clear hypothesis, that fails to deliver any real insight at best or supports biased, and hence flawed results at worst. For this reason, he suggests that SWOT should be discarded, as it offers little value and has become a time-consuming endeavor.
But let’s go back to Martin’s point. To address these shortcomings, Martin recommends conducting an analysis only when the specific purpose is well-defined, allowing for an in-depth exploration rather than a broad overview. Instead of conducting a SWOT analysis upfront, which often leads to vague conclusions, he recommends starting by defining the strategy problem that needs to be addressed. This involves identifying the gap between desired outcomes and current realities. Then, formulate a specific question using the “how might we” framework to guide the analysis process. This question should focus on closing the identified gap. Next, brainstorm potential choices that could answer the question and eliminate the gap. For each choice, evaluate the conditions that need to be met for it to lead to the desired outcome. Identify the least likely conditions or barriers that could hinder the choice’s success.
Only after this thorough process of understanding the question and the required level of proof, should specific tests be designed for critical elements of the chosen strategy. This approach promises more precise analyses and more compelling results. I would add that, at this point, it’s possible to use a slightly modified version of SWOT. Indeed, what is a strength and what is a weakness depends too much on the strategy you actually decide. Instead, I think it’s possible to drop these divisions and just look at internal and external environment, as well as intersections between the two.
Martin is slightly more extreme than myself and advises against conducting SWOT analyses at all. Instead, he emphasizes the importance of generating a hypothesis before engaging in any analysis to prevent aimless data mining.
To summarize, Roger Martin points out that the SWOT framework is of little value and should be eliminated as a time sink and useless practice. Although I tend to agree with his critique and place its value quite low, I still believe there are cases that SWOT could be used. That said, a specific hypothesis is required and separating internal and external environment into subcategories is a waste of time as those add little to no value.
What do you think? Are there any examples when SWOT provided you with some unexpected valuable insight?
