Can intuition yield better decisions?
New research illuminates how managers can combine intuition and experience to make better decisions in moments of high uncertainty
One of the more interesting and yet least understood aspects of managing is intuition's role in decision-making, especially in situations of high uncertainty. With the world seemingly on an inevitable path toward algorithmic management based on the logic of analytics, big data, and artificial intelligence, it remains true that most leaders today make critical decisions through "fuzzy" cognitive processes. It is in this setting that intuition typically operates. Recent research published by Douglas C. West (King's College London), Oguz A. Acar (Cass), and Albert Caruana (Malta), provides valuable insight into this aspect of management performance.
Their study set out to understand how we make decisions not using data or analytics but through experience and intuition. They start by defining intuition as "a fast nonconscious thought process that leads to an outcome." The authors note that when they operate non-analytically, most people combine their intuition with cognitive processes called heuristics, a term coined in the 1950s by Nobel-prize-winning economist and cognitive psychologist Herbert Simon. He suggested that while people strive to make rational choices, human judgment is subject to mental and situational limitations that can prevent them from always making the most rational choice. As a result, people develop methods to make decisions in those situations. These mental shortcuts are heuristics.
With the above definition in mind, imagine a spectrum in which, at one end, we have a fully analytical manager who bases all decisions solely on rigorous logic and, whenever possible, data. Now imagine a manager at the other end of the spectrum who makes every decision based solely on impulsive reactions with no analysis of any kind. Most people exist somewhere between these two extremes, with decisions often based on explicit or implicit heuristics developed over time (or learned from others). While this state may seem sub-optimal to a purely logical decision-maker, one of the intriguing aspects of management heuristics is that studies have shown they often outperform purely analytical decision-making, especially by experienced managers who tend to use only a limited number of heuristic processes.
To understand the role that heuristics play in managerial decisions, the authors surveyed and interviewed managers working in creative fields such as filmmaking and publishing. The rationale for this selection is that managers in these businesses face higher than normal uncertainty and information deficits at critical decision points. It is likely, the authors thought, that leaders in these sectors would be most cognizant of their own heuristic preferences in their decision-making. Moreover, given that these sectors represent multi-trillion-dollar industries worldwide, the value and impact of these heuristics would be significant.
122 executives participated in the study, of whom 80% were male with a mean age of 49.3. Just over 65% were Directors (Account/Development/Digital/Commercial/Creative/Marketing/Sales) and 16% were CEOs. Once recruited, they were shown the following list of nine heuristic models (as well as a tenth option of no decision at all):
Default: the expected/traditional choice is made
Recognition: a choice is made based on a specific past experience ("the movie was a hit so let's make a sequel")
Fluency: a choice with an expected outcome that can most signal the decision as being right or wrong ("innovators should fail fast")
Take‐the‐best: choosing the option that seems most likely to yield the desired outcome, even while admitting that other options might also work
Satisficing: selecting the notion that should exceed all objectives and ignoring the other possibilities
Tallying: give points and pick the highest-rated option
Experience: let the senior person make the decision
Majority: the most popular choice wins
Equality: divide resources among all options and see what happens
Defer: make no choice at all
After receiving an explanation of each of the nine types, the participants were asked to think about the most recent innovation project for which they needed to make an early stage go/no-go decision. They were then asked to “state the extent to which they used the approach described in each item in their decision‐making” on a scale from 1-7, with 7 indicating the highest of use. Lastly, the participants were asked whether their use of a given technique made their final decision (a) better and/or (b) faster.
The authors calculated the commonality of each approach, and found that a few methods dominated the participants’ managerial decision-making:
At over 17% and 19%, the top two heuristics that receive scores greater than five are "take‐the‐best" and "tallying," respectively. These were followed by the "instinct" and "analytic" dimensions at just over 14% and just below 14%, respectively. The "default," "recognition," "fluency," "satisficing," "experience," "majority," "equality," and "hierarchy" heuristics ranged from 8% to 2%, while "defer" received no scores in the 6–7 range.
The study also presents the associations between types, which do exist. For example, there are strong correlations between "fluency" and "recognition" as well as between "experience and hierarchy." There are also connections between people who "lean analytical" to "tallying," "take‐the‐best," and "majority." Managers who lean to the intuitive end of the spectrum more closely correlate to "majority" and tend to avoid "take-the-best." Furthermore, as we would expect, analytically-leaning combinations yielded slower decision speeds, while intuitive-leaning combinations took decisions faster.
Figure 1: Correlations between heuristics and decision-making strategies matrix (Source: Authors)
With the methodological analysis complete, the authors examined all of the data collected to determine if any common approaches emerged within the participants. Five decision-making profiles emerged from this part of their analysis:
"Low engagement" (infrequent use of any heuristic models)
"Full engagement" (frequent use of all/multiple heuristic models)
Of the 122 respondents, only a few fell into the low engagement group, which meant most managers used a heuristic set to make decisions. Furthermore, the executives for the most part felt that their use of heuristic methods had not compromised the quality of their decisions. Quite the opposite was true, for as the authors note: "managers who belong to an ‘instinct–heuristic hybrid’ or a ‘heuristic only’ profile were able to generate decisions that were perceived as accurate as managers who used data analysis in their decision‐making" yet "they were able to do so at a significantly faster rate." In other words, the executives believed that their heuristic-centric decisions were both good and fast.
This is an important finding, note the researchers, because managers usually assume that “heuristics and instinct increase perceived decision-making speed, but that they do so at the expense of accuracy.” The findings from this study show that “in terms of the perceived accuracy‐speed trade‐off, this unfolds differently in the context of innovation screening decisions.” In this context specifically, “when heuristics and instinct–heuristic combinations are used concurrently in decision‐making, perceived speed is increased without compromising perceived accuracy.”
When considering the findings noted above, it is critical to recall the context of this study, which was decisions made in a setting of high uncertainty where analytical supports may be few or even non-existent. These kinds of settings are the ones in which the instinct/heuristic-dominant approaches seem to work best. As the authors note:
It is notoriously difficult to have reliable information on many important factors that would be needed for making an effective screening decision. The information required includes data on customer reactions, competitor moves, and potential problems that might arise throughout the innovation process. Moreover, unforeseeable uncertainties or "unknown unknowns" are likely to be extensive at the front‐end of innovation. This highly uncertain environment together with unreliable information may favor the effectiveness of "instinct" while diminishing that of an "analytic" process approach.
Reading through this paper brought to mind a new paper from marketing professor Robert Mislavsk (Hopkins) and Celia Gaertig (Chicago). In a series of experiments with over 7,000 participants, the researchers presented them with product forecast ranges in both numerical and verbal forms. They then asked the participants to predict a single outcome. Time and again, respondents made more accurate predictions by synthesizing the verbal forecast over the numerical ones. The authors hypothesize that this surprising result is because people use different heuristics when working with words and numbers:
Replicating past research, we find that people act as if they average numeric probabilities. However, we find that people act as if they "count" verbal probabilities, increasing the likelihood that their own forecast is more extreme than each of the advisors' forecasts.
To illustrate their conclusion the authors give the following example:
...imagine that you are purchasing a plane ticket for your next vacation and you check two websites, Kayak and Hopper, to see if they predict any future price changes. If both websites say that there is a 60% chance that prices will increase, your own forecast would be close to the average expert forecast (i.e., close to 60%). However, if both sites say that it is "likely" that prices will increase, you would act as if you are "counting" each prediction as a positive signal, becoming more confident in your own prediction and believing that a price increase is "very likely."
In this case, as with the findings of the first paper cited, the non-quantitative approach yielded a generally (if not specifically) better overall prediction of the future.
Both of these studies note that there is little research to date into how people combine various predictions and opinions from external sources when making decisions about future outcomes. As the first study highlights, in situations such as deciding among various innovation initiatives or new product development efforts to fund, managers rely mainly not on sophisticated algorithms or data sets but on heuristic models developed and refined over time. The good news is that these models can be just as effective (and sometimes faster) than analytically dominant models.
With the rise of AI systems designed to help managers replace their traditionally heuristic-based approaches, it remains to be seen how these two phenomena will be reconciled — especially at the fuzzy end of the management spectrum. Heuristic techniques are already used in computer programming, so could it be that computers develop their own kind of intuition? Will they make the analytical methods easier and faster to deploy, slowly replacing today's human-heuristic primacy? While these questions wait for an answer, managers can benefit from understanding their own heuristic models and preferences, which, as this research shows, can still be highly effective in settings where risk is abundant and data is scarce.
Douglas C. West, Oguz A. Acar, and Albert Caruana. Choosing among alternative new product development projects: The role of heuristics. Psychology and Marketing, 2020;37:1511–1524. https://doi.org/10.1002/mar.21397