Introduction
Although many people hate putting their hand up, in an industry that over-indexes on opinions, we rarely fail for a lack of questions.
We fail, or at least miss the jackpot, because we're asking the wrong kind.
If we're honest, most questioning is confirmatory. It asks how to execute, communicate, or optimise a direction that has already been set. Far less questioning is subversive and intentionally critical, asking whether the direction is right in the first place, or whether it could be proved wrong.
This essay scratches that. So grab a coffee and strap in.
It starts with the default way people choose ideas, and then three weird things that push us towards the wrong questions. Things like the social cost of asking hard ones. The magnetism of whatever idea arrives first. And the habit of chasing familiar patterns.
Ultimately, we take asking better questions for granted, drifting through the day following orthodoxy or failing to slow down, stop, and think. Asking a radical question is harder than it sounds because the forces that suppress hard questions are vast.
Unfortunately, we're stuffed from the start.
Psychologist Raymond Nickerson defined confirmation bias as seeking or interpreting evidence in ways that favour existing beliefs, expectations, or a hypothesis already in hand. He argued that once people take a position, their motivation tends to shift toward defending it. If you’re honest, you probably do it, you’ve certainly been on the end of it. Most likely just this week.
However, it's probably more nuanced than the old confirmation-bias schtick.
Joshua Klayman and Young-Won Ha argue that when people test an idea, they mostly check the cases where they expect it to be true. They called this a positive test strategy. It sounds like confirmation bias, but there is a subtle difference. We intentionally look to test a hypothesis when we expect it to be true rather than those we think might fail.
Seriously, how many times has the best, most creative idea or solution on the table never even been considered for that reason? It's assumed it will fail because it challenges how we think about the problem. It's asking a different, better question outside of the original frame.
Lastly, and more depressingly, Hugo Mercier and Dan Sperber think something else might be going on in our monkey minds: "Skilled arguers are not after the truth, but after arguments supporting their views".
They contend that: “reasoning can lead to poor outcomes, not because humans are bad at it, but because they systematically strive for arguments that justify their beliefs or their actions.”
So, what’s going on?
Problem One: Hard questions carry a social cost
Asking a curly question can be fear-inducing because it can imply someone else's thinking might be wrong. Sometimes the thinking of the person who commissioned the thinking, the person who developed it, or is paying for it. In a client meeting, a pitch, or a boardroom, that makes the hard question a social act with risks attached. The reality is, too many people in senior roles don’t like to be challenged.
People need a safe environment to question
I’ve worked in creative agencies for over 20 years, and despite the label, only the very best of them have nailed a culture where anyone in the business doesn’t just feel comfortable asking better questions or sharing a random idea; they’re encouraged. Naked Communications was an exemplar of this.
More often than not, it’s the creative department that has that license. But god forbid you should challenge an idea. If it’s a strategy-dominant agency, the person who is knee-deep in data, theory, and writing long decks hates to be told the question they’ve spent the last two weeks answering might be wrong.
Then there’s the worst kind. The CEO’s way or the highway. Detached from the problem and the work, they seagull creative reviews (to fly in and shit everywhere), alleging to 'play the role of the client’. But what they’re really doing is telling people what the answer is. To a question of their own making.
Amy Edmondson's research gives a good account of this. She sees team ‘psychological safety’ as a shared belief that makes it OK for a team to take risks.
In her research, she found that psychological safety was associated with learning behaviour, while team efficiency was not. Learning behaviour mediated the link between safety and performance. Essentially, when people feel comfortable and are encouraged to take risks, they learn, it becomes a behaviour, and the organisation’s performance increases. Strange that, hey.
Asking whether a strategy or idea is wrong is exactly the kind of learning behaviour her work describes. It risks embarrassment, both for the person asking and for the person whose idea is being questioned. Where that risk feels high, the question goes unasked.
People self-censor themselves
Across multiple studies, James Detert and Amy Edmondson examined ‘implicit voice theories’. This basically means we have ‘taken for granted’ beliefs about when and why speaking up at work is risky or inappropriate. This leads to workplace silence, even holding back suggestions that could help an organisation.
So, when people are silent in meetings, they might well have something to say, but are too afraid to say it. That’s on the environment you created, not them.
Among the unwritten rules or reasons why include: don't embarrass the boss in public. Never, ever bypass the big dawg, and certainly don't speak up without solid data and a solution. The last rule is especially destructive when you need creative thinking. The most valuable questions are often the ones that can't be answered with data, yet. An unwritten rule requiring evidence before speaking filters the best stuff out.
As a leader, inviting a team to challenge you doesn’t work either if your reaction and behaviour aren’t genuinely welcoming when they do. It’s just lip service. I’ve certainly experienced more than my fair share of pitches that start with the most senior client saying exactly that after the briefing, only to look like you’ve pissed in their almond milk chai latte when you do.
Groups discuss what everyone already knows
Research also shows the social cost problem goes beyond the courage to speak up. In a classic study, Garold Stasser and William Titus found groups tend to discuss and act on information that all members share, at the expense of information only one member holds. When the unshared information points to a better decision, groups often miss it.
This bias is strongest in judgement-based tasks where the aim is actually consensus rather than finding the correct answer. Sadly, discussing information that’s already been shared can make others rate you more favourably, while raising unique information can count against you.
Quite often, the best question that changes the direction of an initiative sits with one person somewhere else: the salesperson who hears the real objections, or the analyst who has seen a worrying number. It is precisely these questions and people that are least likely to be asked.
Problem two: the first answer anchors the questions
The second problem is about timing. Once a draft strategy or idea exists, questions tend to orbit it. They refine it, rather than asking whether it should even exist at all.
Even random starting points skew judgements
We humans really are an odd bunch. In Amos Tversky and Daniel Kahneman's study into ‘judgement in uncertainty’ back in 1974, participants watched a wheel of fortune, rigged to stop at 10 or 65. They were then asked to estimate the percentage of African countries in the United Nations. Those who saw 10 gave an average estimate of 25 percent; those who saw 65 gave 45 percent. Their explanation? People start from the anchor and adjust away from it, but stop too soon.
If a random number can shift judgment by 20 points, a carefully argued strategy deck or briefing document is a far more powerful anchor. The first recommendation in the room becomes the reference point, and later questions stay only a few degrees from it rather than starting fresh.
Plans become forecasts
Kahneman and Dan Lovallo extended the idea to organisational decisions. They argued that decision makers tend to treat each problem as unique, neglecting what happened in similar situations in the past. This leads them to adopt an inside view, resulting in overly optimistic forecasts and predictions. Everything is anchored to the plan, even if the plan is wrong. (Kahneman and Lovallo, 1993).
Commitment makes the anchor heavier
In fairness, it takes a strong leader to backtrack once they’ve committed to something, but once resources and reputations are invested, another force can take hold. Barry Staw's 1976 study found that businesses not only fail to change course, but commitment to a plan can also escalate. Ultimately, a tendency, driven partly by the need to justify earlier choices, commitment to a course of action increases, even after negative results.
In simple terms, asking a question before a recommendation exists is better than asking one after.
Problem three: familiar patterns block better questions
The third issue is the one most rife in marketing. It's the tendency to reach for a familiar framework or a solution that worked before, instead of looking harder at the situation you uniquely face.
Familiar ideas block better ones
Merim Bilalić, Peter McLeod and Fernand Gobet studied expert chess players. They showed them a position that contained a familiar but longer route to checkmate, and a shorter, but less familiar one. Many found the familiar solution and missed the better one.
Eye-tracking showed that, while the players said they were looking for alternatives, their eyes kept returning to the squares linked to the solution they'd already thought of. When the familiar option was removed, players of similar ability easily found the shorter route.
This is pretty much the Einstellung, or ‘set effect’. The first idea triggered by familiar features of a problem directs attention towards information consistent with itself. The researchers suggested that the same mechanism caused scientists to overlook results that didn’t fit their favoured theories. This feels very endemic in marketing to me.
Even though experts genuinely believe they are searching for alternatives, they simply aren’t. Not really. A cascade of decisions is triggered from a familiar frame of reference, meaning lateral, more radical options get missed.
Expertise is a help, but also a hindrance.
Erik Dane’s research described this risk as cognitive entrenchment: as people acquire expertise, they lose flexibility in problem-solving, adaptation, and generating creative ideas. I can hand on heart say this is true at least for me. The older and more experienced I get, the more time I need to invest in learning new things outside the day-to-day; otherwise, I become a deadset dinosaur.
Frameworks can stand in for thinking.
I appreciate that things move faster than ever in our industry, but when it comes to strategy in particular, familiar patterns often arrive as frameworks and templates. Richard Rumelt has argued that template-style planning, with boxes for vision, mission, values, goals, and initiatives, can become a substitute for the hard work of diagnosing the actual challenge.
Among his hallmarks of bad strategy are failing to face the problem and fluff: abstract language that creates an illusion of high-level thinking.
Frameworks also shape questioning. If the template asks for a vision statement, the team asks questions about the vision statement. The questions the template doesn't contain, such as what the business will stop doing or who loses out, rarely get asked.
So what’s a humble human to do?
If awareness doesn't switch biases off, the remedy is to change the process. So, in summary;
Process matters more than analysis
The strongest business evidence comes from Dan Lovallo and Olivier Sibony's study of 1,048 major decisions, including new product investments, acquisitions, and large capital expenditures. They asked managers about the analysis behind each decision and about the process, such as whether major uncertainties were explored and whether viewpoints contradicting the senior leader's were discussed.
After controlling for industry, geography, and company size, they found that the quality of the process explained more of the variation in outcomes than the analysis did, by a factor of six. Moving from the bottom to the top quartile on decision process was associated with a 6.9 percentage point improvement in return on investment. They also observed that a strong process tends to expose weak analysis, while superb analysis is wasted if the process never gives it a fair hearing.
The data is self-reported by managers, so the finding is suggestive rather than conclusive. But it makes a pretty good case to say better questions come from better-designed processes, not from cleverer people.
Imagine the failure has already happened
Gary Klein's premortem asks a team to assume their plan has failed spectacularly, then generate the reasons why. He drew on 1989 research by Deborah Mitchell, Jay Russo and Nancy Pennington, who found that imagining an event has already occurred increased the ability to identify reasons for future outcomes by about 30 percent. Klein's stated aim was partly social: making it safe for people with reservations to voice them.
The premortem tackles two issues at once. It removes the social cost, because everyone is asked to find failure. And it breaks the anchor, because it starts from the opposite outcome.
Build in structured conflict
David Schweiger, William Sandberg and James Ragan compared three ways for groups to make strategic decisions: dialectical inquiry, where the group debates opposing plans; devil's advocacy, where one plan is systematically critiqued; and consensus. Both structured-conflict methods produced higher-quality recommendations and assumptions than consensus, and dialectical inquiry surfaced better assumptions than devil's advocacy. Let’s go with dialectical inquiry but find a better name. Devils playing advocates all the time tend to be a bit annoying.
But genuine dissent beats manufactured dissent
Charlan Nemeth's research adds an important qualification. In her experiments, a genuine dissenter outperformed every form of assigned devil's advocate in stimulating the quantity and quality of solutions. Role-played dissent tended to make people bolster their original view rather than think divergently.
This has a direct bearing on AI tools. An AI that plays devil's advocate is, by definition, role-played dissent. The evidence suggests its challenges may be easier to dismiss than a colleague's real objection.
Take the outside view
Kahneman and Lovallo's remedy for the inside view is to look for different reference points: how comparable plans, in comparable situations, actually turned out. McKinsey's strategy practice recommends calibrating plans against this outside view rather than relying solely on the detail of one's own situation.
If you got this far, thank you.
To finish, here are a few things to try that might just change the conditions in a team or organisation that help better questions get asked.
1. Make the hard question the question.
Start by making tough, destabilising questions part of the standard process. Rather than an individual act of courage, it lowers their social cost. The premortem works this way: everyone is asked to find a different possibility.
2. Question before the recommendation hardens.
Anchoring and the inside view mean the order of work matters. Assumptions and alternatives should be shared early, before a single draft becomes the reference point.
3. Ask what would prove the strategy wrong.
The positive test strategy means teams naturally look to prove why their thinking should succeed. Instead, we should deliberately point questions to where it might fail, and what else might be true.
4. Surface unshared knowledge.
The hidden profile research suggests the most valuable information often sits with one person. By collecting views independently before discussion, unshared knowledge enters the conversation before consensus forms.
5. Break familiar patterns deliberately.
The Einstellung research shows that experts sincerely believe they're searching for alternatives when they're not. Experts need to intentionally introduce structured alternatives, such as competing options, outside-view comparisons, and lateral analogies, and genuinely consider them.
6. Use AI for genuine dissent, not a substitute for it.
Nemeth's findings suggest that an AI agency playing devil's advocate risks the same weakness as any role-played dissenter: people may bolster their original view. So use it to genuinely raise objections to change your thinking, rather than surfacing issues you can overcome to get your thinking through. Everyone does it, don’t lie.
One broader lesson runs through all of this. The barriers to better questions are mostly structural and predictable, so remedies built into the process will usually beat encouragement and motivation. Telling a team to "challenge your thinking" asks individuals to overcome forces that the research shows are far stronger than those good intentions.
Sources
Bilalić, M., McLeod, P. and Gobet, F. The mechanism of the Einstellung (set) effect: a pervasive source of cognitive bias. Current Directions in Psychological Science.
Chapman, G. and Johnson, E. (2002). Incorporating the irrelevant: anchors in judgments of belief and value.
Dane, E. (2010). Reconsidering the trade-off between expertise and flexibility: a cognitive entrenchment perspective. Academy of Management Review, 35(4).
Detert, J. and Edmondson, A. (2011). Implicit voice theories: taken-for-granted rules of self-censorship at work. Academy of Management Journal, 54(3).
Edmondson, A. (1999). Psychological safety and learning behavior in work teams. Administrative Science Quarterly, 44(2).
Kahneman, D. and Lovallo, D. (1993). Timid choices and bold forecasts: a cognitive perspective on risk taking. Management Science, 39(1).
Klayman, J. and Ha, Y.-W. (1987). Confirmation, disconfirmation, and information in hypothesis testing. Psychological Review, 94(2).
Klein, G. (2007). Performing a project premortem. Harvard Business Review.
Lovallo, D. and Sibony, O. (2010). The case for behavioral strategy. McKinsey Quarterly.
McKinsey & Company. Hockey stick dreams, hairy back reality.
Nemeth, C., Brown, K. and Rogers, J. (2001). Devil's advocate versus authentic dissent: stimulating quantity and quality. European Journal of Social Psychology.
Nickerson, R. (1998). Confirmation bias: a ubiquitous phenomenon in many guises. Review of General Psychology, 2(2).
Rumelt, R. Good Strategy/Bad Strategy, via a review in Strategy & Leadership.
Schweiger, D., Sandberg, W. and Ragan, J. (1986). Group approaches for improving strategic decision making. Academy of Management Journal, 29(1).
Schweiger, D., Sandberg, W. and Rechner, P. Experiential effects of dialectical inquiry, devil's advocacy and consensus approaches to strategic decision making.
Sheridan, H. and Reingold, E. (2013). The mechanisms and boundary conditions of the Einstellung effect in chess. PLOS ONE.
Staw, B. (1976). Knee-deep in the big muddy. Organizational Behavior and Human Performance, 16; replication record.
Stasser, G. and Titus, W. (1985). Pooling of unshared information in group decision making, via a review of the hidden-profile research and overview.
Tversky, A. and Kahneman, D. (1974). Judgment under uncertainty: heuristics and biases, via overview of the anchoring effect.