Hypothesis-Driven Problem Solving
MAKE IT UP. THEN MAKE IT BETTER.
When I made the leap to management consulting from chemistry, I (correctly) assumed I’d encounter a lot of challenges. I did not anticipate that one of them would be adopting the hypothesis-driven style of problem solving common to the consulting industry. After all, I was a PhD scientist; who could be better trained in the art of forming and testing hypotheses than someone like me?!
Same Word, Different Meaning: Hypotheses in Science and Business
By “hypothesis,” both scientists and businesspeople mean a supposition or a guess, based on the information at hand that steers or focuses subsequent investigation. Beyond that, the practical use and expectations surrounding hypotheses differ in meaningful ways:
Explanation vs. solution. The subject of scientists’ hypotheses are explanations for the phenomena surrounding us - why does the natural world and those that exist within it work the way they do? In business, however, we are much more concerned with solutions - what does our business need to do in order to drive performance within our environment? These are different objectives: a business person crafts hypotheses that drive action, while a scientist seeks understanding.
Expected depth of prior knowledge. Both the scientific method and the consultants’ hypothesis-driven approach begin by defining the problem. A scientist follows that with exhaustive (I say: exhausting!) research into everything that is already known about the problem, while a business person often jumps straight to forming hypotheses. In fact, one of the main benefits of the consultant’s hypothesis is to narrow and streamline the research that needs to be done; the evidence required to support a business person’s hypothesis is, itself, hypothesized! It’s far from thorough - and may even feel intellectually bereft to, say, a well-trained scientist encountering the approach for the first time, - but it is indisputably efficient.
Proof vs. plausibility. While both the scientist and business person use hypotheses to guide and focus their work, the end-state of that work differs. A scientist seeks to prove definitively that their hypothesis is true or false. The bar for a consultant is far lower: we adhere to the 80:20 rule (you get the answer 80% right with 20% of the work) and merely aim to demonstrate that our hypothesis is a plausible and effective solution.
Pace. A scientist seeking proof and understanding will spend years – even an entire career - investigating a single hypothesis. A business person operates in weeks or months. What drives the difference isn’t impatience (though that particular personality quirk certainly underpinned my own transition out of science), but the pace at which the context of the problem evolves. The natural laws a scientist seeks to understand don’t change - they will be the same 100 years from now - so their work will continue to be relevant… unless they get scooped. Markets, customers and competitors, however, are always evolving. Someone tackling a business problem doesn’t have the luxury of time; move too slowly and their prior work becomes obsolete. We follow 80:20 because it strikes the balance we’re after: knowing enough to make good decisions, but stopping short of knowing so much that what you know is no longer relevant.
Engagement with others. Of all the differences I’m highlighting, this was, by far, the most difficult for me to accept. But it might be the most critical to get right in unlocking the full potential of the consultant’s approach to problem solving. A scientist refrains from sharing their work until it’s airtight. They’ve run the experiment multiple times, quantified the results down to the nth decimal point, anticipated every rebuttal and alternative. In contrast, a consultant will share the first draft of a hypothesis widely (but wisely!), seeking input deliberately from experts, clients, colleagues, and those who just think differently. In a way, this is the consultant’s streamlined approach to the exhaustive research done by the scientist before even forming their hypothesis: get as many smart people as possible to poke holes and continually improve the hypothesis as a result. We call this iteration and the pace at which it advances the thinking on a problem is unbeatable.
Why Hypothesis-Driven is so Valuable
I recall the exact moment of my full appreciation of the consultant’s hypothesis-driven approach. I was about 9 months into my tenure at McKinsey staffed on a large team at a major automotive client. We got word that the global client lead, a 20+ year tenured partner from Germany, would be making the rounds at our Detroit-based client and wanted to spend some time with the team. He would arrive in just a few hours and we needed to figure out what to talk to him about. My harried manager had a topic in mind and asked me to write a few bullet points on one Powerpoint slide for us to discuss before dashing off to whatever meeting would occupy him for the precious hours before the senior partner’s arrival. I don’t remember exactly what the topic was, but I remember staring at my computer screen blankly, clammy palms and rigid fingers hovering over the keyboard, brain swirling senselessly around a singular thought: “I know nothing.”
I managed to cobble a few ideas together, dutifully printed out multiple copies of my slide and sheepishly passed it around as the entire team - three peers, my manager, the lead partner, and the most senior partner I’d ever encountered - a firing line, really - convened to evaluate the output of my morning. Over the next 90 minutes, they eviscerated every last word I’d written. Nothing survived. The carnage was all over the whiteboard, pen scratches on every printed slide. And then, mercifully, it was over.
The shock and shame - how could I have been so wrong?!?!? - must have been evident on my face, because the senior partner pulled me aside.
“You wrote that, right?” I believe I managed a nod.
“That was the best team problem solving I’ve been a part of in quite some time. Thank you.”
Blink. Jaw agape. Stammering: “But… It was all… wrong!”
“Nonsense! There’s no way you could have known what was right, but you made sure that we discussed everything that mattered. That conversation could not have happened without your work.” And he was off to the next meeting he was already late for while I stood dumbly in the hallway.
Then the lightbulbs went on. It had never been about being right. Instead, it was about provoking a conversation that would speedily result in a better answer.
Viewing the session and my work to prepare for it in that light, it was a marvel really. It would have taken me more than a week to come up with that kind of answer on my own (and there were several elements that I probably would have never come to alone). By forcing a fast hypothesis and immediately putting it in front of a group of smart people, we had a client-ready deliverable in just hours.
An Application of “The Wisdom of Crowds”
This approach to solving problems takes advantage of the wisdom of crowds, an idea explored in James Surowiecki’s book of the same title. In the book, he draws the conclusion that large groups of people are smarter than an elite few. The crowd is better at solving problems, promoting innovation, making wise decisions, and predicting future events. In support, he highlights an old story from 1906, where 800 people competed to guess the weight of an ox at a fair. The results were surprising: the average guess of all 800 participants was actually closer than any individual guess (even closer than all the cattle experts). In fact, the average was only 1 pound away from the actual weight of 1,198 lbs.
James adds value to this old story by explaining the crowds must be wise. County fairgoers are likely to have some idea of what an ox weighs (in contrast to us city-folk; does an ox weigh 300 or 5,000 pounds?!?!). As detailed in Exhibit #1, James outlines five criteria mainly centered around making sure that the guesses are informed and truly independent. If these criteria are not met, we instead see a failure of crowd intelligence. History is littered with such failures, from the 2008 market crash, to the rise and fall of NFTs, to the famous Tulip Mania of the 1600s. James usually attributes such catastrophic failures to the damage that conforming opinion has when participants interact too much.
Exhibit #1. James’ Five Criteria for a wise crowd.
More recently, a TED talk described a theory of “Crowds within Crowds” being more intelligent. Researchers had 5,180 participants guess the answer to simple questions such as “how big is the Eiffel Tower?” or “how many people attended the World Cup?”. They found that when participants had briefly discussed in small teams of 5, the average of all the guesses was 49% more accurate than the average of the individual, isolated guesses.
My experience at the automotive client reflected all of these elements. The discussion was amongst a small group of diverse thinkers within a context that prized independence and trust. The business person’s hypothesis-driven approach is about tapping into this phenomenon to speedily develop meaningful solutions to difficult business problems.
Tips for Making the Most of a Hypothesis-Driven Approach
If you’re anything like me, embracing a hypothesis-driven approach to solving problems will involve time and discomfort. Some advice about how to get over the hump, and start enjoying the efficiency, productivity and quality benefits promised by this approach:
Make it up. Going from a blank piece of paper to something (anything!) is the hardest part. Lower the emotional burden by recognizing that something is always better than nothing. Even a wild guess is a step forward. And remind yourself: it’s not about being right, but about beginning a process that will get to right.
Be wrong… in the right way. While you are making it up, concern yourself more with being structured and complete - that is, identifying a good set of categories or elements that represent the full spectrum of the eventual answer - than with getting those elements correct. Furthermore, make your statements provocative: actually say something. When uncertain, we often feel compelled to fill the space with empty words, to be as bland and non-specific as possible, to protect our egos from allowing others to point out our errors. But getting others to identify the errors is exactly the purpose; give them something to react to.
Steal shamelessly. A scientist needs to create something new to the world; you do not. Look for (quick!) inspiration with a Google search or on Wikipedia… or a relevant blog! Great hypothesis-driven problem solving doesn’t care about authorship.
Make it better (aka: iterate, iterate, iterate). If getting to right doesn’t depend so much on where you start, but on how much you iterate, then think of that as your primary job. Drive the pace by getting your work in front of as many smart people as possible in rapid succession and incorporating the feedback thoughtfully after every interaction. “Thoughtfully” is a key word here; indiscriminately applying feedback creates an incoherent monster!
Know your “making it up” will get better! With experience, you will see patterns and develop intuition that help you start off closer to the bullseye even on topics on which you have no expertise. I once had a team that developed “Rule #1,” a phrase they would throw out during team problem solving sessions when we were at a crossroads and needed to choose between alternatives. It was sometimes followed by “Marja’s always right.” After almost 20 years of consulting, my hypotheses tend to be pretty good and my team could see it. Rule #1 helped us stop debating when we didn’t yet have the facts and rally around a hypothesis we could start iterating. Plus, I walked away from that project with a homemade t-shirt emblazoned with “Rule #1.”
What is your experience with hypothesis-driven problem solving? Do you worry about not knowing enough to “make it up”? What tips would you offer to help others embrace the approach? Share your thoughts in the comments and don’t forget to subscribe to the blog!
