2.
GENERATE MULTIPLE HYPOTHESES
“Watson, you can see everything. You fail, however, to reason from what you see. You are too timid in drawing your inferences.”
Sir Arthur Conan Doyle’s fictional detective Sherlock Holmes is famous for his methodical approach to solving criminal mysteries. Hypotheses play a central role in Holmes’ cases as they guide his investigations. In the Hound of the Baskervilles, Holmes is investigating how Sir Charles Baskerville died. Once he carefully observes the corpse, he develops two hypotheses: he was either attacked by a dog or died of a heart attack. Holmes collects data by scanning the surrounding environment. He mysteriously spends time alone in a cave and visits the nearby village. From the new evidence he develops a further hypothesis – that Stapleton killed Sir Charles to get hold of his fortune. To test this hypothesis he sets up a risky experiment, provoking Stapleton to release his hound to attack the younger Baskerville. You’ll have to read the book to find out what happens next.
Holmes formulated hypotheses by interpreting the facts but never accepted his first hypothesis as being the truth. He kept on revising his hypotheses in light of new available data, and was able to hold multiple perspectives without being overly attached to any one. In a way, Holmes demonstrated a case of “beginner’s mind” as he approached problems. He wasn’t bound by his previous knowledge or cases. He learned from every tiny detail, and remained open to the facts – even if they changed along the way.
Dr Thomas Bolte is a boyish-looking 51-year-old modern-day Sherlock Holmes. Based in New York, Bolte is a medical detective specializing in solving medical mysteries. He calls himself a “comprehensivist,” a diagnostician who steps beyond mainstream medicine to solve an evasive health issue.86 Others have described him as a “zebra hunter,” someone who looks for the zebra when he hears hooves, while everyone else is looking for the horse. He is renowned for diagnosing the undiagnosable, picking up cases that haven’t been solved by other doctors, with an astonishing 95% rate of success.
Bolte embodies Holmes’ ability to formulate and weigh different and competing explanations of the evidence. People comment about Bolte’s ability to look at every situation from a new perspective, looking for what others may have missed in unlikely places where nobody has previously looked, asking questions that haven’t been asked. In Bolte’s words “my life is so kooky crazy, nothing surprises me.”
In the business world, we often devalue diagnosis because we’re generally rewarded for leaping to action. However, when we are facing the unknown, a pre-existing solution orientation does not serve us well. Instead, we need to be deliberate in creating a space of possibility, a conscious assessment of what’s going on and what’s possible through a process of diagnosis – observing and collecting data and making a series of interpretations, which Holmes and Bolte are now famous for.
Intuit Inc, the creator of tax software Quicken, is a rare breed of company that champions leadership by experiment and decision-making by hypothesis.87 Scott Cook, the founder of Intuit, explains that instead of being guided by the opinion of the boss in decision-making, as is usually the case, the emphasis at Intuit is on getting people to make decisions based on their own hypotheses and experiments.88
The Intuit lean experiment loop starts with an idea, like Intuit India setting out to create new businesses that improve the financial lives of poor Indian farmers. Their vision was to raise farmers’ incomes by 10%. Once the idea was articulated, the team set out to find an important, unsolved customer problem. They immersed themselves in the lives of the farmers to gain a deep understanding of the issue, a process Intuit calls “deep customer empathy.”
One problem they discovered was that the farmers didn’t know which market to take their produce to in order to get the highest price. This seemed to be a good opportunity for Intuit, who thought that they could build a system where they could text farmers today’s best wholesale prices and which market agent was offering them. However, unlike many companies who would at this point go ahead and implement the “solution,” Intuit has added a few more steps to their process.
Next comes the “leap of faith,” where the team comes up with a series of hypotheses. In the Indian farmers’ example, the team’s hypotheses to be tested included:
• Enough market agents will share the prices with Intuit
• Market agents will honour the prices they give
• Typically illiterate farmers can read SMS
• Farmers will change behaviour based on SMS
• Farmers will perceive that they acquired a better price
• Intuit can monetize the opportunity
• Revenue will exceed cost89
Cook explains that seven weeks after discovering this opportunity and developing the hypothesis, the Indians started testing their hypotheses and running a series of experiments, including:
• pilot test of 15 farms
• data collection tests
• farmer acquisition tests
• push versus pull messaging tests
• alternate crop test
• price test
• advertising tests
• outsource sales test
The experiments drove the decision-making so that when they started delivering to farmers the prices at their local wholesale options on a simple mobile phone, they found out the method worked. After 13 more experiments, the results showed that farmers reported a 20% improvement in their farm income. Cook explains: “Now, for a poor farmer, for many of them, it’s the difference between their children going to school and not going to school. Yet this is a business that the bosses, including myself, would have said no to.”90
Sherlock Holmes always started with a dead body. Since he didn’t know the cause of the death, he generated a range of hypotheses about it, even ones that seemed unlikely. When new information came up, he applied it to the case and adjusted it as he went along. Hypothesizing prevented Holmes from jumping to conclusions too soon, which, as we’ve seen earlier (see part I), is what we tend to do, to bad effect, when we come to the edge of what we know.
The Intuit example shows how a company can adopt a Not Knowing approach when they enter a new market and test a new product. By setting up a business system that requires clear articulation of hypotheses, employees can move into the unknown with confidence because they have internalized the acceptance of the unknown. A culture has been built up that depoliticizes the process of forming views about the future (e.g. “this will be a good market”). In most organizations, people put forward a view and then spend their energy advocating it, to the detriment of the learning process. These are provisional ideas, best guesses, tentative explanations rather than fixed answers or solutions. When we feel we have to advocate for a position, we have a vested interest in proving it is right. The beauty of hypotheses is that no one needs to have a vested interest in the answer. Instead of becoming a champion for one possible explanation or model, the interest of the group is in collecting as many as possible and proving or disproving them. The focus is on discovery and revision, on considering all hypotheses plausible until new evidence is found that causes us to rule one of the hypotheses out.
However, many of us don’t work in an environment like this. Faced with tremendous pressure to deliver results, and fast, it may prove challenging to take the time to hold and play with multiple interpretations and perspectives on an issue. Providing alternative frames for looking at problems rather than solutions may frustrate, even infuriate, those who are expecting us to come up with a quick answer. It takes courage to be interpretative.