A Thorough Explanation of 'Hypothesis-Driven Thinking': The Art of Reaching Conclusions Fastest with Minimal Information

Hello! This is Ryota😊

[Those who think only after gathering information will never reach an answer] A thorough explanation of Kazuaki Uchida's 'Hypothesis-Driven Thinking': The art of reaching conclusions fastest with minimal information

"Let's gather a bit more data before making a decision."

"Let's talk to everyone involved before thinking about a conclusion."

"I don't have enough information, so I can't say anything yet."

In business, a cautious attitude toward gathering information is often praised.

Read as many documents as possible.

Research past cases.

Analyze every possibility.

Check the numbers in detail.

Wait until enough information is gathered to provide the correct answer.

At first glance, this seems like a very serious and rational way to work.

However, when you actually proceed this way, you may find that your research and analysis never end.

Gathering data on every product to investigate why sales have dropped.

Conducting surveys for every single customer to improve customer satisfaction.

Organizing all conceivable data—traffic volume, speed, OD, traffic lights, accidents, weather, roadside facilities, etc.—to investigate the cause of traffic congestion.

The more you research, the more new questions arise.

It would be better to check these numbers too.

It would be better to compare it with another period as well.

It would be better to look at other regions too.

Additional interviews are also necessary.

And then, right before the deadline, faced with a mountain of information,

"In the end, what can actually be said?"

you begin to wonder.

What fundamentally changes this way of working is Kazunari Uchida's masterpiece,

"Hypothesis-Driven Thinking: The BCG Approach to Problem Discovery and Solving"

.

This book is a 240-page business book published by Toyo Keizai Inc. in 2006. Starting with the introductory chapter, "What is Hypothesis-Driven Thinking?", it is structured around the flow of "Start with a Hypothesis," "Using Hypotheses," "Formulating Hypotheses," "Verifying Hypotheses," and "Enhancing Hypothesis-Driven Thinking Skills." The publisher introduces the book's central argument as: "It is not analytical ability, but hypotheses that determine the speed and quality of work." (Toyo Keizai STORE)

The author, Kazunari Uchida, joined the Boston Consulting Group after working at Japan Airlines, and served as the Japan representative and senior advisor. This book systematizes the methods for problem-solving within limited time and information, based on his approximately 20 years of consulting experience. (Toyo Keizai STORE)

In simple terms, hypothesis-driven thinking is

a way of thinking where you do not wait to gather all information before considering an answer, but instead set a tentative answer first and collect only the information necessary to verify that answer

.

Of course, this does not mean deciding on a conclusion based on a whim without any basis.

Create a tentative answer first.

Think about what facts should be found if that answer is correct.

Investigate the necessary data.

Correct it if it is wrong.

Verify it again.

By repeating this process, you increase the accuracy of the answer within a limited amount of time.

This time, we will explore the contents of "Hypothesis-Driven Thinking" by

Work.

Data analysis.

Document creation.

Meetings.

Problem solving.

Generative AI.

Job hunting.

I will explain thoroughly in a way that can be applied to these and more.

1. A hypothesis is a 'provisional answer'

When you hear the word hypothesis, you might feel it is something difficult used by researchers or scientists.

However, we unconsciously form hypotheses in our daily lives as well.

In the morning, the train is delayed.

'Maybe there was an accident.'

A friend hasn't replied.

'Maybe they are busy.'

The computer isn't working.

'Maybe it's a network issue.'

These are all provisional answers at the current moment.

Hypotheses in business are basically the same.

Sales are down.

Is the primary cause not a decrease in new customers, but rather the churn of existing customers?

Overtime is increasing.

Is the cause not an increase in workload, but rather rework due to waiting for approvals?

Bus ridership is declining.

Is the problem not the number of runs, but rather the lack of connections to major destinations?

It has not been proven yet.

However, having a tentative answer makes it clear what needs to be investigated.

2. People without hypotheses make gathering information their job

If you start an investigation without a hypothesis,

What information is necessary?

How much research is sufficient?

Which results are important?

You do not know.

Therefore, you end up gathering as much information as possible.

Organizing data from the past 10 years.

Comparing all products.

Interviewing all stakeholders.

Analyzing every conceivable angle.

The workload increases.

Materials also become thicker.

However, having a lot of information is not the same as getting closer to the answer.

The act of researching itself becomes the goal,

'What is this research intended to clarify?'

becomes obscured.

People who have a hypothesis think before they gather information.

What is the most likely cause of this problem?

If that hypothesis is correct, which numbers should be moving?

What facts, if found, would disprove the hypothesis?

Because the necessary information is determined, the scope of the research can be narrowed down.

3. Excellent people do not have a large amount of information, but rather 'gather while thinking'

When looking at people who are good at their jobs,

one feels that:

They possess a lot of industry knowledge.

They can process large amounts of data.

They have many examples in their heads.

Of course, knowledge and experience are important.

However, the real difference is not just the amount of information.

Every time they receive information,

they ask, 'Does this support my hypothesis?'

Is this not a fact that contradicts it?

Is there no need to consider an alternative hypothesis?

That is what I think.

People without a hypothesis receive 10 pieces of information as 10 separate pieces.

People with a hypothesis divide those 10 pieces of information into:

Information that supports the hypothesis.

Information that refutes the hypothesis.

Information that does not affect the judgment.

They categorize them accordingly.

It is not just about increasing information, but organizing it while considering its relationship to the answer.

That is hypothesis thinking.

4. 'Gathering facts first' is not necessarily neutral

Some people worry that if you form a hypothesis first,

your preconceptions will become stronger,

and you will end up gathering only information that is convenient for you.

That is why,

you might want to say,

'I will gather facts first without any preconceived notions.'

That is what they want to say.

However, it is difficult for humans to gather information in a completely neutral state.

Which data to look at.

Who to talk to.

Which time periods to compare.

Which questions to ask.

Those choices already contain some kind of idea.

In other words, even if you intend not to have a hypothesis, you may be acting according to an unconscious hypothesis.

If that is the case, it is better to clearly put your hypothesis into words.

What do I currently consider to be the cause?

What information would make me change my mind?

If you make it visible, you can have others provide counterarguments.

It is more dangerous to not recognize your unconscious assumptions as hypotheses than it is to have a hypothesis.

5. Hypotheses and jumping to conclusions are completely different

If you misunderstand hypothesis thinking,

it becomes 'a method of deciding on a conclusion first and then gathering data that fits that conclusion.'

This is not a hypothesis, but a conclusion jumped to.

People who jump to conclusions:

Only look at information that fits their own ideas.

Ignore data that contradicts them.

This is a biased approach.

Trying to defend one's own hypothesis.

People who use hypothesis thinking

think that it is A at the present moment.

However, if the fact B emerges, they discard A.

They revise it to a new hypothesis.

They maintain this attitude.

A hypothesis is not meant to prove that you are right.

It is meant to discover your mistakes as quickly as possible.

6. A good hypothesis changes actions and decisions

Not just any hypothesis will do.

When you reach an answer with a good hypothesis, your subsequent actions change.

For example, in a company where sales are declining,

suppose you think,

'Isn't it because the economy is bad?'

This holds up as a hypothesis.

However, after finding out that the economy is the cause, it is unclear what the company can do about it.

On the other hand,

'Isn't the decline in sales of the main product due to a drop in the renewal rate of existing customers, rather than a contraction of the entire market?'

With a hypothesis like this, the countermeasures change depending on the result.

If the cause is market contraction, consider entering a different market.

If the cause is existing customer churn, implement product improvements or ongoing support.

When formulating a hypothesis,

check whether your next action would change depending on whether the answer is A or B.

If the action remains the same in either case, the priority for investigating that hypothesis is not high.

7. A good hypothesis can be verified concretely

'If we work harder, won't sales increase?'

'Isn't the problem that customer satisfaction is low?'

These are vague.

It is unclear what needs to be checked to determine if they are correct.

To change them into verifiable hypotheses,

'If we increase the number of visits to new customers, won't the number of business negotiations increase within three months?'

'Isn't the main cause of the decline in customer satisfaction the response time to inquiries, rather than product quality?'

is how you should phrase them.

This makes the figures to check and the objects of comparison visible.

A good hypothesis includes:

Target.

Cause.

Result.

These are the components.

Conditions.

specifically includes one of the following.

The ideal state is one where the person reading the hypothesis thinks,

“In that case, I can verify this by looking at this data.”

The ideal state is one where the person reading the hypothesis thinks,

8. A good hypothesis has a reason.

The difference between a mere hunch and a hypothesis is whether you can explain the reason behind it.

“I think the number of users has decreased because the fares are high.”

Why do you think so?

There are many complaints about fares in the survey.

The number of users has decreased since the price hike.

The decrease in short-distance usage is significant.

It is more expensive than competing modes of transportation.

If you have such observations or past knowledge, it is worth verifying as a hypothesis.

A hypothesis does not need to be strictly proven from the start.

However,

the starting point of “why do you suspect that possibility?” is necessary.

9. Have multiple hypotheses, not just one.

9. Have multiple hypotheses, not just one.

If you fixate only on the first hypothesis that comes to mind, your perspective will become narrow.

As causes for a decline in sales,

The number of customers has decreased.

The average spend per customer has dropped.

Purchase frequency has declined.

The composition ratio of high-priced products has fallen.

There are multiple possibilities.

Even for the cause of traffic congestion,

Traffic demand has increased.

Right-turning vehicles are obstructing straight-moving traffic.

Signal timing does not match demand.

Capacity has decreased due to on-street parking and stopping.

There is a temporary impact from construction or events.

These are some things that can be considered.

Do not commit to just one from the start; create multiple highly probable hypotheses.

From among them, prioritize those that:

Seem to have a large impact.

Are easy to verify.

Lead to countermeasures.

Prioritize these.

Hypothesis thinking is not about blindly trusting your initial intuition.

It is about comparing multiple possibilities to narrow down to the most plausible one.

10. Think about 'what results will emerge' before conducting analysis

In many analyses, people aggregate data, look at graphs, and then search for patterns.

In hypothesis thinking, you predict the results before performing the analysis.

For example,

Suppose you form the hypothesis that 'an increase in right-turning vehicles is the primary cause of traffic congestion.'

If this is correct,

Traffic speed will decrease during times when there is more right-turning traffic.

Right-turn queues will extend into the straight-through lanes.

On days with less right-turning traffic, congestion will be lighter even with the same total traffic volume.

The impact will be smaller at intersections that have dedicated right-turn lanes.

You can predict results such as these.

Once you have thought this far, you know what data is necessary.

Conversely, if any result you get could be used to claim the hypothesis is correct, then it is not a valid verification.

Before analyzing,

What will happen if it is correct?

What will happen if it is incorrect?

These are the questions you must ask.

It is important to think about.

11. Setting a hypothesis first reduces the amount of analysis.

The better someone is at data analysis, the more they want to try various analyses.

By time of day.

By day of the week.

By age group.

By region.

By product.

By customer attribute.

As the number of items that can be analyzed increases, the workload also increases.

However, what is necessary for decision-making is not to discover every single characteristic.

It is to answer the hypothesis.

If you want to verify the impact of right-turning vehicles,

Right-turn traffic volume.

Queue length.

Travel speed.

Signal timing.

Prioritize these.

If you want to confirm whether the decrease in usage is due to changes in demand for hospital visits,

Number of users for the purpose of hospital visits.

Relocation of medical facilities.

Usage by time of day.

Boarding and alighting locations.

Look at.

It is more valuable to perform three analyses that influence a hypothesis than to perform 100 analyses without a purpose.

12. Confirm the direction before increasing precision too much

In analysis, it is important to make the numbers accurate.

However, it is not always necessary to calculate precisely to the decimal point from the beginning.

Mr. Uchida explains that the purpose of using statistics in consulting is not to make the numbers themselves extremely accurate like in academic research, but to assist management in decision-making. He also shares an experience from his younger days when he was precisely plotting points on a scatter diagram, and a senior colleague pointed out that 'it is enough to know whether there is a correlation or not.' (Diamond Online)

What is needed first is to:

Check if this direction seems relevant.

Check if this cause seems significant.

Check if this measure has potential.

Confirm these things.

It is meaningless to increase precision when the direction is wrong.

First, make a rough guess.

After narrowing it down to a promising hypothesis, analyze it to the necessary level of precision.

This is the efficient order.

13. If the hypothesis is wrong, is the work a failure?

Sometimes you verify a hypothesis and find out it is wrong.

'I thought the cause was existing customer churn, but the churn rate had not changed.'

At this moment,

the hypothesis was off.

I did a useless analysis.

I failed.

You might feel this way.

However, having a hypothesis disproven is also a result.

If you know it is not existing customer churn, you can move on to the next possibility.

The problem is not that a hypothesis is off.

It is continuing to believe it without verifying it.

It is hiding the evidence that it was wrong.

It is learning nothing and clinging to the same hypothesis.

In good hypothesis thinking,

you formulate.

you verify.

you discard.

you revise.

You repeat these movements quickly.

14. The faster you can discard a hypothesis, the stronger you are

You develop an attachment to the hypotheses you create yourself.

You spent time thinking about it.

You explained it to your boss.

You even wrote it in your documents.

Because of this, it becomes difficult to admit you were wrong.

To protect your hypothesis, you start making excuses for data that contradicts it.

This time, there were special circumstances.

The sample size is too small.

The field team didn't implement it correctly.

Of course, there are times when there really are special circumstances.

However, you must check whether protecting your hypothesis has become the goal itself.

A hypothesis is a tool.

It is not your personality.

Even if a hypothesis is wrong, it does not mean your value has been denied.

Those who notice their mistakes quickly and move on to the next step get closer to the correct answer faster.

15. People who fail are too obsessed with having the 'correct hypothesis'

When you learn hypothesis thinking, you try to create an excellent hypothesis from the very beginning.

A sharp hypothesis that no one else has thought of.

A hypothesis that is almost certain to be correct.

A hypothesis with perfect logic.

However, you do not need to hit the correct answer from the start.

The value of a hypothesis lies

not only in being correct from the beginning,

but also in

moving investigations and discussions forward.

It lies in that.

Even if your initial hypothesis is rough, its accuracy will improve as you verify it.

Rather than worrying for a week without being able to form a hypothesis, it is better to create a current answer in one hour and verify it the next day to make progress.

A hypothesis is not a finished product.

It is a draft intended to be updated.

16. Even when conducting interviews, ask questions with a hypothesis in mind.

When asking someone for their input,

'Is there anything you are having trouble with?'

'Please feel free to share your opinions.'

If you only ask questions like these, you are likely to get generic answers.

The other party may not necessarily have fully verbalized their own problems.

If you have a hypothesis, your questions will become more specific.

'Is the reason you stopped using the morning flight because the schedule no longer aligns with your working hours?'

Is the time spent waiting for approval more burdensome than the input work itself?

Are you feeling more anxious about post-implementation support than the price?

If your hypothesis is correct, you can draw out more detailed information.

If it is wrong,

they will correct you by saying, 'That's not it; the real problem is...'

and you can adjust accordingly.

However, you must be careful not to lead the witness.

Maintain an attitude of, 'This is what I think, but there is a possibility I am wrong,'

and change your hypothesis to match what the other person says.

17. Change meetings from 'places to voice opinions' to 'places to verify hypotheses'

In meetings where the purpose is vague, participants just talk about whatever opinions come to mind.

I think it is A.

Based on my experience, isn't it B?

In the field, there is talk of C.

The number of opinions increases, but nothing gets decided.

If you set a hypothesis before the meeting, you can organize the discussion.

'The main cause of the sales decline is a drop in average spend per customer rather than a drop in customer numbers.'

Regarding this hypothesis,

what data supports it?

what facts contradict it?

What should be confirmed additionally?

What should be done if the hypothesis is correct?

We will discuss these.

The role of meetings will shift from just speaking to advancing the answer.

18. Include a hypothesis when consulting with your supervisor

When a young employee asks their supervisor,

'What should I do?'

the supervisor has to think from scratch.

If you consult them with a hypothesis,

'I think Plan A is best. The reason is that the cost is low and it can be executed in a short time. However, I am concerned that the effect is limited.'

you convey this.

It does not matter if the hypothesis is wrong.

The supervisor can give specific advice such as,

'We should prioritize effect over cost.'

'You are overlooking that risk.'

'A comparison with Plan B is necessary.'

and so on.

A person who consults with a hypothesis is not dumping the answer on someone else.

They have thought about it themselves and are coming to borrow help for the weak parts of their thinking.

19. Create documents to prove your hypothesis

A common mistake in document creation is including everything you have researched.

Research overview.

Aggregation results.

Comparison with the past.

Interview results.

Reference cases.

The page count increases, but it is unclear what you want to convey.

If you have a hypothesis, the role of the document becomes clear.

Conclusion.

Key evidence supporting the conclusion.

Data showing the evidence.

Responses to opposing opinions.

Next steps to take.

You can include only what is necessary.

Hypothesis thinking not only speeds up analysis but also shortens documents.

It is not about reducing what you write.

It is about reducing content that is unnecessary for decision-making.

20. Hypothesis thinking and the Pyramid Principle are connected

In 'The Minto Pyramid Principle', you place the conclusion at the top and arrange supporting evidence and facts beneath it.

However, you cannot build a pyramid without a conclusion.

That is where hypothesis thinking becomes useful.

Set a tentative conclusion based on the current situation.

Think about why you can say that.

Confirm the necessary evidence.

Revise the conclusion based on the data.

Finally, place the verified conclusion at the top of the pyramid.

In other words,

Hypothesis thinking is a technique for creating conclusions.

The pyramid principle is a technique for structuring and communicating those conclusions.

Combining the two creates consistency from the thinking stage to the communication stage.

21. The difference between hypothesis thinking and 'Issue Driven'

In 'Issue Driven',

you identify,

'What is the question that truly needs an answer right now?'

In 'Hypothesis Thinking', for that question,

you consider,

'What is the most likely answer at this moment?'

For example,

There is a major theme of

how to increase sales.

First, narrow the issue down to

"Is the primary cause of the sales decline the number of customers or the average transaction value?"

.

Next,

"Isn't the primary cause a decrease in the purchase frequency of existing customers, rather than the number of customers?"

formulate a hypothesis such as.

After that, verify the necessary data.

Choosing the question and setting a tentative answer are different, but they are closely linked.

22. Relationship with 'Issue-Driven Thinking'

Kazunari Uchida has also written 'Issue-Driven Thinking' as a companion volume to 'Hypothesis-Driven Thinking'.

The publisher describes 'Issue-Driven Thinking' as a technique for finding the problems that truly need to be solved from among the piles of issues in companies and workplaces, and focusing on them.(Toyo Keizai STORE)

To summarize the two books,

Issue-Driven Thinking deals with what to solve.

Hypothesis-Driven Thinking deals with what the likely answer might be.

Verification deals with whether that answer is truly correct.

is what they handle.

Even if you formulate an excellent hypothesis for the wrong problem, you will not achieve results.

Even if you choose the right problem, if you continue to investigate without having any hypothesis, it will take time.

Both problem setting and hypothesis setting are necessary.

23. In the era of generative AI, people who have hypotheses become stronger.

If you use generative AI, you can obtain a large amount of information in a short time.

List possible causes.

Summarize data.

Propose analysis methods.

Create drafts of text or documents.

However, precisely because AI can produce many ideas,

which possibility to prioritize,

what should be verified,

and which data will change the conclusion,

must be decided by humans.

If you ask AI questions without having a hypothesis, you will get a large amount of general theory back.

If you ask questions while having a hypothesis,

you can use it to ask, 'What data is needed to confirm if this cause is correct?'

'I want you to list facts that might refute this hypothesis.'

'I want you to organize the judgment criteria for whether to choose Plan A or Plan B.'

and so on.

AI accelerates hypothesis verification.

However, deciding what to use as a hypothesis and which answer to take responsibility for is a human task.

24. You can also use hypothesis thinking in your job search.

When you keep getting rejected during your job search,

you might think:

I lack the necessary qualifications.

My age is a disadvantage.

It's impossible because I have no experience.

However, those are still just hypotheses.

The real cause might be:

The connection between your experience and the company you are applying to is not being communicated.

Your resume only lists your job duties.

The experience required by the job posting does not match your background.

You are unable to explain your reasons for applying during the interview.

These are also possibilities.

Break down your hypotheses and verify them in small steps.

Apply to different companies using the same resume.

Improve your resume.

Apply for job types where you can easily leverage your experience.

Get feedback from recruitment agents or people with hiring experience.

By comparing results, you can get closer to the cause.

"I have no market value"

Before jumping to such an overly broad conclusion, it is important to break it down into verifiable hypotheses.

25. Having hypotheses in daily life makes it harder to judge based solely on emotion

A friend hasn't replied.

"Maybe they dislike me."

I'm not being evaluated at work.

"My boss has no intention of evaluating me."

I'm not feeling well.

"Maybe it's a serious illness."

People sometimes believe a single hypothesis as if it were a fact.

However, there are other possibilities.

They are busy.

They forgot to reply.

The evaluation criteria haven't been communicated.

The results of the work are hard to see.

I have been suffering from a lack of sleep or fatigue.

If you have multiple hypotheses, you are less likely to have your emotions swayed by a single interpretation.

Of course, for issues related to health or safety, you should not rely solely on your own judgment and should consult the necessary professionals.

Hypothesis thinking is not a substitute for professional judgment.

26. The greatest enemy of hypothesis thinking is confirmation bias

People have a tendency to easily gather information that supports their own ideas.

I thought this product would sell.

Therefore, I prioritize the voices of favorable customers.

I thought this subordinate had low ability.

Therefore, I only remember the times they failed.

I thought this investment target would grow.

Therefore, I only look at the good news.

This is not hypothesis verification, but hypothesis defense.

As a countermeasure,

Consider what the reasons might be if the hypothesis is wrong.

Is there any data that supports the opposite conclusion?

Who could argue against my idea?

Think about these things.

Construct the strongest possible counter-argument to your own hypothesis and check if it still holds up.

By doing so, the accuracy of your hypothesis will improve.

27. In jobs where safety is critical, do not prioritize speed alone

Hypothesis thinking makes work faster.

However, faster is not always better.

Medicine.

Legal affairs.

Structural safety.

Personal information.

Large-scale investments.

Decisions involving human life or significant losses.

In fields like these, narrowing the scope of investigation too much based on a hypothesis carries the risk of leading to serious oversights.

While using hypotheses to design investigations,

Legally required verifications.

Safety standards.

Third-party audits.

Unforeseen risks.

It is necessary not to omit these.

Hypothesis thinking is not a technique for skipping necessary verifications.

It is a technique for concentrating time on important verificationsis.

28. It is natural for the hypotheses of inexperienced people to be off the mark

Knowledge and experience are useful for forming good hypotheses.

Having seen similar problems in the past.

Knowing the structure of the industry.

You understand the characteristics of the data.

The more experience you gain, the easier it becomes to create accurate hypotheses.

However, having little experience does not mean you should not form hypotheses.

In fact, forming and testing hypotheses increases your learning speed.

Write down your own predictions.

Look at the actual results.

Think about why you were wrong.

Reflect that in your next prediction.

Rather than just being taught the answer,

the gap between your prediction and reality

is what helps knowledge stick when you check it.

Hypothesis-driven thinking is not a technique only for the experienced.

It is also a technique for accumulating experience quickly.

29. Three habits to improve your hypothesis-driven thinking skills

1. Write down your own answer before looking it up

Search for it.

Ask your boss.

Ask an AI.

Before doing that,

ask yourself, 'What do I think at this moment?'

Write it in one sentence.

It is fine if you are wrong.

By predicting before seeing the answer, you can identify the perspectives you were lacking.

② Predict the results

Before analysis or meetings,

think about what the results would be if your hypothesis were correct

as you consider it.

If you only create an explanation after seeing the results, you can interpret any outcome in a way that suits you.

③ Record the reasons why you were wrong

When the outcome differs from your prediction,

The hypothesis was poor.

Your background knowledge was insufficient.

There was a problem with how the data was collected.

The situation changed along the way.

record the reasons, such as these.

Failed hypotheses also become material for improving your next hypothesis.

30. Five points where this book is easily misunderstood

Misunderstanding 1: 'You should just decide on a conclusion based on intuition'

A hypothesis requires reasons such as observation, experience, and knowledge.

It does not mean adopting an idea without verifying it.

Misconception 2: 'Proving the initial hypothesis'

The goal is not to defend the hypothesis.

If it is wrong, discard it and revise it toward a better answer.

Misconception 3: 'You only need a little data'

It means reducing unnecessary data, not reducing the evidence needed for judgment.

Misconception 4: 'Reach a conclusion as fast as possible'

It requires not just speed, but also verification and revision.

At the stage of forming a hypothesis, it is not yet a fact.

Misconception 5: 'Only experienced people can use it'

While experienced people can easily form high-probability hypotheses, beginners can also grow by repeating the process of prediction and verification.

31. 'Hypothesis Thinking' 7-Day Program to Start Today

Day 1: Set a tentative answer for your current work

Regarding the problem you are currently facing,

'Isn't the cause X?'

Write it down in one sentence.

Day 2: Create two alternative hypotheses

Think of other possibilities so you do not get stuck on your initial idea.

Day 3: Narrow down the necessary data

For each hypothesis,

What should I check to confirm it is correct?

What would appear if it were to be disproven?

I will write.

Day 4: Verify on a Small Scale

Before starting a massive investigation, verify using a small amount of data, an interview with one person, or a portion of the work.

Day 5: Formulate Counterarguments

Think about what reasons might exist if your hypothesis were wrong.

Day 6: Revise the Hypothesis

Based on the information obtained, choose whether to keep, change, or discard the initial hypothesis.

Day 7: Summarize the Conclusion and Rationale in One Sentence

"Because of XX, we should do YY."

Organize your current conclusion in this format.

Summary

Kazunari Uchida's "Hypothesis-Driven Thinking" is,

not a book about making decisions on a whim even when information is insufficient.

It is also not a book that argues there is no need to collect large amounts of data.

What this book teaches is,

a way of working where you first have a tentative answer and then collect the information necessary to verify that answer

.

Many people gather information before they start thinking.

Read every document.

Interview everyone involved.

Aggregate all available data for analysis.

Think about the conclusion only after all the information is gathered.

However, if you haven't decided what you want to know, information gathering never ends.

Research one thing, and a new question arises.

Additional analysis increases.

Documents become thicker.

Yet, the answer remains invisible.

In hypothesis-driven thinking, you reverse the order.

First,

think about what the most likely answer is at this moment.

Next,

predict what facts should be found if that answer is correct.

Gather the necessary data.

If the results differ, discard the hypothesis.

Formulate a new hypothesis and verify it again.

This cycle allows for rapid decision-making.

It is the key to efficiency in business.

Through this cycle, you improve the quality of your answers while starting with minimal information.

A hypothesis does not need to be correct from the start.

It may sometimes be wrong.

In fact, what you learn when a hypothesis is wrong is what matters.

I found out that this is not the cause.

The likelihood of another possibility has increased.

I lacked this specific knowledge.

All of these become material to improve your next hypothesis.

A hypothesis is different from an assumption.

People who make assumptions only gather information that is convenient for them.

People who use hypothesis thinking also look for information that contradicts their own ideas.

People who make assumptions do not admit their mistakes.

People who use hypothesis thinking correct their mistakes quickly once they are identified.

A hypothesis is not meant to prove that you are right.

It is a tool for finding your mistakes as quickly as possible.

Also, you need to have multiple hypotheses, not just one.

Sales have dropped.

Has the number of customers decreased?

Has the unit price gone down?

Has the purchase frequency declined?

Create multiple possibilities, prioritize them, and investigate.

Before forming a hypothesis, you also need the right question.

Issue thinking is about deciding what to solve.

Hypothesis thinking is about creating a tentative answer to that question.

The Pyramid Principle is about organizing the evidence for your answer and communicating it to others.

These are not separate techniques, but are connected as a single problem-solving flow.

In the era of generative AI, the importance of hypothesis thinking is increasing even further.

AI can summarize vast amounts of information.

It can also generate numerous potential causes.

However, as the number of options increases,

you need the ability to decide which possibilities to prioritize,

what to verify,

and which results would cause you to change your judgment.

This ability is essential.

When someone without a hypothesis uses AI, the amount of information only increases.

When someone with a hypothesis uses AI, the speed at which they approach the answer increases.

People who work quickly are not just fast at the tasks themselves.

They are fast at deciding what to investigate.

They are fast at discarding unnecessary analysis.

They are fast at admitting mistakes and making corrections.

As a result, you reach the final answer faster.

It is not that you cannot think yet because you lack information.

Even when information is scarce, you think based on what is currently known.

You create a tentative answer.

And then, you act to verify that answer.

The answer will not simply appear one day just because you collect information.

It is because you have a hypothesis that you can find the facts that lead to the answer from within a vast amount of information.

People who wait for perfect information before acting cannot move forward unless all the information is gathered.

People who have a hypothesis can take a step forward even from an incomplete state and update their answer while taking action.

That is what I believe can be learned from 'Hypothesis Thinking',

a way of thinking that simultaneously increases the speed and quality of work with minimal information and limited time.

▽Finally, an earphone recommendation: the sound quality is incredible, it has noise cancellation, and it's under 15,000 yen!

▽X (https://x.gd/RUCL71)

Following me on social media would be encouraging!
I would be happy if you could leave a comment on this note with your thoughts on this article.