What’s it about?

Bulletproof Problem Solving (2019) delves into one of the most important yet consistently neglected skills in the modern workplace: problem-solving. With routine jobs declining around the world, more and more employees are being tasked with tackling open-ended challenges. As we’ll see in these blinks, you don’t need an advanced degree in statistical analysis to be a great problem solver – you just need a dash of creativity and the right strategies. 

About the author

Robert McLean studied finance before joining New York RAND as an analyst. He later worked on strategic and organizational issues for McKinsey in the United States, Australia, and Asia. McLean also served as the dean and director of the Australian Graduate School of Management. His current focus is social enterprise, philanthropy, and conservation. 

Charles R. Conn is a Canadian-American CEO and conservationist. He previously headed up Oxford Sciences Innovation, a venture-capital firm that invests in science and technology businesses. Conn is also the former warden and global CEO of the Rhodes Trust, the organization responsible for administering the Rhodes Scholarship.


You need to define problems correctly first to find valuable solutions:

When you face a problem, you can start to think about how you will solve it immediately. You are eager to collect information, consult experts and analyze the results. Soon, with answers, you’re coming up. There is only one problem – a crucial step has been missed.

The solution of problems works only when the right questions are answered. Your work will be meaningless if you don’t do it. It could even be counterproductive, worse.

This is why the problem-solving process is so important to start by taking a moment to think about which question you want to answer carefully.

If a problem is not properly defined, the consequences can be disastrous. Take, for example, the journal industry.

Until the mid-1990s, local news was dominated by newspapers. Then a new competitor was born out of nowhere: the Internet.

At first, online publications such as blogs spoke to top managers in the industry, but the more they looked at the issue, the easier it was. Since new technologies such as radio and television have survived, why does the Internet have to be different? And no blog would match the content that large and experienced newsroom publishing teams produce.

Naturally, that’s not how it panned out things. How have they gone so wrong? Well, their problem had not been adequately defined.

There was no need to wrestle readers online platforms – they needed the people to place ads in newspapers. In other words, the managers thought about their content’s quality, but the main issue was the amount of advertising revenue. Hundreds of newspapers began to bust when advertisers moved online.

Who are the key decision-makers determining whether my solutions are adopted or ignored? What will success look like, and how will I know when I’ve achieved it? More importantly, how will key 7decision makers gauge whether my approach is working or failing? What’s my time frame? Do I need a solution by next month or in a decade? And finally, are any potential solutions off-limits? 

These questions won’t just help you define your problem more accurately – they’ll also prevent you from wasting your time coming up with great answers to the wrong questions!


Breaking problems into smaller pieces facilitates solving them:

A few years ago, writer Robert McLean began to consider installing solar panels in his home to reduce the carbon footprint. Living in Australia, going to the solar system seemed like nonsense.

But was it economically meaningful? It was more challenging to answer this question. Government grants were progressively phased out at that time for sustainable energy. But panel prices fell, and feed-in tariffs were to be taken into account – the price that electricity businesses buy from individual houses an excess of energy.

McLean needed a tool that could defuse this nude issue.

McLean has learned to tackle problems with logic trees at McKinsey, a leading management consulting firm. This tool works like this.

The first step is a hypothesis. In the case of McLean, the statement was “I should install solar panels.” It was the following. Then ask yourself what evidence this hypothesis would support.

McLean identified two criteria. Installation of solar panels was a good idea if it 7could reduce its carbon footprint by 10 percent and recover its investment in ten years. Criteria indicate the type of information you have to collect.

Let us, therefore, take this reduction of 10% of CO2. You must first know how much CO2 you emit to calculate a possible decrease in your carbon footprint. McLean examined, for simplicity, the extent to which the average Australian talks every year. The following is how much carbon he can avoid by switching to solar panels. He found his carbon footprint could be reduced by over 20 percent.

What about payoff? What about payoff? McLean increased the panel and installation cost. Then, he worked out how much he savings every year by using less external power and selling excess energy – a simple analysis by solar installed online calculators. The results indicated that in less than a decade, he could recover his initial investment. McLean had solved his problem with a bit of online research – he was supposed to install solar panels.

That’s how beautiful logic trees are. You will find out what kind of information can solve your problem by indicating your hypothesis and the criteria needed to support it.


The priority is to evaluate your influence and the possible impact of the solutions:

Atlantic salmon are not yet at risk, but the stocks of wild salmon are damaged by pollution, overfishing, and mismanagement.

For some time, the author Charles Conn was employed by a charity to prevent wild Pacific salmon from happening. For the North Pacific rainforest ecosystem, these fish are significant. They did better than their Atlantic peers, but they didn’t promise long-term forecasts.

The goal of charity was to increase the number of wild Pacific salmon, but it was possible to use its limited resources with so many possible solutions and uncertainty. Conn came in, that’s where. He dealt with one of the critical aspects of problem-solving: priority.

How do you increase the stocks of wild fish? A lot of answers are available. Ocean conditions could be improved, or habitat damaged restored. It could help to reduce fisheries quotas or tighten up sports fisheries regulations. The real question, however, is which is the most stringent strategy for you.

The best way to give priority to solutions is to look at the interaction between two factors: their impact scale and their ability to influence results.

Let us begin with low-influence, high-impact solutions. Improved ocean conditions would be suitable for salmon stock, but the efforts of many states and international organizations would be needed in a coordinated way. It is very effective but beyond your influence, in other words.

Low impact, low influence solutions are also available. A non-profit organization may only lobby politicians for decades if it does not grant licenses for sport fishing. But even if this is done, the evidence suggests that it does not boost wild salmon stocks in particular.

However, imagine the head, the Minister of State responsible for issuing fishing licenses, was also the charity’s head. You would have a great deal of impact now, but you could still look at a solution with a low blow.

This leads to high-impact solutions. This leads us to increased influence.

Pacific salmon are ocean dwellers and upstream in Alaska, British Columbia, and the Kamchatka Peninsula. Conn’s team was given an insight: go to the problem’s source and focus on improvements in the primary breeding rivers. The outcome? A manageable project limited to three or four streams that allow maximum efficiency for the limited resources of the charity.


Equalitarian work can help your team to overcome individual prejudices:

Problem-solving can be highly complex, but you don’t have to start with advanced statistics or sophisticated mathematical models. However, what you have to do is eradicate prejudices.

There are more than 100 common cognitive errors, according to the experts, that can be made by any one of us. Take confirmatory preference – this tends to concentrate on evidence that strengthens our existing convictions and ignore information that contradicts them. Then there’s the mistake of the sunk cost, which doubles losses because we don’t want to admit we’re wrong. This list continues.

So what are these pitfalls best avoided? Teamwork, in a nutshell.

The key message here is: equal treatment can help your team to overcome individual differences.

Take it from Supervision’s author Philip Tetlock. The book focuses on the art of predicting and making collaboration clear. Tetlock’s data indicate that the most talented people are always better organized by well-organized teams in anticipating future developments. In some cases, they can handle large quantities of raw data better than computers.

What does “good organization” mean here, however? To encourage a fair hearing in a fair and equitable atmosphere, the best teams optimize their problem-solving processes.

This idea is profoundly ingrained in the culture of the McKinsey consultancy, which has a policy called dissent.

This policy does not just encourage junior team members to voice discrepancies with senior officials; they have to make their differences known. In the meantime, superiors are committed to listening. Why does this matter so much? McKinsey believes that poor problems are often the result of a particular form of partiality: ideas do not value their merits but the status of the person who puts them forward. On the other hand, if everyone has a voice, the team is far more likely to act on the best ideas.

Attributing ten votes to team members represented by adhesive notes is a practical way to promote this openness and prevent senior team members from dominating discussions. Put each proposal on a whiteboard, and then put everybody in your working group next to your favorite idea. As an additional bonus, you can ensure that senior members vote last and do not influence anybody else’s vote.


Treat the information well, and valuable insights will reward you:

One thing is the collection of data; another thing is to use it to provide valuable solutions.

This is the workings of data. It can’t tell you anything alone, so crucial as data to solve problems – you have to speak it.

There are good and bad ways to do this, so the old statistics say that poor analysts torture the data until it tells them what they want to hear.

This approach will undoubtedly mislead you, but what is the alternative? It’s time for a heuristics conference.

“heurical” is a Greek heuristic, meaning “find” in ancient Greek.

A heuristic objective is to aid you in finding something – a solution that mesh with the data before you, in particular, as suggested by the word etymology. Let’s look at a few handy examples in greater detail.

Occam’s razor: first off. A French philosopher named William of Ockham developed this analytical tool in the 14th century. The solution is usually the right one. It says. Regardless of your problem, your best bet is to deal with the minor assumption.

Take an easy example of math. Tell us that you have four hypotheses which each have an 80% chance of being right. Execute the numbers, and you will see that all four are more than 40% likely to be correct. By contrast, it is only 64 percent if you make two assumptions. In other words, the lower you assume, the higher your chances are.

This is a heuristic addition: the rule of 80:20. The Italian economist Vilfredo Pareto of the 20th century developed it, which is why it is also known as the Analysis of Pareto. It says that 80% of results are often determined by 20% of causes. It is not unusual, for instance, to find that 20% of buyers of a product are 80% driving sales.

You will need to list your problems to perform a Pareto analysis – things like customer complaints or missed orders, or products that could be damaged. Next, calculate each issue on the basis that it makes a difference. Now that your problems are listed identify the root causes of your pain – things like training lack, equipment broken, or unclear processes. Finally, group the issues according to their roots and add the scores. The higher the total, the greater the effect or cause of solving the problem.


If you take the time to look, you can find lots of helpful information in the real world:

Take governments for example. Do tax cuts stimulate business? Ideally, to find out, you would run an experiment. How? Well, a particular income group could be chosen, its tax rates untouched, everyone else slashed, and you could see what’s happening. But such experiments in the real world are ethically questionable and, in many respects, completely illegal.

This is just one example of how data collection from an organization can be prevented. Budget restrictions have a similar effect in other contexts. But these obstacles are overcome.

The key message in this link is: If you take the time to look, you can find a lot of helpful information in the real world.

Take from the two political scientists Evan Soltas and David Broockman, who wanted to know if American voters were discriminating against candidates who had been elected as minorities. They could not create their experiment to answer this question and therefore turned to a natural experiment.

Natural experiments are worldwide experiments that generate the data after which you are accidental. In Soltas and Broockman, this was the Republican Party’s voting procedure during the Illinois state presidential primaries.

The voters chose delegates representing them instead of casting votes for candidates such as Trump or Romney. In the United Kingdom, it’s not uncommon, but Illinois has two quirks. First, on ballots, the names of these delegates, who are politically unknown and often ungoogleable. Secondly, voters need not vote for the entire slate of their preferred candidate – they can check the two Trump delegates, say.

It means that voters are familiar with the ethnicity of delegates; for instance, José will be Latino, while Tom and Dick are probably white. This also allows voters to choose between representatives on the same platforms. If voters discriminate, it is reasonable to believe that the number of votes cast for candidates such as Trump or Romney is lower than for representatives with names like Dick and Tom.

This is a great natural experiment since it provides the scientists with the information they need to answer their questions. All they need to do is screen the data – a task that requires considerably fewer resources than conducting their experiments.

The story’s moral? If you look long enough, you will probably find that data from somebody else can answer your question!