The Six Parts of a Problem Statement

This post is also recorded as a video. You can listen to Michael here.

Scenario planning is a very common tool used by most companies in the world. Almost every large company has run one. And scenario planning is built on a physical limitation: you need one or two facilitators with about twenty people in a room. That is the physical constraint. Then there is the intellectual constraint, which is that the human mind can only comfortably hold two uncertainties in it at the same time. So you take an uncertainty on oil prices, an uncertainty on inflation, and you draw four quadrants.

But there are many more uncertainties. There always were. The four quadrants were never enough. The four quadrants were the number of things a human being could hold in their head while standing at a whiteboard.

Almost everything you were taught about corporate strategy has a constraint like that buried inside it. Nobody thinks about it because for a hundred years there was no reason to.

So the question Michael put to our team, and the question this video is built around, is, “What would strategy and corporate finance look like today if it was reinvented for a world that did not consider the limits of human physical capability and the limits of human intellectual capabilities?”

I am going to show you what came out of asking that question. The tool we created and why. The skill everyone is telling you to learn that we think is the wrong skill to focus on. The six parts of a problem statement that most people get wrong before they do a single hour of work. And the one question that changes a board meeting.

Why We Stopped Calling It Scenario Planning

In our system, it is not called scenario planning. It is called scenario living. There is a very specific reason we chose that word, and Michael will get to it in a follow-up episode, so I will not spoil it here. What I will say is that the change is not cosmetic. Once you remove the two-uncertainty ceiling, the thing you are doing stops being an exercise you plan to do in a room with 20 colleagues and clients and starts being something else.

That is what happens when you take a constraint out of a process that was designed around it. You do not get a faster version of the old tool. You get a different tool that needs a different name.

For those of you who have worked with us for many years, you have used our books and our training video programs and audio programs. There are literally thousands of hours of streaming episodes, and at one point we had thirty books out in the market. But when we wrote those books and prepared those videos, we had to write them for the audience, knowing the audience would not have the full capacity and capability and time and even budget to implement the most accurate and advanced version of what we do.

It is as if you go to a university program on aeronautical engineering. The textbook you use is not going to be the most advanced bible on aeronautical engineering. They are going to give you a simplified version of it, because that is what you can process at the time, given your skills and capabilities. It is the same with our books. They solve the problem accurately, and they are the entry level.

We have far more advanced thinking than what is in the videos and in the books. The freedom we have with Michael AI is that we can build that advanced thinking into the system, so that any user does not have to learn how we have done things. You do not have to spend months reading a book trying to figure out all the different steps. You do not have to watch all the videos on how to reconstruct a restructuring for a major pharmaceutical company. The system does it.

Prompt Engineering Is Not the Skill That Will Give You an Edge

Everywhere you look today, people are telling you that you have to master prompt engineering. You must know how to do prompt engineering. Everything must work according to prompt engineering.

Here is the problem with that. Another word for prompt engineering is defining the problem statement. And when Michael and I work with clients around the world, even today, the biggest problem is helping clients define the problem.

So when you ask someone to master prompt engineering, you are basically asking them to master the art of devising a problem statement. Most of us are never going to master that skill, because it takes a lot of time and effort to be great at developing a problem statement.

And it goes further than that. When you set up any system to solve a problem, first you have to look at the finances and assets that you have at your table. That is number one. Two, and this is the part that is woefully ignored by most AI tools, you have to consider the person who is going to take the recommendation and implement it, because depending on that person’s capabilities, skills and bandwidth, they may not be able to run with the recommendation given. Three, you do not know if you have the bare minimum to solve the problem, so most people overcompensate. They do analysis they do not need to do. They do not know how it links to the problem statement. They do not know if it solves the problem statement. They just do it. AI leads to a lot of unnecessary reporting.

Then there is the fourth one, competitors, and workwithmichael.ai has a module coming on that. The way we do competitor analysis is something you have never seen before, and it will make eminent sense when we show it to you.

The Ferrari

Think of your AI system as a Ferrari. You open the back, and the engine is there. That is the way Michael AI works. The AI is the engine, and the engine just produces the power. But all the thinking, all the decisions about what to do with that power, where to shuttle it, where to move it, that is algorithm and code that we wrote and that sits in the system.

In every other AI system, the AI does the thinking for you. For us, the AI is only the engine. We specifically have rules in place where the AI cannot do the thinking for you.

That is a design decision, and it took years and decades of methodology to be able to make it. Michael AI is coded to think the way we think. If you have worked with us, you know the way we think, and the system is coded to work that way. But what does that mean in practice? You can ask a general system to use our published prompts. You cannot replicate the algorithm running behind the answer, and more importantly, with a general system you do not know the assumptions, you do not know the weighting, and you do not know how all the data is coming together.

That last part matters more than people realize. If you take three inputs, are we weighting them equally? Are they correlated? Are they countercorrelated? That is the kind of thing running in the background that you never see, and it is the thing that determines the answer.

The Six Parts

Here is something you can use today, whether or not you ever touch our system.

As we have taught for many years, in case interviews and with executive clients like heads of banks and divisions and management consulting partners, the way you structure a problem statement is probably the biggest area of leverage, because most people structure their problem statements incorrectly.

A problem statement has parts. The problem. By when you need to solve it. The metric you are going to use. The target, because you must have a target, and you have to say what you want to achieve. The unit of measurement, because metric and units are different. And the baseline, meaning which year you are comparing against.

If you look at most people who structure a problem statement, they do not break it into these six parts. And there are four other elements our system is working with in the back that the user never sees, because they do not need to see them.

The majority of individuals in the world will have the wrong problem statement. They will spend days, and weeks, doing the research to identify the right problem statement. Michael AI does that in a matter of minutes, and we know it works, because this is the system we use ourselves. Everything we do runs through it, including the investment opportunities.

The Four Things That Decide What You Can Actually Do

In business, what you can do is a function of four things. One is your personal capability to see this through. Can you actually execute the plan? Two, if you know the laws, are there some completely legal and ethical loopholes your company could get through to find an opportunity no one else has seen? Three, what are your financials and assets like? And four, the big one, what are your competitors doing and going to do?

Notice that only one of those four is about the quality of the idea.

This is also why we built the system to refuse to give generic advice. If you ask a general AI system to prepare a strategy for Novartis, it is not going to take into consideration whether you have C-suite level experience, whether you are SVP level, whether you have no experience, whether you are an MBA student. It is going to give you a recommendation that you cannot implement for your level, for your skill, for your capability. That is a serious deficit.

Geography works the same way. If you are living in the Czech Republic and you work for Pfizer, you do not have a lot of say over what Pfizer can do, because they are based in the United States. Where you are based determines what control you can have in your company, and giving you recommendations for things you do not control is a really bad idea, because you can’t do anything with it.

There is a second layer to this. Regional materiality. If the person doing a study is based in Bulgaria, or Indonesia, or the Philippines, we would probably say the work is not material, because the bulk of Pfizer’s assets are not in those countries. Trying to fix an org design problem there in order to lift Pfizer’s overall valuation is not really going to matter.

And then there is an entire module of the system focused on how your actions will impact your career. It looks specifically at whether this piece of work or decision, even if you implement it perfectly, is going to be good for your career or bad for your career. No AI system does that, because they are not trained to focus on an individual, and they don’t have decades of experience of working with senior leaders, consulting partners and up and coming executives. They are consumer grade. They know a lot of things, but they don’t know anything in a lot of depth.

The Question

The system also produces a meeting guide. What you can lead with. What you can hold back. And the key question to ask.

In the Pfizer example, where the client’s idea was that the company should restructure its organogram to lower costs, this is the question that came back:

“If the redesign impacts every headcount and cost target on the current plan, what is the specific return on invested capital improvement you’re committing to the board, and how does that number change if the cuts are made in the wrong parts of the organization?”

It forces the client, or the investment target, or your own organization, to think about what they are actually trying to achieve, how they are going to do the cuts, and how they know the cuts they are going to make in headcount will lead to a good outcome.

Does fixing this area actually increase the value of the business? Because if it does not, why are you doing it? A lot of things you do in business look good on the surface, at least initially. You increase profits. You reduce time to market. You reduce shipping costs. But you may not have increased the value of the business. Those things don’t necessarily become valuable if you cut costs to increase revenue.

And when the answer is yes, our system does not stop there. It says, hold on a second, there’s an even bigger opportunity, and here is what that one looks like.

That is what senior partner grade thinking sounds like. Not just consulting grade but senior partner grade.

Final Thoughts

Michael AI is not a chatbot. There is no chat interface, because he is not there to chat with you and he is not going to make you feel better. He is there to help you solve the most complex business problems in a way that increases the value of a company and supports your career. We wanted to call him Michael Jr., but kids today want their own naming convention, so he picked AI.

There is more coming. But the question worth asking about your own work this week is not which AI tool to buy. It is this. Which part of how I work exists only because of a constraint that is already not relevant?

Take care,
Kris Safarova

P.S. The work of deciding which of your own processes are built on limits that no longer exist is exactly the kind of thing we do with a small number of leaders, alongside Michael, in our executive coaching. Deliberately tiny groups, a handful of senior leaders, about fifty minutes of focused work with each client on every call, with both Michael and me on the call. If you have been meaning to get serious about this, simply reply to this email. We also have more affordable coaching options via the MasterPlan Annual and Speak Without Limits.