POB-CRA: How to decide well

Comments are now enabled. Challenge assumptions, offer alternatives, or add your own perspective. Join the discussion. Click here to add your thoughts.

We worry about making the wrong decision. There is rarely enough information to make a decision with complete confidence.

The tempting response is to wait. Another report, proof of concept, or month of operational data. Perhaps then the correct path will become obvious?

However, not deciding is itself a decision.

While we wait, competitors move, projects stall, teams lose momentum, and opportunities disappear. The hidden cost of not deciding can be far greater than the cost of making an imperfect choice.

So how do we make good decisions when the information we need is incomplete?

The cost of not deciding

Everyone is busy. Deferring a decision can feel productive because it avoids commitment while giving us the impression of progress.

But every deferred decision implies a cost. Markets evolve, business priorities shift, and team members move their attention elsewhere. What originally looked like prudent caution can become expensive delay.

Your organisation may already be paying a price long before a decision is finally made.

What do we do?

Most significant technology decisions are made under uncertainty. Incomplete information is the norm.

We rarely have all the information we would like, and waiting often fails to remove that uncertainty completely, and delay becomes another variable to contend with.

What we need is a way to assess the risks, costs, trade-offs, and assumptions behind each option. A structured approach allows us to make the best decision available, with the information we have today, explaining the reasoning clearly, to build the momentum needed to move forward.

A decision is a trade-off

Technology decisions are not a search for the correct answer. Most of the time there is no such answer. A decision has to be appropriate giving what is known today.

A decision to build means not buying, optimising for organisational speed means accepting additional risk. A decision to standardise means sacrificing some local flexibility.

Every option creates benefits, costs, and risks.

Therefore, our goal is not to find the perfect decision. There isn't one. Our goal is to understand the situation clearly enough to choose well.

Many decisions require you to grapple with a large number of variables. The following framework helps you make sense of them.

  • What is the problem facing the business?
  • What are the available options?
  • What benefits does each option create?
  • What costs does each option introduce?
  • What risks are there and which ones are we accepting?
  • What assumptions must be true for the decision to succeed?

This framework is known as POB-CRA: problem, options, benefits, costs, risks, and assumptions.

Problem

In making any decision it is important to first understand clearly what problem the business faces.

Five-whys can help here but it has its limitations: it assumes a single chain of issues causing the problem; it can become subjective; and different people can reach different root causes.

For complex business and technology decisions, combining several approaches can lead to a clearer problem identification.

5 whys

The starting issue is:

Delivery is too slow.

Why?

Statement Underlying Cause. This is uncovered by asking Why?
Delivery is too slow. Testing takes too long.
Testing takes too long. Releases are manually coordinated.
Releases are manually coordinated. Teams do not trust deployment automation.
Teams do not trust deployment automation. Previous automation failures caused outages.
Previous automation failures caused outages. There is no safe rollback mechanism. The root cause.

The above highlights the linear nature of 5-whys. There may be multiple answers to why delivery is too slow, other than "Testing takes too long", e.g., work stalls in deployment. If this is the case, you can apply subsequent "Whys?" to each identified answer.

In the table, the original problem of delivery is too slow is now seen more accurately as teams lack confidence in deployment automation because recovery mechanisms are weak. This is a very different problem, and is a concrete statement that you can start to address. Problem statements tend to be much more abstract requiring attention to surface the many reasons behind them.

When posing initial problems (delivery is too slow) it is important to acknowledge that many factors may be involved, not least business process and the structure of the business. Adequate resourcing may also be an issue. Very rarely are the staff the cause: the staff work within a business culture and follow processes in an environment (e.g., regulated) that is resourced. The environment, resources, process, culture, and structure can be the cause of issues.

Other than 5-whys, there are many other ways to surface the underlying problems that your business faces:

Technique Approach
Problem Statement Testing Validate whether the problem is clearly defined, specific, evidence-based, and focused on the real issue rather than a proposed solution.
Symptoms vs Causes Analysis Distinguish visible symptoms from underlying root causes to ensure effort is directed at fixing the source of the problem.
Jobs-to-Be-Done Analysis Understand what users or stakeholders are trying to achieve and identify obstacles preventing successful outcomes.
Constraint Identification Surface limitations such as budget, time, technology, regulation, or skills that shape or create the problem.
Stakeholder Perspective Analysis Examine how different stakeholders experience, perceive, and are affected by the problem to uncover hidden needs and priorities.
First Principles Analysis Break the problem down into fundamental facts and assumptions, then rebuild understanding from the ground up.
Reframing Questions Restate the problem from alternative perspectives to reveal new insights, opportunities, and solution paths.

Using more than one approach lets you illuminate multiple underlying factors. The goal is to achieve clarity at the start. Solving a "problem" that the business does not have will be costly.

Options

Once you are crystal clear on the problem, ask yourself "what are the plausible ways to address this problem?".

It is recommended that you then consider across categories to help you reduce decision-making biases:

  • Solution fixation where you become attached to the first idea
  • Expert bias so you view every problem through your own discipline
  • Sunk cost bias that can lead to favouring existing investments
  • False dichotomy where you believe there are only two choices

Eight options may present themselves:

Option Category Description
Do Nothing Accept the problem and its consequences.
Improve Current Approach Make incremental changes to existing processes, systems, or structures.
Buy Purchase a product or service to address the problem.
Build Create a bespoke solution internally.
Outsource Transfer responsibility to a third party.
Reorganise Change teams, responsibilities, governance, or decision-making.
Eliminate the Requirement Challenge whether the underlying need actually exists.
Hybrid Combine elements of multiple options.

Once your options are visible, you can move away from which option feels right to calmly comparing options based on visible trade-offs.

Writing down all the variables allows you to focus on each one, rather than trying to juggle them in your thinking. It shows the size of the issue. You might feel it is large, but making it explicit captures the complexity and likely impact, helping you to start to address it. Making the issue concrete gives you the opportunity to delegate parts of the decision to others: making time for you to focus on other, more operationally pressing items.

Benefits

Benefits can be harder to identify than problems, as we tend to jump straight to solutions. But it is worth considering the benefits of your options to help you rule out any options you feel would not lead to enough benefit.

You can identify benefits by asking questions, from a number of perspectives:

Outcome-based Questions

You can ask:

  • What becomes easier?
  • What becomes faster?
  • What becomes cheaper?
  • What becomes safer?
  • What becomes more reliable?
  • What becomes possible that is not possible today?

For add deployment automation with rollback, the benefits might be:

  • Faster releases
  • Reduced operational effort
  • Lower outage recovery time
  • Increased deployment confidence

Stakeholder-based Questions

Different stakeholders will value different benefits. The CEO will value growth and competitive advantage, whereas the CTO will focus on scalability and stability, and customers will value better service and fewer defects.

Other questions you can ask are:

Approach Purpose
Time-Horizon Questions Identify benefits that emerge in the short, medium, and long term.
Quantitative vs Qualitative Benefits Distinguish measurable outcomes from intangible improvements.
"What Good Thing Happens?" Analysis Trace immediate benefits to business outcomes elsewhere.
Capability-Based Analysis Identify new capabilities the option enables.
Compare Against Doing Nothing Focus on benefits that are unique to the option being considered.

Costs

In considering costs do not only consider financial cost. Most significant costs in business decisions are not monetary. Every option you consider will consume something, be it money, time, or the opportunity to do something else.

Systematically ask what must be spent, sacrificed, or endured for an option to succeed.

Cost Type Description
Financial Direct expenditure on products, services, licences, infrastructure, or contractors.
Opportunity cost Benefits or initiatives foregone by choosing one option over another.
Time Time required to implement, adopt, and realise value.
People Effort, skills, capacity, and attention required from individuals and teams.
Transition Temporary disruption caused by migration, change, training, or adoption.
Complexity Additional operational, technical, or organisational burden introduced.
Maintenance Ongoing effort and expenditure required to sustain the solution.
Reversal Cost and difficulty of undoing the decision if it proves unsuccessful.

Risks

A risk is something that might happen.

To surface risk you can ask "What could cause this option to fail, underperform, or create unintended consequences?"

Risk Type Question
Delivery What could prevent successful implementation?
Technical What could fail technically?
People How could people, skills, or adoption become obstacles?
Financial What could increase costs or reduce expected value?
Operational What could disrupt day-to-day business operations?
Strategic How could this weaken future positioning or flexibility?
Legal & Regulatory What compliance or contractual issues could arise?
Reputational How could trust or credibility be damaged?
Pre-Mortem If this failed completely, what would the causes be?
Success Risk If this succeeded beyond expectations, what problems could emerge?

Because a risk may not materialise, different people will asses it differently. To quantify this it can be useful to consider the likelihood of a risk happening, and then asking, what is the impact assuming it occurs.

Some risks are low: they are truly unlikely, but not impossible. However, if they do occur, their impact can be huge.

Some risks are high but should they occur, the impact is low.

For any risk, you break it into likelihood and impact, and give a score between 1 and 10 to each. You then multiply these values together to get a single risk score.

This score will be between 1 and 100. As a business, you define a risk appetite, also out of 100.

Any risk score at or above this appetite is something that your business is not prepared to tolerate.

Assumptions

Explicitly considering assumptions is key as they are often invisible. People may not givn them much - if any - conscious thought; seeing them as part of the problem or its possible solution. But our assumptions can be the key to analysing a business problem and understanding the likely solutions.

These techniques will help you surface your assumptions so that they can be assessed and discussed within the context of the rest of your analysis.

Technique Purpose
Success Preconditions Identify conditions required for success.
Challenge Every Benefit Reveal assumptions hidden inside expected outcomes.
Challenge Every Cost Estimate Expose assumptions embedded in budgets and timelines.
Stakeholder Assumption Analysis Surface assumptions held by different stakeholders.
Dependency Analysis Identify assumptions about external factors and dependencies.
Inversion Expose fragile assumptions by considering their opposite.
Evidence Testing Distinguish proven facts from beliefs and unknowns.
"What Must Be True?" Analysis Systematically uncover assumptions required for success.

Conclusion

The POB-CRA framework makes your thinking explicit so that you can calmly discuss trade-offs. It supports your thinking, enabling you to systematically work through the detail of a problem and its possible solutions.

Not deciding is itself a decision, one that might become too costly for your business.

Need support with a decision? I can help with a one-on-one session: £149 for 60 to 90 minutes

Read next: Your board wants a strategy. You have three weeks.
How do you decide on a good strategy when you only have three weeks to figure it all out?

Your Comments

Please enter a nickname. You do not need to register. A nickname is only required the first time you comment.

Table of Contents

\