Facing an important business decision?

Use our free Business Decision Tool to compare your options, challenge assumptions and assess the risks. Create a clear decision record with practical next steps and a plan to review the outcome.

Introduction – The Search for a Certain Answer

One of the things I have noticed from working with business owners is just how difficult it can be to make an important decision when there is no obviously correct answer.

• Should I take on another employee?
• Should I increase my prices?
• Should I buy the new equipment?
• Should I accept this large contract?
• Should I launch a new service?
• Should I borrow money to fund the next stage of growth?

These are important decisions. They can affect the profitability, cash flow and future direction of the business, so it is entirely reasonable to think carefully before making them. The problem is that many business owners are not simply looking for enough information to make a sensible decision. They are looking for certainty.

They want the forecast to prove that the employee will pay for themselves. They want the business plan to confirm that customers will buy the new service. They want the spreadsheet to demonstrate that the investment will produce a return.

Most of all, they want somebody to tell them that they are definitely making the right decision.

I completely understand why.

When it is your business, your money and your reputation at risk, uncertainty is uncomfortable. Making no decision can feel safer than making the wrong one.

But certainty is not normally available.

No accountant, consultant, forecast or business plan can tell you exactly what is going to happen. We can analyse the information, challenge the assumptions and compare the alternatives, but we cannot remove uncertainty from the decision.

Imagine you are considering employing someone for the first time.

You can calculate their salary, employer’s National Insurance, pension contributions, equipment, training and other employment costs. You can estimate the extra work they will allow the business to complete. You can calculate how much additional revenue will be needed to cover their cost.

All of that information is useful.

What you cannot know for certain is how quickly they will become productive, whether customer demand will continue, how well they will fit into the business or whether employing them will genuinely free you to spend more time on higher-value work.

You still have to make a judgement. The same applies when pricing a large contract.

You might carefully calculate labour, materials, subcontractors, transport, overheads and the profit margin you want to achieve. But there will still be uncertainty.

• Will the project run to schedule?
• Will material prices change?
• Will the customer alter the specification?
• Will the work require more management time than expected?
• Will the customer pay on time?

The fact that the spreadsheet calculates a profit does not guarantee that the project will produce one. This does not make the spreadsheet pointless. Far from it. Good financial information gives us a much stronger basis for making the decision.

But it is important to recognise what that information can and cannot do. It can help us make a better decision. It cannot promise us a particular result.

That distinction matters because the search for certainty can easily become an excuse for doing nothing.

We ask for another forecast. We hold another meeting. We conduct more research. We decide to look at it again next month. Sometimes more information is genuinely needed. But there comes a point when another version of the spreadsheet will not materially improve the decision. We are no longer gathering useful evidence. We are simply hoping that the uncertainty will disappear.

It rarely does. Meanwhile, the cost of delay continues to grow.

The business owner who will not employ anyone remains overloaded and turns work away. The company that keeps postponing a price increase continues accepting unprofitable jobs. The business that spends six months discussing a new service watches a competitor launch it first.

Doing nothing is not a risk-free option. It is still a decision, and it still has consequences.

There is another danger here. Detailed forecasts can create the appearance of certainty where none exists. A forecast showing that a new service will generate £427,500 of revenue can look impressively precise. But where did that number come from?

Perhaps it assumes that the business will attract 15 new customers every month, that each customer will spend an average of £2,375 and that 80% of them will remain for a full year.

The formula may be perfect. The assumptions may not be.

Whenever I look at a forecast, I am less interested in whether the spreadsheet adds up than I am in the assumptions sitting behind it.

• Why do we believe 15 customers will buy?
• What evidence do we have?
• What happens if we attract only ten?
• What happens if customers spend less than expected?
• How much money could we lose before we realise the plan is not working?
• Could we test the idea on a smaller scale before committing fully?

These questions do not give us certainty, but they do give us a much better understanding of the decision.

That is where probabilistic decision-making comes in.

Rather than asking: “Will this definitely work?”

I believe we should ask: “What are the possible outcomes, how likely is each one, and what happens to the business if we are wrong?”

This is not about turning every business decision into a complicated mathematical exercise. Nor is it about attaching an artificial percentage to every possible result. It is simply about accepting that we are making decisions in an uncertain world.

Some outcomes are more likely than others. Some decisions offer greater potential rewards. Some carry consequences that the business can comfortably absorb. Others could place the entire business at risk.

We need to consider all of these things together. We also need to recognise that not every decision deserves the same amount of time and analysis. Jeff Bezos describes decisions as either one-way or two-way doors.

A two-way door is a decision we can reverse fairly easily. We can walk through it, see what happens and step back if it does not work. Testing a new advert might be a two-way door. So might trialling new software, introducing a price change to a small group of customers or using a contractor before recruiting a permanent employee.

If the experiment does not work, the loss is limited, and we can change direction.

A one-way door is different. Once we walk through it, returning is difficult, expensive or perhaps impossible. Signing a ten-year lease, taking on substantial debt, acquiring another company or agreeing to a large fixed-price contract could all be one-way doors.

These decisions deserve much greater scrutiny because the cost of being wrong is considerably higher.

One of the most common mistakes I see is businesses treating every decision as though it were a one-way door. They spend weeks or months analysing choices that are cheap, reversible and easy to test.

The opposite can be even more dangerous. A business makes a large, irreversible commitment as casually as it would run a small marketing experiment.

The amount of evidence we require should reflect the size of the commitment, the potential downside and how easily we can reverse the decision.

We must also be careful about how we judge decisions after the event.

In her book Thinking in Bets, Annie Duke explains that a good decision can produce a bad result, while a bad decision can sometimes produce a good one. Suppose I accept a large contract without properly checking the costs. I have underestimated the labour, allowed nothing for delays and agreed to poor payment terms.

The job happens to run smoothly. The customer makes no changes and pays immediately. Against the odds, I make a profit.

That does not mean I made a good decision. It means I got away with a bad one. If I repeat the same process often enough, eventually one of those projects will go badly wrong.

Now consider the opposite situation. I carry out a thorough credit check on a customer. I review their accounts, obtain references, set a sensible credit limit and monitor their payment record. Six months later, the customer unexpectedly becomes insolvent, and I suffer a bad debt.

The outcome is poor, but that does not automatically mean the original decision was unreasonable.

A good decision can still be followed by bad luck.

This is why we should not judge the quality of a decision solely by what eventually happened. We need to judge it by the information available at the time, the assumptions we made, the risks we considered and the steps we took to protect the business.

Good business decision-making is not about being right every time. That is impossible. It is about making sensible decisions with the information available, understanding what could happen if we are wrong and adjusting our position as new evidence emerges.

Every important business decision is, to some extent, a bet. The aim is not to avoid making bets. It is to understand the odds, protect the downside and make better ones.

1. The Search for a Certain Answer

Most business owners understand that running a business involves risk. Yet when an important decision arrives, many still act as though a completely safe option must exist if they analyse the situation for long enough.

We look for the decision that guarantees success. Unfortunately, it rarely exists.

Suppose you are considering taking on your first employee. You are working too many hours, turning opportunities away and spending time on work someone else could do. Logically, employing someone makes sense.

Then the doubts begin.

What if the work dries up? What if the employee is not good enough or leaves after six months? What if you cannot manage them, or the additional revenue does not cover the cost? These are reasonable questions. The problem comes when we believe we must answer every one before deciding.

We prepare forecasts, calculate employment costs, estimate the required revenue and consult accountants, business owners and recruitment consultants. But none can promise that employing the person will work.

Eventually, we have gathered most of the useful information available. Further analysis is unlikely to change the decision materially. We are no longer seeking information; we are seeking reassurance.

That is understandable. Business decisions are personal. A failed investment may affect your savings, income, reputation, employees, customers and family. It is hardly surprising that owners become cautious.

But caution and certainty are not the same.

Caution means understanding the risks, protecting the downside and making a considered decision. Seeking certainty means refusing to act until every risk has disappeared. The first is sensible. The second is impossible.

More information does not always create a better decision

More information can improve a decision, but only up to a point.

If I am considering buying equipment, I should understand its cost, useful life, finance charges, maintenance requirements and likely effect on productivity. I should know how much additional work the business needs to justify the investment.

But producing five cash-flow forecasts, testing six growth assumptions and researching every competing machine may eventually add little value. Analysis can stop improving the decision and start postponing it.

I still will not know exactly how much work the business will win, whether a competitor will cut prices, whether a major customer will leave or whether the equipment will perform as expected. No analysis can create facts about events that have not happened.

The useful question is not: “Do I have every piece of information?”

It is: “Do I have enough reliable information to make a reasonable decision?”

That is a much better standard.

Forecasts are assumptions written as numbers

I use forecasts regularly and consider them essential to good business management. They help us understand cash requirements, break-even points and possible outcomes.

However, a forecast is not a fact.

A spreadsheet might show that a new service will generate £250,000 in its first year, based on detailed assumptions about sales, customers, prices and conversion rates. The spreadsheet can calculate the result precisely, but the result is only as reliable as the assumptions.

Where will the customers come from? How many enquiries will be generated? What percentage will convert? How quickly will customers decide? Will they buy once or return? How long will delivery take? What if a competitor launches a similar service?

The £250,000 is one possible result, not a fact. If the business attracts fewer customers, converts fewer enquiries or launches late, the outcome will change. That does not make the forecast useless. It means we should use it to explore possibilities rather than predict the future.

What does the expected outcome look like? What if sales are 20% lower or costs 15% higher? How much cash will the business need during the first six months? How severe would the downside need to be before the business came under serious pressure?

A useful forecast does not remove uncertainty. It helps us understand it.

“Do nothing” is still a decision

Delaying a decision can feel safer than making one. If I do not recruit, sign the lease or buy the equipment, I cannot suffer the consequences of choosing badly.

But doing nothing has its own costs.

Consider a business owner who spends 15 hours a week invoicing, chasing paperwork, answering routine emails and updating records. They avoid employing someone because they cannot be certain there will always be enough work.

Yet what does it cost to keep doing that work? Those 15 hours might be used to meet prospects, improve pricing, develop a service or strengthen customer relationships. If the owner’s time could generate £100 an hour in higher-value work, the opportunity cost is potentially £1,500 a week.

That does not automatically mean they should employ someone. It does mean the current position is not free. The business is already paying a price, even if it is less visible than a salary. The same applies to pricing. A business may know its margins are too low but delay increasing prices because it cannot predict customer reactions. It remains busy for another six months while generating insufficient profit and placing pressure on cash flow. The delay has not removed the risk; it has allowed the problem to continue.

“Do nothing” should therefore be assessed alongside every other option:

• What is likely to happen if we make no change?
• What will it cost? 
• What opportunities might we lose? 
• How long can the current situation continue?

Sometimes doing nothing is correct. But it should be an active, evidence-based decision, not the automatic result of failing to find certainty.

Waiting has a cost

Caution is appropriate for genuinely irreversible commitments, such as buying another company, taking on a major loan or signing a long commercial lease.

Many everyday decisions are easier to reverse. Testing a different price, trialling software, running a small advertising campaign or hiring a contractor for excess work does not require certainty.

Taking action may be the best way to obtain the missing information.

Instead of debating whether customers will pay more, quote the higher price to the next ten suitable prospects and monitor the response. Rather than researching a new service for months, speak to existing customers, create a simple offer and seek the first three orders. Instead of wondering whether another person would free up your time, engage a freelancer one day a week and measure the effect.

Small tests turn assumptions into evidence. They do not guarantee success, but they allow us to learn without exposing the business to an unacceptable loss.

The honest answer is often “I don’t know”

Business owners, advisers, managers and employees often feel pressure to appear confident. As a result, we may present estimates as facts:

“We will win this contract.”

“The new employee will pay for themselves.”

“Customers will accept the price increase.”

“The new service will be profitable within six months.”

These statements sound decisive but hide the assumptions behind them.

A more useful answer might be: “I believe we have a good chance of winning the contract, but the margin will be too low unless we control these three costs.”

Or:

“I believe the employee should pay for themselves within 12 months, provided demand remains steady and I use the time they release to generate new work.”

Or:

“I think most customers will accept the increase, but we should test it first and watch our most price-sensitive accounts.”

These answers are less certain but more useful. They explain what we believe, why we believe it and which assumptions matter. They also show what to monitor after the decision.

The aim is not certainty; it is a sound decision

Good business owners are not those who always make the right decisions. That standard is unrealistic because it assumes we can know what will happen before it happens.

Good owners use the best information reasonably available. They understand their assumptions, consider what could go wrong and protect the business from outcomes it could not afford. They also accept that uncertainty will remain.

The objective is not to find an answer that cannot be wrong. It is to choose the course of action most likely to produce a worthwhile result, with a downside the business can understand and withstand.

That allows us to be careful without becoming paralysed, and confident without pretending to predict the future.

2. Business Decisions Are Made Under Uncertainty.

Every important business decision is made with incomplete information.

We may have management accounts, cash-flow forecasts, customer research and years of experience. We may speak to employees, advisers and other business owners. All of this can improve the quality of the decision.

But none of it allows us to know exactly what will happen.

Suppose you are considering launching a new service. You can research the market, speak to customers and estimate the likely cost of delivery. You can look at competing services, test different prices and prepare a detailed forecast.

That work is valuable because it replaces some of the unknowns with evidence.

However, uncertainty remains.

Customers who express interest may not buy. Competitors may respond. Delivery may take longer than expected. Marketing costs may be higher. The economy may change before the service becomes established.

The decision must still be made before the result is known.

Better information reduces uncertainty

There is a danger of taking this argument too far and concluding that research and planning are pointless because the future cannot be predicted. That would be wrong.

Uncertainty is not an excuse for guesswork.

If I am considering investing £100,000 in a new project, I want to understand the likely return, cash requirement, risks and alternatives. I want to know which assumptions are supported by evidence and which are based mainly on judgement.

Better information may not remove uncertainty, but it can reduce it.

If ten existing customers have already agreed to trial a new service, that is more useful than simply believing customers will want it. If similar projects have consistently achieved a 30% gross margin, that gives us a stronger starting point than an untested estimate.

If a cash-flow forecast shows that the business can fund the downside for 12 months, we can make the decision with greater confidence. The aim of analysis is not to produce certainty. It is to make the remaining uncertainty visible.

Different decisions contain different levels of uncertainty

Not all unknowns are equal.

Some decisions involve risks we can estimate reasonably well. A business that has completed 100 similar projects may have reliable information about labour hours, material costs and likely delays. Other decisions involve much greater uncertainty. Launching an entirely new product into a new market gives us less evidence to work with.

This distinction matters.

When we have reliable historical information, our estimates should carry more weight. When we are doing something new, we should be more cautious about precise forecasts and more willing to test our assumptions.

A 12-month forecast for an established service with recurring customers is very different from a 12-month forecast for an idea that has never generated a sale. The spreadsheets may look equally professional. The certainty behind them is not equal.

Uncertainty cannot be delegated

Business owners sometimes hope an adviser will remove uncertainty for them.

They ask the accountant whether they should employ someone, the marketing consultant whether a campaign will work or the solicitor whether they should accept a contract. A good adviser can clarify the facts, identify risks and challenge assumptions. They may also explain consequences the owner has not considered.

But they cannot guarantee the result.

The accountant can calculate the cost of employing someone and the revenue needed to cover that cost. They cannot promise that customer demand will remain strong. The marketing consultant can use experience and data to design a better campaign. They cannot guarantee that every prospect will respond.

The solicitor can identify an unfavourable contractual term. They cannot know whether the commercial relationship will succeed. Advice improves judgement, but the decision still belongs to the business owner.

Make uncertainty part of the decision

Rather than trying to hide uncertainty, I believe we should include it openly in the decision-making process.

We can ask:

• What do we know?
• What are we assuming?
• Which assumptions matter most?
• How confident are we in them?
• What could make the outcome better or worse?
• What information would materially change the decision?
• What can we monitor after deciding?

This creates a more honest discussion.

Instead of saying, “The project will make £50,000,” we might say: “Our expected profit is £50,000, provided labour remains within 2,000 hours, material costs do not increase by more than 5%, and the customer pays according to the agreed schedule.”

That does not weaken the proposal. It tells us what needs to be true and what must be monitored. Good decision-making does not require us to eliminate uncertainty before acting. It requires us to understand the uncertainty, decide whether the potential return justifies it and protect the business if events do not go as planned.

3. Every Decision Is a Bet

Describing a business decision as a bet may sound uncomfortable. We tend to associate betting with gambling, guesswork and unnecessary risk. 

That is not what I mean. A bet is simply a decision made before the outcome is known. Seen this way, business owners make bets every day.

• When you employ someone, you are betting that the value they create will exceed the cost of employing them. 
• When you increase prices, you are betting that the additional margin will outweigh any sales you lose.
• When you grant a customer credit, you are betting that they will pay in full and on time.
• When you accept a fixed-price contract, you are betting that the cost of delivering it will remain within your estimate.

Recognising these decisions as bets does not make them reckless. It makes the uncertainty visible.

Decisions are not predictions

We often present business decisions as though they are predictions:

“This salesperson will generate £300,000 of revenue.”

“This campaign will produce 100 enquiries.”

“This new service will be profitable within six months.”

These statements sound confident, but they hide the assumptions behind them. 

A more useful statement would be: “Based on their experience, our current pipeline and the size of the market, I believe this salesperson has a reasonable chance of generating £300,000.”

That leads to better questions.

What evidence supports the target? How long will it take them to become productive? What gross profit will the revenue produce? What happens if they achieve only £200,000? At what point will they cover their full employment cost?

• A prediction says, “This will happen.”
• A considered bet says, “Given what I currently know, I believe this is the best course of action.”

Doing nothing is also a bet

A business does not avoid making a bet by deciding not to act.

Suppose a business is operating at full capacity but decides not to recruit because the economy feels uncertain. The owner may believe they have avoided risk. In reality, they have bet that remaining understaffed is the better option. They are betting that customers will tolerate delays, employees can continue managing the workload, and the opportunities being turned away will not matter.

That decision may prove correct if demand falls. But demand may continue growing, service may deteriorate, and good employees may leave.

There is no risk-free position. There are only different choices with different possible outcomes.

The same applies to pricing. Keeping prices unchanged is a bet that protecting sales volume is more valuable than improving margin, and that costs will not rise faster than the price charged.

Once we recognise this, doing nothing becomes another option that must justify itself.

State the odds honestly

In many small businesses, we will not have enough information to calculate an exact probability. Saying there is a 63% chance of success may sound scientific, but the number could be little more than an opinion.

That does not make probabilistic thinking useless. We can still describe an outcome as unlikely, possible, reasonably likely or highly likely. The purpose is not to attach false precision to the decision. It is to expose the reasoning behind our confidence.

If I say a new service is reasonably likely to succeed, someone can ask why.

Perhaps customers have already requested it. Perhaps we have delivered it successfully on a small scale. Perhaps we know the likely price and cost. Alternatively, I may discover that my confidence is based on enthusiasm rather than evidence.

That is valuable information.

Improve the odds

We should also ask what we can do to increase the probability of success.

• Before developing a new service, speak to customers or secure the first few orders.
• Before employing a full-time member of staff, test the need using a freelancer or contractor where appropriate.
• Before accepting a major project, clarify the scope, price the risks and negotiate staged payments.
• Before purchasing equipment, hire it and measure the effect on productivity.

These steps either provide better information or limit the downside. They improve the quality of the bet. 

Annie Duke’s idea of “thinking in bets” helps move the conversation away from arguing over who is right. Instead, we ask:

“What do we believe? How confident are we? What evidence supports that belief? Which assumptions matter most? What would cause us to change our minds?”

We can then commit to a decision without pretending the outcome is guaranteed. The aim is not to stop making bets. That would mean stopping investment, change and growth. It is to recognise the bets we are already making, and become better at choosing which are worth taking.

4. One-Way Doors, Two-Way Doors and the Cost of Getting Decisions Wrong

Not every business decision deserves the same amount of time, analysis and caution. Some decisions are easy to reverse. We can make them, observe the result and change direction if they do not work. Others create commitments that are difficult or expensive to undo.

Jeff Bezos describes these as two-way doors and one-way doors. A two-way door allows us to step through, see what happens and step back if necessary. A one-way door closes behind us, making the return journey much more difficult.

This provides a useful way to decide how much evidence we need before acting.

Two-way doors

A two-way door might include:

• testing a new price with a small group of prospects;
• running a limited marketing campaign;
• trialling new software;
• using a contractor before recruiting permanently;
• offering a new service to existing customers;
• buying a small quantity of stock to test demand.

These decisions still require thought, but they do not require certainty.

Suppose a business is considering spending £1,500 on an advertising campaign. It understands its target customer, has a clear offer and can measure the enquiries generated. If the campaign fails, the business loses £1,500. That is disappointing, but assuming it can comfortably afford the loss, it is not disastrous. The campaign can be stopped or changed.

The business will probably learn more by running a controlled test than by discussing it for another two months.

One-way doors

A one-way door creates a more significant commitment. Examples include signing a long lease, acquiring a business, taking on substantial debt, purchasing specialist equipment or accepting a large fixed-price contract.

Consider a company signing a ten-year lease.

The commitment is not simply the monthly rent. It may include a deposit, service charges, business rates, fitting-out costs, repairs and dilapidations. If the company later decides it does not need the premises, it cannot necessarily hand back the keys and stop paying.

The new premises may still be the correct decision, but stronger evidence is needed because the cost of being wrong is greater.

We should understand the total commitment, effect on cash flow, sales required to cover the cost and whether the business could survive a downturn. We should also explore break clauses, subletting and alternative premises.

A one-way door is not a door we should never enter. Businesses could not grow without long-term commitments. We simply need to recognise it before walking through.

Match the process to the decision

Businesses often spend too long analysing small, reversible decisions.

Imagine a software tool costs £100 a month, could save five administrative hours a week and can be cancelled with 30 days’ notice. Holding several management meetings may cost more than trialling it.

The opposite mistake is more dangerous: treating a one-way door casually.

A business buys equipment based on one customer’s promised workload, accepts a large contract without modelling the cash requirement or signs a lease because the premises feel right.

Before making a large commitment, I would ask: “If this goes badly, how do we get out?” If the answer is unclear or unaffordable, we are dealing with a one-way door.

Make large decisions more reversible

We can often redesign a decision to reduce the cost of being wrong.

• Instead of signing a ten-year lease, negotiate a break clause.
• Instead of buying equipment immediately, hire it first.
• Instead of recruiting an entire team, add capacity in stages.
• Instead of building a complete service, sell a small pilot.
• Instead of accepting unclear fixed-price work, agree on a discovery stage and clear variation terms.
• Instead of funding a major contract entirely, request a deposit and staged payments.

These changes do not remove risk. They limit the commitment while the business gathers evidence.

A company may pay more to rent equipment for its first three projects, reducing the initial margin. However, it learns whether demand is real and whether the work can be delivered profitably without committing £100,000 to an asset it may not need.

The higher short-term cost buys valuable information and keeps the return journey open.

Before deciding, ask whether the choice is reversible, what the credible downside is, how quickly the result will become clear and whether the commitment can be tested or staged. Low-cost, reversible decisions should usually be made quickly. Large, difficult-to-reverse decisions deserve stronger evidence and greater protection.

We cannot remove the possibility of being wrong. But we can often control how much being wrong will cost.

5. Probability, Payoff and the Cost of Being Wrong

When assessing an opportunity, business owners often begin by asking: “How likely is this to work?”

That matters, but it is not enough. We must also consider what we gain if it works and what it costs if it does not.

A decision with a high probability of success may offer very little return. Another with a lower probability may offer a much larger reward. A third may look attractive until we discover that one bad outcome could threaten the entire business.

A sound decision considers three things together:

• the probability of success;
• the potential payoff;
• the cost of being wrong.

A likely success may still be a poor decision

Suppose a business is offered a £100,000 contract and is confident it can deliver the work successfully. The contract value sounds attractive, but it tells us little about the payoff. What will the materials and labour cost? How much management time will it consume? Will additional people or finance be required? What other work will the business be unable to accept?

If the project occupies the team for three months but produces only £3,000 of profit, successful delivery does not make it a good commercial decision.

Revenue is not the payoff. The payoff is the profit, cash and wider value remaining after the business fulfils its obligations.

The opposite can also be true. A £5,000 investment with only a 40% chance of success may still be attractive if it could create £100,000 of recurring revenue and the loss is limited to the initial £5,000.

The most likely option is not always the best one.

Timing matters

We must also consider when the payoff will arrive.

A project might eventually produce £50,000 of profit but require the business to spend £300,000 before receiving most of the customer’s payment. It may be profitable on paper while creating a dangerous cash-flow problem.

Whenever I assess an opportunity, I want to know how much cash must be committed, when it will leave, when payment will arrive and whether the business can fund delays or disputes.

A contractor may expect a £100,000 gross profit from a large project, but still need to fund wages, materials and subcontractors while waiting 60 days for payment. If the business cannot finance that gap, the eventual profit becomes irrelevant.

Calculate the credible downside

The cost of being wrong is often greater than the obvious investment.

If a new employee does not work out, the cost includes more than salary. There may be recruitment fees, employer’s National Insurance, pension contributions, equipment, training, management time and the cost of finding a replacement.

If a new service fails, the loss may include development, marketing, software, unused stock and time diverted from profitable work.

I prefer to consider the maximum credible loss: a realistic downside rather than an extreme disaster.

What happens if sales reach only half the forecast? What if costs are 20% higher? What if the customer pays two months late? What if the employee leaves after six months?

The purpose is not to frighten ourselves into rejecting the opportunity. It is to understand what the business may genuinely need to absorb.

Can the business survive being wrong?

The same opportunity can be sensible for one company and reckless for another.

An investment may have a 70% chance of making £200,000 and a 30% chance of losing £150,000. That may be attractive to a well-capitalised business with £500,000 available. For a company with £30,000 in the bank, the unfavourable outcome could be fatal.

The probability and payoff have not changed. The business’s ability to carry the risk has. Businesses can recover from many poor decisions. They cannot recover from one that destroys them.

That is why I would apply a simple rule: Never allow one decision to expose the business to a loss it cannot survive.

This does not mean avoiding every large opportunity. It means structuring the risk through deposits, staged payments, insurance, credit facilities, spending limits or firm customer commitments.

The best decision is not always the one most likely to succeed or the one offering the largest potential return. It is the one where probability, reward and downside combine to create an attractive and survivable risk for that particular business.

We cannot guarantee the outcome. But we can make sure success is worth having, and that being wrong does not cost more than the business can afford.

6. Is This a One-Way or Two-Way Door?

One of the most useful questions we can ask before making a decision is: “How difficult will this be to reverse?” Jeff Bezos describes decisions as either one-way or two-way doors.

A one-way door is difficult or expensive to reverse. Signing a long commercial lease, acquiring another business, giving away equity or buying highly specialised equipment could all create lasting commitments.

A two-way door allows us to try something, assess the result and step back if it does not work. Testing a new price, trialling software, running a small marketing campaign or outsourcing a task for three months are usually more reversible.

The distinction matters because the two decisions should not be approached in the same way.

A one-way door deserves careful analysis because the cost of being wrong may be substantial. A two-way door should usually be approached more quickly because delay may cost more than an unsuccessful experiment.

Match the analysis to the decision

Businesses sometimes apply the same approval process to every decision, regardless of its size or reversibility.

Imagine a company is considering software costing £100 a month. It could save several hours of administration each week and can be cancelled with 30 days’ notice. If four managers attend several meetings, request a detailed business case and postpone the trial for three months, the cost of deliberation may already exceed the possible loss.

The business does not need to prove that the software will work. It needs to define what success looks like, trial it and review the evidence.

The potential downside is limited. The decision is reversible. We can afford to be imperfect. A ten-year property lease is different. It may involve rent, rates, service charges, repairs, fitting-out costs and dilapidations. Leaving early may be difficult.

That decision justifies deeper financial modelling, professional advice and serious consideration of the downside. The amount of analysis should be proportionate to the commitment, not to how nervous the decision makes us feel.

Use action to obtain information

With reversible decisions, action is often the fastest and most reliable way to learn.

A business owner could spend months debating whether customers will pay a higher price. Alternatively, they could quote the new price to the next ten suitable prospects and measure the response.

They could commission extensive research into a new service or offer a simple paid pilot to three existing customers.

They could spend weeks discussing whether more administrative support would help or engage a freelancer for one day a week and measure the time released.

These tests will not answer every question, but they replace opinions with evidence.

Excessive deliberation has a cost. Opportunities are missed, inefficient processes continue, and management time is consumed. Meanwhile, the business learns nothing because no action has been taken.

Make one-way doors more reversible

Some decisions initially appear to be one-way doors but can be restructured. Instead of signing a ten-year lease with no exit, negotiate a break clause.

• Instead of purchasing equipment immediately, rent it for the first few projects. Instead of recruiting an entire team, appoint one person and add capacity as demand becomes clearer.
• Instead of investing heavily in a complete product, develop a limited version and test whether customers will pay.
• Instead of funding a large contract until completion, require a deposit and staged payments.
• Instead of accepting an unclear fixed-price project, begin with a paid discovery stage and agree how variations will be charged.

These arrangements may cost more initially or reduce the first stage’s profit. Renting equipment can be more expensive than owning it. A short lease may carry a higher monthly rent. A small pilot may be less efficient than a full launch.

That additional cost can still be worthwhile because it buys flexibility and information. The business limits its exposure while testing the assumptions behind the larger commitment.

Commit in stages

Large decisions do not always need to be made in one step.

A new service could progress through stages:

1. Speak to customers and confirm the problem.
2. Secure several paid pilot customers.
3. Measure the delivery cost and customer response.
4. Refine the service and pricing.
5. Invest in the people and systems needed to scale.

Each stage requires further commitment, but only when the previous stage has produced enough evidence to justify it.

Before proceeding, ask:

• How easy is this decision to reverse?
• What is the maximum credible loss?
• How quickly will we know whether it is working?
• Can we test the main assumption more cheaply?
• Can we reduce, stage or delay the commitment?
• What would allow us to exit?
• What is the cost of waiting?

The purpose is not to avoid one-way doors. Growing a business eventually requires commitments that cannot be tested forever.

The aim is to move quickly when failure is affordable, slow down when it is not, and keep the route back open wherever possible.

7. Think in Bets, Not Predictions

Business discussions are often full of absolute statements: 

• “We will win this contract.”

• “The new employee will pay for themselves.”

• “Customers will accept the price increase.”

• “This project will make £50,000.”

These statements sound confident, but they disguise the uncertainty involved. They also discourage useful questions. If someone claims an outcome will definitely happen, challenging the assumptions can appear unnecessarily negative.

Annie Duke’s idea of “thinking in bets” gives us a better approach.

Instead of pretending to know the result, we state what we believe, how confident we are and what evidence supports that confidence.

We might say:

“I believe we have a good chance of winning the contract because we have a strong relationship with the customer and our price is competitive. However, I would not include the revenue in our cash-flow forecast until the contract is signed.”

Or:

“I expect the employee to cover their cost within 12 months, provided demand remains steady and I use the time they release to generate new work.”

This is not weaker decision-making. It is more honest and useful because it exposes the conditions required for success.

Confidence is not certainty

We do not need a precise percentage for every decision. Describing an outcome as possible, reasonably likely or highly likely may be enough.

The important point is to explain why.

If I believe there is a 70% chance that a new service will succeed, what supports that confidence? Have customers asked for it? Have we tested the price? Do we understand the delivery cost? Have we already delivered it successfully on a small scale?

Equally, what creates the remaining 30% of doubt? Perhaps demand has not been proven, a competitor may respond, or the business lacks the capacity to deliver consistently. Stating confidence in this way improves the discussion. It allows other people to challenge the reasoning, contribute information and identify assumptions we may have overlooked.

It also helps us decide what to monitor after acting.

Do not judge the decision only by the outcome

Once the result is known, another problem appears.

We naturally judge the quality of the decision by what happened next. Annie Duke calls this “resulting.”

If the outcome was good, we assume the decision was good. If the outcome was bad, we assume somebody made a mistake.

But good decisions sometimes produce bad outcomes, and bad decisions sometimes produce good ones.

Suppose a business accepts a large contract without properly checking the scope, calculating the cost or agreeing how variations will be charged.

The project happens to run smoothly. The customer requests no changes and pays immediately. The business makes a profit.

That does not mean accepting the contract was a good decision. The business was fortunate. Repeating the same process will eventually produce a very different result.

Now consider a business granting credit to a customer. It reviews the accounts, obtains a credit report, checks the payment history and sets a sensible limit. Six months later, the customer unexpectedly fails.

The bad debt does not automatically make the original decision unreasonable. The available evidence may have supported granting credit, but an unfavourable outcome still occurred.

Separate judgement from luck

When reviewing a decision, I would ask:

• What information was available at the time?
• Were the assumptions reasonable?
• Did we consider realistic alternatives?
• Did the potential return justify the risk?
• Did we protect the business from an unacceptable downside?
• Did we respond appropriately as new information arrived?
• How much of the result came from judgement, execution, external events or luck?

This helps us learn the correct lesson.

If a poorly priced project makes money because delivery was unusually easy, the lesson is not that the pricing method works.

If a well-planned marketing campaign fails, it does not prove that investing in marketing was wrong. The audience, offer, message or timing may need to change.

A successful outcome should not protect a weak decision from scrutiny, and an unsuccessful one should not automatically condemn a sound process.

The goal is not to make decisions that always produce good results. That is impossible.

The goal is to make decisions that were reasonable based on the information available—and to learn from the outcome without allowing hindsight or luck to rewrite the story.

8. Keep a Decision Record

One of the difficulties with reviewing a decision is that our memory changes once we know the outcome.

If the decision succeeds, we remember feeling more confident than we really were. We downplay the risks and convince ourselves that the result was obvious.

If it fails, we often do the opposite. Warning signs that seemed minor at the time suddenly appear impossible to have missed. We tell ourselves that we “knew it would happen,” even if we supported the original decision.

This is hindsight bias, and it makes it difficult to learn the right lessons.

A simple decision record can help.

Before making an important commitment, write down:

• the decision being made;
• the options considered;
• the information available;
• the assumptions being relied upon;
• the possible outcomes;
• the estimated likelihood of each outcome;
• the potential payoff;
• the maximum credible downside;
• the reasons for choosing the preferred option;
• the results expected at specific review points;
• the evidence that would cause the decision to be changed.

This does not need to become a long report or another layer of bureaucracy. For most decisions, one or two pages will be enough. The purpose is to capture what we genuinely believed before the result became known.

Record the assumptions

Suppose a business decides to employ a new salesperson.

Its decision record might state:

• The full annual employment cost will be £55,000.
• The salesperson is expected to take three months to become productive.
• The target is £300,000 of additional annual revenue.
• The expected gross margin on that revenue is 35%.
• The existing pipeline and market demand support the target.
• Progress will be reviewed after three, six and nine months.
• If qualified opportunities are significantly below target after six months, the sales approach and role will be reconsidered.

This creates something against which the decision can later be assessed.

If the salesperson produces only £180,000 of revenue, the business can examine why. Was the revenue target unrealistic? Was the gross margin lower than expected? Did the salesperson receive enough support? Did market demand change?

Without a written record, the original assumptions are easily forgotten or changed to fit the outcome.

Include confidence, not just numbers

A decision record should distinguish between facts, estimates and assumptions. The employee’s salary is a known cost. The expected sales value is an estimate. The belief that the existing pipeline will support those sales is an assumption.

Treating all three as equally reliable creates false confidence.

I would also record the level of confidence behind important assumptions. This does not require precise percentages. We might describe them as low, moderate or high confidence.

For example:

“We have high confidence in the employment cost, moderate confidence in the gross margin and lower confidence in the time required to build the new sales pipeline.”

That immediately shows where the greatest uncertainty lies and where monitoring should be concentrated.

Set the review point in advance

A decision should include a date on which it will be reviewed.

Without a review point, businesses tend to do one of two things. They abandon an idea too quickly because early results are disappointing, or continue indefinitely because they have already invested money and effort.

Neither response is particularly rational.

A review point gives the decision enough time to produce meaningful evidence while preventing it from continuing without challenge. The record should also say what the business expects to see by that date. “Review in six months” is less useful than:

“After six months, we expect the salesperson to have created £200,000 of qualified opportunities and secured at least £75,000 of new work.”

The decision record is not designed to prove later that somebody was right or wrong. It is there to improve learning.

It allows us to compare what we expected with what actually happened, identify which assumptions were sound and make better decisions next time.

9. A Practical Probabilistic Decision Framework

Probabilistic decision-making does not need to involve complicated mathematics.

For most business decisions, a structured conversation and a single page of notes will provide more value than an elaborate model containing assumptions nobody has challenged.

The purpose is to slow down enough to understand an important commitment without allowing analysis to prevent action.

I would use the following process.

1. Define the decision clearly

Begin by stating exactly what is being decided.

“Should we grow the business?” is too vague.

“Should we recruit an administrator at a total annual cost of £35,000 from January?” is a decision we can assess.

So is:

“Should we increase prices by 8% for new customers from next month?”

A clear question prevents the discussion from drifting and identifies the point at which a decision must be made.

2. Identify the realistic options

Most decisions involve more than a simple yes or no.

If the business needs additional capacity, the options might include:

• employing someone full-time;
• recruiting part-time;
• using a contractor;
• outsourcing the work;
• improving the process;
• postponing the decision;
• continuing with the existing arrangement.

Doing nothing should appear on the list because it has costs and consequences of its own.

We should compare the genuine alternatives rather than assess one proposal in isolation.

3. Separate facts from assumptions

Write down what is known and what is being estimated.

If we are considering an employee, the salary and employer’s costs can be calculated reasonably accurately. The additional revenue they will help create is an assumption.

If we are assessing a large contract, the quoted selling price may be known. The labour hours, variation risk and payment timing may be estimates.

This distinction matters because a spreadsheet can make assumptions look like facts.

Ask:

• What evidence supports this assumption?
• How reliable is that evidence?
• What would happen if the assumption were wrong?

Concentrate on the assumptions that have the greatest effect on the result.

4. Describe the possible outcomes

I would normally consider at least three scenarios:

• Expected case: what we reasonably believe will happen.
• Upside case: what happens if results are better than expected.
• Downside case: what happens if the important assumptions prove too optimistic.

For larger decisions, I would add a survival case: the combination of credible events that could seriously threaten the business.

The aim is not to create endless versions of the future. It is to understand the range of outcomes rather than relying on a single forecast.

5. Estimate the probability

We may be able to use historical figures, industry data or previous experience. In other cases, the assessment will rely more heavily on judgement.

Either is acceptable, provided we are honest about it.

We might use percentages, but simple descriptions can work just as well:

• unlikely;
• possible;
• reasonably likely;
• highly likely.

The important question is why we believe one outcome is more likely than another.

If we cannot explain the reasoning, the probability is probably based on optimism, fear or instinct rather than evidence.

6. Assess the payoff and downside

Consider what the business gains if the decision succeeds.

That may include profit, cash, recurring revenue, additional capacity, time saved or a stronger market position.

Then consider the maximum credible loss.

Do not limit the downside to the obvious expenditure. Include management time, working capital, lost opportunities, disruption and any effect on employees or customers.

For a credit decision, the downside may be the unpaid invoice and the profit lost on the goods or services already supplied.

For a new contract, it may include cost overruns, slow payment and the profitable work displaced while capacity is occupied.

The central question is: “Can the business withstand the downside without placing its future at risk?”

7. Decide whether the commitment is reversible

Is this a one-way or two-way door?

• If the decision is low-cost and easily reversed, decide quickly and learn from action.
• If it is difficult to reverse, require stronger evidence and more protection.

Then ask whether the commitment can be redesigned. 

Could we run a pilot, use a contractor, request a deposit, introduce staged payments, negotiate a break clause or invest in phases?

Reducing the initial commitment allows the business to obtain information before accepting the full risk.

8. Make the decision and record the reasoning

State what has been decided and why.

Record:

• the evidence available;
• the important assumptions;
• the expected outcome;
• the principal risks;
• the level of confidence;
• the maximum acceptable loss;
• the measures being used to protect the business.

This creates a fair record of the decision before the result is known.

It also ensures that everybody understands what must be true for the decision to succeed.

9. Set review points and warning indicators

Decide in advance when the decision will be reviewed and what evidence will matter. 

• If prices are increased, monitor conversion rates, margin and customer losses.
• If an employee is recruited, review productivity, capacity released and progress against the objectives of the role.
• If credit is granted, monitor payment behaviour, outstanding balances and changes in the customer’s circumstances.

Agree what would cause the business to continue, invest further, make changes or stop.

10. Update the decision

When new information arrives, use it.

If the evidence improves, the business may increase its commitment. If the assumptions weaken, it may reduce its exposure or change direction.

This does not necessarily mean the original decision was wrong. It means the next decision is being made with better information.

The framework can be summarised in ten questions:

1. What are we deciding?
2. What are the realistic alternatives?
3. What do we know, and what are we assuming?
4. What are the possible outcomes?
5. How likely is each outcome?
6. What do we gain if it works?
7. What is the maximum credible downside?
8. Can the decision be tested or reversed?
9. What evidence will we monitor?
10. What would cause us to change course?

The framework will not provide certainty. That is not its purpose.

It helps us make a defensible decision using the information available, while protecting the business from risks it cannot afford.

That is usually the best any decision-making process can do.

Final Word – You Do Not Need Certainty to Move Forward

Business owners are often expected to have the answers.

Employees want direction. Customers want confidence. Banks, investors and suppliers want reassurance. Even when we are uncertain, we can feel pressure to behave as though the correct course of action is obvious.

But good decision-making is not about knowing exactly what will happen. It is about making the most sensible choice using the information available.

That means understanding what we know, identifying what we are assuming and considering the range of possible outcomes. It means weighing the probability of success against the potential payoff and the cost of being wrong.

It also means recognising the type of decision we are facing.

If it is a two-way door, we should normally act, test and learn. If it is a one-way door, we should take more time, challenge the assumptions and protect the downside.

Where possible, we should make large commitments smaller and more reversible. We can use pilots, deposits, staged investment, break clauses and clearly defined review points to gather evidence before risking more.

Most importantly, we should stop judging every decision solely by its eventual result.

A well-considered decision can still produce a disappointing outcome. An ill-considered decision can occasionally succeed through luck. The quality of the decision depends on the reasoning, evidence and risk controls used at the time—not simply what happened afterwards.

We will never remove uncertainty from business. Nor should we allow it to prevent us from acting.

The aim is not to be certain.

It is to make better bets, limit the cost of being wrong and remain willing to change course when the evidence changes.