3.2.2 - Evidence, intuition and decision making

Data-based decisions and intuition

Scientific decision making uses evidence and systematic analysis, while intuitive decision making relies more directly on judgement, experience or an instinctive sense of what will work.

A scientific approach identifies the problem, defines relevant criteria, gathers information, compares alternatives and checks the consequences of the chosen action. “Scientific” does not mean that an exact answer is guaranteed. It describes a reasoned, evidence-based process. Quantitative data, such as sales, costs and waiting times, can be combined with qualitative evidence, such as customer comments or staff observations.

Illustrative case: a café considers extending its opening hours. It records evening footfall, surveys potential customers and estimates extra staffing and energy costs. Comparing expected additional revenue with those costs gives a more transparent basis for deciding than assuming that extra hours must produce extra profit. A short trial could help test the assumptions before the café makes a long-term commitment.

Data can improve consistency, reveal patterns and let other people challenge the reasoning. It can also make progress easier to assess after implementation. However, collection and analysis take time and money. Past sales may not predict a new market; a small or unrepresentative survey may distort customer demand; missing costs may overstate profit. A spreadsheet makes assumptions easier to calculate, but does not make them true.

Managers should consider relevance, accuracy, representativeness and timeliness. A national survey may be accurate yet poorly matched to one café's local customers. Online comments may overrepresent very satisfied or dissatisfied customers. Old cost quotations may no longer apply. A number with several decimal places can still be based on a weak estimate, so precision should not be confused with reliability.

Intuition can be valuable when the manager has relevant experience, the decision is urgent or useful data are unavailable. An experienced buyer may recognise a likely fashion trend before it appears in past sales figures. The judgement may reflect knowledge that is difficult to express in a formula. It can be quick and useful for developing ideas.

Intuition is also vulnerable to overconfidence, personal preference and memorable but unusual experiences. Success in one context may not transfer to another. An enthusiastic manager might interpret only favourable comments as support for a preferred plan. Calling a decision intuitive does not establish that the manager possesses useful experience.

The approaches can complement each other. Intuition can suggest an option; evidence can test it. Evidence can indicate a pattern; experienced judgement can investigate whether the pattern is plausible in the current context. Under time pressure, managers may use a smaller set of relevant data and an informed judgement, then monitor results. The appropriate balance depends on the decision's consequences, urgency and available evidence.

Netflix's engineers explained in September 2021 that its “Top 10” feature began with a hypothesis and an A/B test. Randomly assigned groups received different product experiences, and Netflix compared engagement alongside other measures. This is scientific decision making: a plausible idea becomes a testable choice rather than relying solely on an executive's taste. A test remains dependent on the chosen measures and whether the tested users represent the intended customers. Netflix's account of A/B testing.

Risk, reward and opportunity cost

A decision must consider possible gains, possible adverse outcomes and what the business gives up by choosing one option.

Reward is a benefit that may result from a decision, such as higher profit, growth, stronger customer loyalty or greater social impact. Risk concerns exposure to adverse outcomes, taking account of their likelihood and consequences. A proposed outlet may earn additional profit, but could lose money if demand is weak. A low-probability outcome can still matter if its impact would threaten survival.

Uncertainty exists when future outcomes or their probabilities are not known reliably. In decision analysis, risk can sometimes be estimated using probabilities, while uncertainty limits confidence in those estimates. Estimates based on past experience can be useful, but an unfamiliar product or disruptive competitor may make the future less predictable. A manager should not treat an estimated probability as a guaranteed frequency or hide uncertainty behind a precise forecast.

Illustrative case: a manufacturer chooses between upgrading existing equipment and launching an unfamiliar product. The upgrade may have a modest, more predictable cost saving. The launch may offer higher rewards but uncertain demand and a possibility of large losses. Choosing the highest possible reward alone ignores the chance and consequences of failure. Choosing the lowest risk alone may overlook a valuable opportunity. The firm's objectives and ability to absorb losses influence the balance.

Limited resources make opportunity cost unavoidable. If the manufacturer can fund only one project, choosing the launch means giving up the benefit of the next best use of those funds, perhaps the equipment upgrade. The opportunity cost is that forgone benefit, not the sum of every possible project and not automatically the money paid for the chosen one.

Opportunity cost can concern time, space or people as well as money. If a manager spends a day negotiating a supplier agreement, the next best use of that day might have been solving a production problem. The opportunity cost is the value of that alternative. Some benefits cannot be expressed reliably in pounds, but they can still matter to a decision.

Risk management can include a pilot, a smaller initial commitment, obtaining more information or developing an alternative response. Such actions also have costs and can delay rewards. A trial may reduce uncertainty about local demand without resolving future competition. Managers should consider what new evidence would actually change the decision and whether it is worth obtaining.

Decision trees, taught in the next lesson, provide a way to represent alternatives and estimated probabilities. They help structure a decision, but the distinction between risk, uncertainty, reward and opportunity cost remains essential when interpreting their numerical results.

Objectives, ethics and feasible choices

A good decision is judged against the business's purpose and the options it can realistically carry out, rather than one isolated forecast.

Mission and objectives define what the decision should achieve. A profit-focused business may prefer the highest sustainable financial return, while a social enterprise may choose an option that serves more beneficiaries with enough income to remain viable. A survival objective can favour a lower-reward option whose potential loss is manageable. The same evidence can therefore lead to different choices in businesses with different priorities.

Ethics influence which methods and consequences are acceptable. A purchasing option may offer lower costs through conditions the business regards as unfair. Rejecting it can reflect an ethical constraint even if it lowers expected profit. Ethical conduct may also support trust and repeat business, but the principle should not be assumed to matter only when it increases returns. Managers need to identify whose interests are affected and what commitment the business has made.

The external environment, including competition, changes the likely consequences. A rival launch may reduce expected demand; changing borrowing costs may alter affordability; new technology may shorten a product's useful life. Forecasts should be reassessed when their conditions change. Waiting for complete certainty is impossible, but ignoring credible new evidence weakens a decision.

Resource constraints set the feasible options. Finance, skilled staff, time, production capacity and access to supplies all matter. A project with an attractive profit forecast cannot be carried out if the firm lacks required capacity or cannot obtain funds in time. Additional resources may be possible, but their costs and delay must be included. Choosing a plan that assumes unavailable skills is different from choosing to recruit and train those skills within a realistic schedule.

Illustrative case: a food producer compares a rapid expansion with a smaller local launch. The larger project forecasts more profit but requires £300,000; only £120,000 of funding is available. It also depends on an unverified supplier whose working conditions conflict with the producer's ethical mission. The local launch forecasts lower profit but uses current equipment and approved suppliers. The larger forecast alone does not establish the better choice. The producer should assess whether finance and acceptable supply can be secured, or whether the smaller launch better fits its objectives and constraints.

Business activities are interrelated. Marketing must establish credible demand, operations must be able to deliver, finance must support the cash commitments and employees must have the relevant skills. Technology can improve information access and speed of analysis, but a decision system using poor data can distribute an error more quickly. Monitoring actual outcomes allows managers to learn and adapt without pretending the original forecast was certain.