Occam's Razor

Law 18 / 20 Question Easy Answers

When two explanations fit the evidence equally well, start with the one that needs fewer assumptions, then test it, because simpler is not the same as true.

  • Adage or reasoning rule
  • 5 min read
  • Illustrative scenario
  • Reviewed by Ali Dezhami, November 8, 2026

The dilemma

Illustrative scenario

Daniel is head of marketing at a city bike-sharing operator. On Thursday morning the ad dashboard shows new-member sign-ups attributed to paid social down by more than a third week over week. The brand manager already has a story: a rival e-scooter company launched in the same week, and its discount is pulling riders away. She wants to double paid social spend and cut the first-month price before the weekend.

The data engineer points out two other facts. A new cookie-consent banner went live on Monday, and the app's own database shows new accounts down only slightly. Those facts fit a duller explanation: fewer sign-ups are being tracked and attributed, not necessarily fewer are happening.

Both stories could be partly true. One requires an emergency price cut costing real margin; the other requires a few hours of checking. Daniel has to decide what to believe first, and what evidence would change his mind.

The short answer

Occam's Razor is the principle that, when several explanations account for the evidence comparably well, you should prefer the one that makes fewer assumptions or postulates fewer entities. It is named after William of Ockham, a fourteenth-century English philosopher, although the famous wording "entities must not be multiplied beyond necessity" is a later paraphrase. In practice it is a rule for ordering investigation: try the explanation that needs the least extra machinery first, and keep it only if the evidence holds up.

How strong is the evidence?

Adage or reasoning rule

A memorable saying or reasoning rule, not an experimental finding. Treat it as a prompt to check, not as evidence.

This is an adage in the sense used in this course: a methodological principle from philosophy and science, not an empirical finding about the world. Philosophers of science distinguish two kinds of simplicity it can refer to: elegance, meaning fewer or less complicated hypotheses, and parsimony, meaning fewer kinds or numbers of things assumed to exist.

Why simplicity should guide us toward truth is genuinely disputed. Some defend it as reflecting how the world tends to be, others as a practical rule that makes theories easier to test and to reject, and others doubt there is a single justification at all. What is not disputed is the condition built into the razor: it only applies when the competing explanations fit the evidence comparably well. It never licenses ignoring evidence that the simple story cannot explain.

In marketing analysis the razor is most useful as a triage rule. Measurement changes, tagging errors, seasonality and site releases are frequent, cheap to check and require no new assumptions about competitors or customer psychology.

Investigating an example

Illustrative scenario

The conversion drop that was a tag change

An online furniture retailer sees conversion rate in its analytics tool fall sharply overnight. The paid-media agency suspects a competitor's sale, and the creative team suspects ad fatigue. Before anyone changes budgets, an analyst compares analytics purchases with orders in the shop's back office. Orders are flat; tracked purchases are not. A site release the previous evening had moved the order-confirmation page, and the purchase event no longer fired on one payment method.

Fixing the tag restores the reported numbers. The competitor's sale was real, but the data gave no sign that it had changed customer behavior that week.

This is a teaching scenario, not a real company. The useful move was not "the simple answer is right" but "the simple answer is cheapest to test and would explain every fact we have". Had orders also fallen, the analyst would have moved on to market explanations.

A teaching scenario. The people and numbers are invented; it is not a company case.

Choose your move

You are in the dilemma above. Which move would you make? Open each one to see its trade-offs.

A Act on the competitor story: double paid social and cut the first-month price before the weekend.

If the competitor is the cause, acting fast protects share. If the drop is mostly a tracking artifact, the company gives away margin and spends more against a number that was never real.

B Spend a few hours on the cheapest discriminating checks first: database sign-ups versus attributed sign-ups, consent rates, release notes.

Delays any response by a day and may frustrate leadership. If the checks show real sign-ups falling too, the team responds with better evidence; if not, it avoids an expensive mistake.

C Assume it is the consent banner and change nothing.

Uses the razor as a verdict instead of a starting hypothesis. If the competitor really is pulling riders away, the company loses time it cannot recover.

No move wins in every business. The law tells you what to check, not what to choose.

How to apply it

  1. List rival explanations. Write down every plausible cause, including measurement, release, seasonality, competition and pricing.
  2. Count the assumptions. For each explanation, note what else would have to be true. Prefer starting with those that need the fewest new assumptions.
  3. Write what each one predicts. A tracking fault predicts that back-office numbers hold; a real demand drop predicts they fall too. Look for evidence that separates the explanations.
  4. Run the cheapest discriminating test first. Compare sources of truth, check release and tag logs, segment by device or browser. Time-box it.
  5. Escalate when the simple story fails. If a fact contradicts the simple explanation, drop it and move to the next one rather than stretching it to fit.

What to measure

Agreement between independent sources (analytics, ad platform, back-office orders or sign-ups), time from anomaly to confirmed diagnosis, and the share of past alarms traced to measurement issues. Guardrail: no budget or price change on an anomaly until at least one independent source confirms it.

Worked example

Worked example: ordering explanations for a drop in attributed sign-ups (illustrative)
ExplanationExtra assumptions neededWhat it predictsCheapest check
Consent banner reduced tracking Few: a known change went live that week Database sign-ups roughly stable Compare database and attributed counts
Competitor discount pulled riders Several: awareness, switching and timing Database sign-ups fall too, most where rival operates Split sign-ups by neighborhood
Ad fatigue on paid social Some: creative wore out this exact week Falling click-through before the drop Check creative frequency and click-through trend

Illustrative scenario, not data from a real company. Explanations are listed in the order the razor suggests testing them, not ranked by how likely they are to be true. More than one can be true at once.

Where it breaks

  • Simplest is not automatically true. The razor ranks where to start, not which explanation wins.
  • It needs comparable fit. A simple story that leaves facts unexplained does not beat a complex one that explains them.
  • Real causes are often multiple. A tracking fault and a competitor launch can coincide; stopping at the first cause found can hide the second.
  • "Simple" can be in the eye of the beholder. The explanation that fits your existing beliefs can feel simplest; write assumptions down to keep yourself honest.
  • Ethics: using the razor to dismiss a customer complaint or a harm report as "probably nothing" without checking is a misuse, not a method.

What if you apply it incorrectly?

A subscription box company sees churn rise and settles on the simplest story, a recent price increase, without checking anything else. It rolls the price back. Churn keeps rising, because the real driver was a courier change that made deliveries late in two regions. The company has now lost margin and months, and still has the delivery problem.

Myth versus evidence

The social-video version

The simplest explanation is usually the correct one.

What the evidence supports

Occam's Razor prefers the explanation with fewer assumptions only when explanations fit the evidence comparably well; it tells you what to test first, not what is true.

Five-minute field assignment

Take the last metric drop or spike your team discussed. Write three possible explanations, the assumptions each needs and what each predicts in a second, independent data source. Then check that source. Did the explanation the team acted on survive?

Checkpoint

Two questions: one about the idea, one about a decision. Answer both correctly to master this law. Questions are drawn from a larger bank and graded on our server; a retry brings different questions.

Evidence ledger

Every factual claim in this lesson, where it comes from, and how far it has been checked.

Further reading

Which other law could explain this decision?

Editorial review: Ali Dezhami, November 8, 2026. Last substantive update: November 8, 2026. Our methodology