Which assumptions are already spending next year’s budget?

Plans often inherit beliefs that no longer receive a proper hearing. Identify the assumptions carrying the most money and consequence, then decide which need fresh evidence before renewal.

Next year’s plan begins with last year’s spreadsheet. Budgets are adjusted, campaign names change and a new target appears at the top. Underneath, the same beliefs continue to direct the money.

The audience wants this offer. The channel reaches the right buyers. The team can deliver the promised experience. Expansion will offset customer losses. Each belief may be reasonable, but each deserves to be visible.

An assumption review asks which parts of the plan depend on conditions the business has not recently examined. It is a practical way to improve the next decision without reopening every settled fact.

Extract the belief from the activity

Take a major planned investment and finish the sentence: “This should work because…” Keep asking until the answer describes a customer behaviour, market condition or operational capability.

“We need the event because it is important” is circular. “We expect this event to create conversations with buyers who cannot be reached efficiently through our existing routes” is an assumption that can be investigated.

Separate different links in the argument. The right people must attend, your team must reach them, the conversation must be useful and the follow-up must happen. Failure at any link changes the case for investment.

The Magenta Book’s evaluation guidance makes assumptions and causal mechanisms explicit when explaining how an intervention should work. A marketing plan can benefit from the same discipline at a proportionate scale.

Distinguish evidence from inherited confidence

Ask where the belief came from. It may rest on a recent experiment, repeated customer observations, an old campaign, an industry convention or a senior person’s experience.

Those sources are not equally strong for every decision. Experience can generate a valuable hypothesis while still requiring fresh evidence when the audience, offer or operating conditions have changed.

Record when the supporting information was collected and what it actually covered. A successful campaign in one segment may not justify expanding the same approach to a different segment without adjustment.

Do not interpret a long history of repeating an activity as proof of its effect. The organisation may have become good at delivering it without learning whether it creates the intended outcome.

Prioritise the assumptions that matter

Focus on beliefs with substantial commercial consequence and meaningful uncertainty. A minor design preference should not consume the same attention as an assumption that determines a large acquisition budget.

Consider how reversible the decision is. A small, easily stopped pilot can proceed with more uncertainty than a long commitment that would be costly to unwind.

Also consider whether evidence can arrive before the choice must be made. An ideal study that finishes after a contract renewal may be less useful than a narrower investigation completed in time.

Choose a small number of priority assumptions. A list of fifty unknowns can make the organisation feel thoughtful while leaving it unable to act on any of them.

Write what would change your mind

For each priority belief, describe an observation that would weaken it and the decision that would follow. This prevents the team from interpreting every result as support for the original plan.

For a hypothetical premium service, the assumption might be that buyers value faster implementation enough to pay a higher price. Positive comments about speed are weaker evidence than an actual choice involving price and delivery terms.

Be careful not to set an impossible standard for disconfirming a favourite idea. “We will reconsider only if every customer rejects it” protects the belief from learning rather than testing it fairly.

Discuss the decision rule before collecting the next round of evidence. The team can revise it if circumstances change, but the reason should be explicit.

Review mistakes without rewriting history

Look at a few decisions that produced disappointing outcomes and ask what information was available at the time. Separate a weak decision process from an unlucky result under genuine uncertainty.

Google’s discussion of blameless postmortems offers a useful principle: examine contributing conditions and corrective actions rather than turning the review into personal blame. The marketing adaptation is to improve how the next choice will be made.

The same discipline applies to successes. A favourable market shift may have helped a campaign whose original assumptions were weak. Winning does not remove the need to understand the mechanism.

Convert the review into next steps

Assign an owner, a decision date and an evidence task to each selected assumption. Some tasks will be analysis, some customer research and some a controlled or staged trial.

Record which assumptions remain unresolved and how the plan accommodates them. A spending cap, staged commitment or fallback option can make uncertainty manageable without pretending it has disappeared.

End with a revised plan, not simply an interesting discussion. Change the allocation, scope or learning sequence where the review justifies it.

The most valuable question is often not what new tactic to add next year. It is which familiar belief has been directing substantial resources without enough recent evidence, and what practical decision becomes better once that belief is examined.