A marketing plan can begin with activity — a channel, a campaign, a tool — and then backfill a reason. This brief argues for the reverse order. Write down the decision you are actually trying to make, state what evidence you would need to make it, and only then describe the initiative. The point is not to slow planning down. It is to make the plan falsifiable before money and time are committed.
Why the decision question comes first
A proposal becomes reviewable when it has an owner, a cost, a time constraint and a question attached to it. Without the question, a plan is difficult to argue with, because there is nothing to test: attention drifts to whether the deck is convincing rather than whether the premise is sound.
If the decision question is “should we fund a second content programme this quarter?”, a review has something concrete to resolve. If it is only “do content marketing”, the same meeting has no decision in it, and will likely produce a preference. Preferences are legitimate inputs in marketing, but they should be labelled as such rather than disguised as analysis.
Advertising claims carry a proof burden
Any plan that will eventually make claims about outcome, performance or effectiveness faces a real external standard. The US Federal Trade Commission states that under the FTC Act advertising must be truthful and non-deceptive, that advertisers must have evidence to back up their claims, and that ads cannot be unfair (FTC).
The FTC describes a “reasonable basis” as objective evidence that supports the claim, with the kind of evidence depending on the claim itself (FTC). It also notes that the FTC looks at express and implied claims, and that advertisers must have proof for both (FTC). Two further cautions in the same guidance are worth carrying into marketing plans: statements from satisfied customers usually are not sufficient to support a claim that requires objective evaluation, and a money-back guarantee is not a substitute for substantiation (FTC).
For planning, the inference is direct although the FTC guidance does not address marketing plans as such: if the finished campaign will say something specific, the plan should record what evidence already exists and what would have to be gathered before the claim is used. The FTC also states that its staff cannot clear ads in advance (FTC), so an internal review has to do that work rather than relying on a regulator’s sign-off.
If AI is in the plan, handle the uncertainty explicitly
AI-assisted marketing widens the number of unknowns a decision has to survive. NIST describes its AI Risk Management Framework as intended for voluntary use and to improve the ability to incorporate trustworthiness considerations into the design, development, use and evaluation of AI products, services and systems (NIST). NIST also states the framework is being revised as part of the White House AI Action Plan (NIST) — a reminder that a document’s status is time-bound.
More directly relevant to marketing may be the Generative AI Profile, NIST AI 600-1, released 26 July 2024, which NIST says can help organizations identify unique risks posed by generative AI and proposes actions for risk management aligned with their goals and priorities (NIST). NIST’s technical reports listing describes the profile as centring on 13 risks and more than 400 suggested actions and as produced with a public working group of 2,500 participants (NIST AIRC). The same listing notes the ARIA programme pairs people with AI applications to study behaviours and impacts in scenario-based interactions, and observes that current approaches to evaluation often do not account for real-world risks and impacts (NIST AIRC).
Inference, not a NIST instruction: a decision brief can borrow the shape of that guidance by separating what you know from what you assume, and by naming the risks the initiative would introduce alongside its expected benefit.
Content that meets a decision test
Google states its automated ranking systems are designed to prioritise helpful, reliable information created to benefit people rather than content created to manipulate rankings (Google Search Central). It also states that E-E-A-T is not itself a specific ranking factor, that trust is the most important of those aspects, and that fabrication of creator profiles is treated as deception and a signal of a low-quality page (Google Search Central). Google adds that interactive features such as calculators or online tools can count as main content where they directly help a page achieve its purpose (Google Search Central).
These are statements about Google’s own systems, not a performance promise. The planning implication is modest: describe the asset in terms of what it lets a reader decide, and avoid borrowed authority you cannot substantiate.
Put the decision into a plan
Open the decision brief worksheet and set up to three proposed initiatives side by side, entering only what you can defend: the evidence you hold, the assumptions you are making, cost, owner, time constraints and the question being decided. The worksheet produces no score and promises no outcome; it exports a private Markdown brief with the unknowns left visible. Then read keeping evidence separate from assumptions, comparing initiatives without fake precision and setting a useful review cadence.