How do I know if marketing advice is good?

Run four tests: what evidence sits underneath it, whose context produced it, who benefits if you follow it, and whether any outcome could prove it wrong. Advice that passes all four still deserves a small, reversible, time-boxed trial against a recorded baseline before it earns real budget.

Why does marketing advice contradict itself?

Because most advice is situated experience stripped of its situation. When two credible practitioners disagree, each is usually reporting a different context — company size, market position, stage, era — with the context left unstated. The contradiction typically marks a hidden variable; find it, then work out which side of it you sit on.

How to evaluate marketing advice

Q: Should I distrust advice from vendors and agencies?

Not automatically — expertise and incentive usually travel together, and the people who know a channel best are often paid to run it. But incentive-aligned advice has not cleared the bar on its own: it needs a second, unconflicted source before it earns belief, and standing suspicion when the recommendation always matches the seller's service.

Q: What makes marketing advice falsifiable?

It predicts something that could visibly fail: do this, expect roughly that, within roughly this horizon. Before adopting advice, write down what success and failure would look like and by when. Advice constructed so that every outcome confirms it — success proves it worked, failure proves you executed badly — is a slogan, not a strategy.

Guide · Frameworks · 4 min read · last verified 2026-07-27

Reviewed before publication Editorial board Independent commercial review

In shortFour tests separate usable marketing advice from autobiography: evidenced or vibes, whose context produced it, who benefits, and is it falsifiable — plus why contradictory advice can both be true, and how to trial it cheaply.

Most marketing advice is autobiography: a report of what once worked, somewhere, for someone, delivered in the grammar of universal law. That is why the discipline seems to contradict itself constantly — post daily, post rarely; gate content, never gate; brand is everything, brand is a luxury. The contradictions are not a scandal. They are what you get when situated experiences are stripped of their situations and sold as rules. Evaluating advice, then, is mostly the craft of putting the situation back. Four tests do most of the work.

Test one: evidenced or vibes?

Ask what actually sits underneath the claim. Documented results with method visible? Named cases you can examine? Or confidence, charisma, and a large following — which are evidence of the advisor's marketing, not of the advice?

Apply the same source discipline you would apply to any research input: primary observation outranks secondhand summary, disclosed methods outrank asserted conclusions, and recency matters in a field where conditions shift under your feet. How to tier your research sources formalizes the hierarchy, and the habit of reading claims skeptically — who measured this, how, and what would count against it — is the same muscle described in How to read AI answer evidence like an analyst.

One structural distortion deserves standing suspicion: advice tends to be published by the people it worked for. The many who tried the identical tactic and failed rarely write the retrospective. What reaches you has been filtered for survival, so even honest advice usually arrives with its failure rate missing.

Test two: whose context produced it?

Every tactic that worked, worked somewhere — inside a specific company size, market maturity, price point, sales motion, team, and moment. Advice is data about the advisor's situation before it is guidance about yours, and its transfer value depends on how much your situation overlaps theirs.

The unglamorous questions do the sorting. Was this a consumer play or a considered B2B purchase? A famous brand harvesting existing demand, or an unknown one creating it? A tactic from an era whose conditions still hold? A team of forty executing, or a founder with a spare afternoon? Advice that fails these questions is not wrong — it is simply about someone else. The costliest advice is rarely bad; it is good advice from a context that does not resemble yours.

Test three: who benefits if you follow it?

Run the incentive audit. An agency recommending the channel it sells, a vendor whose category the advice happens to justify, a creator whose income depends on the tactic staying fashionable, a consultant whose diagnosis always matches their service — none of this proves the advice wrong. Expertise and incentive usually travel together; the people who know channels best are often the people paid to run them. But incentive-aligned advice has not cleared the bar on its own. It needs a second, unconflicted source before it earns belief. Advice is a product; ask what the seller gains from the sale, and notice when the answer is "your dependence."

Test four: is it falsifiable?

Good advice makes a prediction that could visibly fail: do this, expect roughly that, within roughly this horizon. Bad advice is constructed so that no outcome can embarrass it — if you succeed, the advice worked; if you fail, you executed poorly, or stopped too soon, or lacked commitment. If no result could prove it wrong, it is a slogan, not a strategy.

So before adopting anything, write down what success and failure would look like and by when. Advice that resists this exercise — that goes vague exactly when you ask what failure looks like — is telling you what it is. The habit of demanding falsifiable claims from others pairs naturally with making them yourself, which is the core of What is evidence-backed marketing.

Why contradictory advice can both be true

Here is the reframe that makes the contradictions useful instead of maddening. When two credible practitioners give opposite advice, the likeliest explanation is not that one is lying but that each is correctly reporting a different context, with the context left unstated. "Invest in brand" and "focus everything on capturing demand" can both be true — at different stages, budgets, and market positions. The contradiction is a signpost marking a hidden variable.

So treat contradictions as research prompts: find the variable on which the answer flips, then work out which side of it you sit on. Note also what happens when a whole category obeys the same advice at once — tactics that worked partly by being uncommon stop working when universal, and everyone converges into indistinguishability, a dynamic examined in Why B2B brands sound the same. Popularity is not validation. Sometimes it is expiry.

From evaluation to cheap trial

No filter substitutes for contact with your own reality, so the final move is to make trials small enough to survive. Take the advice that passed the four tests, define the outcome and time horizon in writing, record your baseline before you start, then re-measure the same way at the deadline — the locked-benchmark habit that Magrios builds into its re-scans applies just as well to testing borrowed wisdom. Keep the bet reversible; the goal is information, not commitment.

Then let your own results outrank everyone. The purpose of evaluating marketing advice was never to find the guru worth obeying. It was to need one less and less: borrowed judgment is scaffolding, and the structure it is supposed to support is your own evidence about your own market.