In most marketing meetings, decisions are won by whoever argues with the most confidence, has the most seniority, or remembers the most recent anecdote. “It worked for me at my previous company,” “yesterday’s customer told me,” “I think the audience wants to see.” None of that is evidence; it is opinion with seniority.
Evidence-based marketing proposes something uncomfortable and liberating at the same time: the team’s beliefs—including the manager’s—are hypotheses, and hypotheses are tested before being funded at scale. This is the practical guide to installing it.
The cycle: hypothesis → experiment → evidence → decision
1. Turn opinions into falsifiable hypotheses
A useful hypothesis takes the form of a specific bet: “If we change the campaign message from technical attributes to financial outcome, the landing-page conversion rate will rise by at least 20% in 30 days.” It has a variable, expected direction, minimum magnitude, and timeframe. “We should communicate better” is not a hypothesis; it is well-intentioned vagueness.
2. Design the minimum experiment
The key question: what is the cheapest and fastest way to find out whether this is true? An A/B message test before rebuilding the entire site. A $200,000 campaign in one segment before committing the annual budget. Five customer interviews before launching the new service. The minimum experiment protects the large budget with small bets.
3. Define the success criterion before seeing the data
This is the step that separates evidence from self-deception. If you decide after seeing the results which number was “enough,” you will always find a favorable reading. Write the threshold in advance: which metric, what minimum value, and what timeframe. We cover the biases that operate when this is missing in interpreting customer data correctly.
4. Decide based on the result—and document it
Three possible outcomes: scale (it worked), iterate (promising signal, adjustment, and a new test), or close (it did not work, the learning is recorded, and the budget is released). The experiment log—what was tested, what happened, and what was decided—becomes the area’s intellectual asset: it prevents repeating expensive mistakes and accumulates knowledge of your own market.
System operating rules
- 70/20/10 budget: 70% on proven activities, 20% on optimizing what exists, and 10% on new experiments.
- Maximum 2–3 experiments at once: more than that and no one analyzes anything rigorously.
- Sufficient sample before concluding: no one decides anything from 30 visits; wait for volume or extend the timeframe.
- Failures are presented like successes: an experiment that rules out a bad idea saved money—celebrate it, do not hide it.
The minimum infrastructure
None of this requires a data warehouse. It requires three pieces: reliable baseline metrics (traffic, conversion, cost per lead, sales by source—typically your CRM plus web analytics), the ability to segment audiences for tests, and a dashboard where results are visible without manual heroics. With well-built analytics, an experiment can be read in minutes.
The cultural change: from “I think” to “let’s test it”
The initial resistance is always the same: “this will make us slow.” Experience shows the opposite: evidence-based teams decide faster because discussions become shorter. Faced with two opposing opinions, the response stops being a two-hour meeting and becomes “let’s design the test.” Data ends the debate—and depersonalizes the error: when an idea does not work, the hypothesis failed, not the person.
Management must play by the same rules: if the manager’s ideas are not tested, the system dies in the second week. Data-driven leadership is practiced by example, a topic we expand on in making better decisions using data.
Start this month
Choose one pending marketing decision. Formulate it as a hypothesis, define the success threshold, run the minimum experiment, and decide from the result. One complete cycle teaches more than any training—and the second is already culture in formation.
Want to build the complete system with a method? Book a free diagnosis and let us design your first portfolio of experiments.