Bayes' rule
Weigh a belief by how likely the evidence is if it were true.
How we think
The maths behind the marketing.
Every equation on this site and what it means, because marketing is a decision made under uncertainty.
The seven on the walls
From David Sumpter's The Ten Equations That Rule the World, which is where this habit started.
Correlation, the advertising equation
How two attributes move together across people. It says nothing about why.
The betting equation
Win probability as a function of the odds, with fitted parameters alpha and beta.
The confidence equation
After n trials with mean h and spread sigma each, the 95 percent range for the total.
The influencer equation
The steady state of influence is fixed by the structure of the network, not by any one node.
The market equation
A price or metric moves with a signal h, feedback f of its own level, and noise.
The reward equation
An exponentially weighted memory of rewards. Short-term wins can mislead it.
Decision science
What we reach for when the question is where to spend, what to test, and what a result is worth.
Kelly criterion
The share of your bankroll to stake on a bet that pays b to 1 and wins with probability p, so it grows fastest over many bets.
Adstock and saturation
Advertising carries over, then saturates. The core of marketing mix modelling.
Bayesian conversion rate
After s successes and f failures, the whole distribution of the rate rather than a point estimate.
Incremental lift
What the ads caused, measured against a holdout group, not what they touched.
Weighted-sum MCDA
Options scored against weighted criteria, so trade-offs are explicit.
Expected value of perfect information
How much it is worth paying to remove uncertainty before deciding.
Little's law
On average, items in a steady system equal the arrival rate times the average time each spends inside. Queues, pipelines, backlogs.
Learning and AI
Two lines under most of what is called AI in marketing.
Bellman equation
The value of a decision includes the value of the decisions it leads to.
Attention
How a transformer weighs each word against every other. The core step in most large language models.
Try the method
Is this change real?
Your conversion rate went up. Is that enough evidence to change the budget? The example is a landing page test most dashboards would call a win, up by half. Run the numbers and there is about an 80 percent chance it is real, which is not enough to act on. Put your own numbers in.
How it works
Each version gets a curve of plausible true rates, a Beta distribution, from its visitors and conversions. The tool draws from both curves twenty thousand times and counts how often the after version wins. It runs in your browser and nothing is sent anywhere. Formulas it rests on.
The longer version, with the thinking behind each one:
Why there are equations on a marketing websiteSources worth reading. David Sumpter, The Ten Equations That Rule the World. Evan Miller, Formulas for Bayesian A/B Testing, which the test tool above rests on. Evan Miller, How Not To Run an A/B Test, on why peeking at a running test misleads.
Next step
Got a question that needs one of these?
Tell us what you are trying to decide. We will tell you which method fits and what evidence it needs.