Each entry defines a term in one sentence, then says what it changes in a media plan and where this site goes further: the pillar, the method sheet or the page for brands. Terms with a page of their own are linked; the others are sections of this page.
An advertising A/B test randomly splits an audience in two and compares outcomes under two versions of an ad or offer, or with and without the advertising.
An A/B test randomly assigns people to two groups and compares their outcomes under different conditions; an advertising holdout, where one group sees no advertising, is one version of this design. It estimates the effect of one tested condition relative to the other, for that audience and spend level, and a platform running the test is measuring inside its own audience and rules. Where the unit is a region rather than a user, the same logic becomes a geo-lift experiment, which works without any user-level data and estimates the effect of the tested advertising change on total sales in that region.
Advertising effectiveness measurement assesses how advertising changes awareness, consideration or sales, from a model, experiments and benchmarks.
AI for ad operations supports tasks such as writing ad copy, monitoring campaigns and adjusting bids, up to an agent acting within limits the team sets.
AI in ad operations ranges from writing ad copy to running bids. The case where the word agent matters is media planning carried out by an agent: software that pursues the plan’s goals over time, observes what the campaigns are doing, proposes or takes the next action within limits the team has set, checks the result and continues. What the numbers are made of matters as much as the automation: the Kuwalyst Media Planner bases its recommendations on a media mix model, incrementality experiments and audited benchmarks rather than platform attribution, with every figure tagged measured, projected or benchmark. The evidence requirements and the three levels of autonomy are explained on the page about agentic media planning.
Brand lift is a campaign's incremental effect on awareness, consideration or another brand outcome, estimated against a comparable group that did not see it.
A cross-media strategy coordinates several media channels, online and offline, in one plan and one measurement, so the budget follows what each adds.
A cross-media strategy can bring television, radio, out-of-home and print together with search and social, in one plan and a shared measurement framework. Platform reporting alone cannot establish what each channel adds: each platform reports within its own tracking and attribution rules, and offline media stay outside them, so a television campaign shows up as an unexplained rise in branded search. Cross-media measurement comes from a media mix model, which reads online and offline channels in one model, weekly, and from experiments that measure what a channel caused. The Media Planner builds the plan on that basis, with cross-platform measurement for online and offline channels alike.
Data-driven marketing is marketing whose decisions rest on evidence, measured or modelled, with every figure showing where it came from.
Incrementality is the part of a result that advertising caused: the sales that would not have happened without it, as opposed to the baseline.
Marketing budget allocation divides a budget across channels and periods; done on evidence, it follows the return each channel's next euro should bring.
Marketing ROI is the return on marketing investment: the incremental profit that marketing generated, after marketing costs, divided by those costs.
Marketing science applies statistical models and experiments to marketing decisions: measuring what marketing causes and deciding where the next euro goes.
The media mix is the set of channels a plan spends on and the share each one gets; media mix modeling measures what each of them contributed.
MMM, marketing mix modeling, is a statistical method that explains sales as a baseline plus the contribution of each marketing channel, from aggregate data.
ROAS, return on ad spend, is revenue divided by advertising spend; platform ROAS uses attributed revenue and describes an average return, not the next euro.