Skip to main content

Marketing measurement glossary

Adstock

Adstock is a modeling transformation that represents the possibility that advertising has an effect beyond the period in which it ran.

What does adstock mean?

Adstock describes modeled advertising carryover. A campaign that runs this week may influence awareness, consideration or demand in later weeks, even after spend stops. An adstock transformation carries some prior-period media forward with a decreasing weight so a media mix model can test delayed response instead of assuming every effect starts and ends immediately.

How carryover and decay work

A simple conceptual version is:

Adstock today = media today + decay × adstock yesterday

The decay rate controls how much of the prior effect is retained. A lower rate fades quickly; a higher rate retains more history. This is a model parameter to evaluate, not a universal property of a channel. Different models may also use lag windows, geometric decay or other transformations.

How to interpret adstock

Short carryover

The model places more weight on recent media and suggests the relationship fades relatively quickly.

Longer carryover

The model retains more prior-period media, which may fit a delayed response pattern over the chosen time grain.

Uncertainty matters

A fitted carryover pattern can be unstable when the data has little variation, channels move together or the time window is too short.

Adstock is not proof of causality

A longer modeled decay does not prove that advertising alone caused later sales. Promotions, seasonality, brand demand, distribution and other factors may move at the same time. Treat adstock as an assumption that helps represent timing inside a model, then compare specifications and validate important decisions with appropriate experiments or other causal evidence.

Adstock versus saturation

Saturation describes how response may flatten as investment increases. Adstock describes how an effect may carry over through time. They answer different questions and may be used together in Marketing Mix Modeling: one changes the time profile of media, while the other changes the response shape.

Why adstock matters for budget decisions

Carryover assumptions affect when a team attributes response, how it reads a recent spend change and how it frames a forecast. If a channel appears to be performing after a campaign ends, the team should ask whether delayed media response, demand timing or another factor explains the pattern. Hypermacx uses this kind of evidence as part of a broader Marketing Mix Modeling workflow, alongside contribution, saturation, marginal return and validation planning.

Common mistakes

Treating decay as a fixed fact

Carryover depends on channel, creative, market, time grain and data quality; it should be tested rather than copied from a benchmark.

Confusing timing with impact

A delayed relationship describes when an effect may appear, not how large or causal the effect must be.

Ignoring the business context

Promotions, stock, pricing and seasonality can explain timing patterns that a media-only interpretation misses.