How does the adstock and carryover work?
Borrowed from Pharmacokinetics: Geometric adstock: A(t) = x(t) + θ × A(t-1)
Pharmacologists model how a drug dose lingers in the body and fades at a steady rate. Advertising effects follow a similar pattern, which is why marketers borrowed the geometric decay curve.
A higher θ means the effect lasts longer, so you can pulse spend instead of running it every week. A low θ means you need steady presence.
Worked example
A US retailer spends $10,000 in one week with θ = 0.60. The half-life is about 1.4 weeks, and the lifetime effect is 2.5 times the first week, since 1 ÷ (1 - 0.60) = 2.5.
How should I read my result?
Effects fade within a week or two. Keep spend steady, as in search ads and promotions.
Effects carry for several weeks. Alternate flights of spend with short gaps and test the gaps.
Effects last a long time, typical of brand and video. Pulse spend and avoid judging a burst by its first week.
Embed this calculator
Questions about adstock and carryover
How do I find my own θ?
Fit it with a media-mix model on weekly sales and spend data. Until then, test a range and compare.
Is adstock the same as frequency?
No. Frequency counts exposures per person, while adstock describes how the effect of exposure decays over time.
Does adstock apply to search ads?
Search ads usually have a low θ because intent is immediate. Brand and video channels show higher values.