AAT - Advertising adstock theory

Advertising adstock theory was first introduced by Simon Broadbent in the Journal of Market Research Society in 1979. Back then he wanted to explain how TV advertising works in real life and now I want to share some of the conclusions of AAT and marketing implications for brands in the 21. century.

The basic idea of advertising is simple. It should take potential prospects from point A to point B. Point B is our final goal and it can vary depending on the market and/or brand context but when talking about TV it is most often related to the raise of awareness of a brand.

AAT theory suggests that there is a delay of this trip from point A to point B. So, the speed of getting consumers from point A to point B is not immediate but in fact, it has a lag. Furthermore, this lag comes with diminishing returns, as with each incremental rise of advertising, the rise of effectiveness (awareness/sales) is progressively lesser. So, let's touch each point in a bit more detail.


Carry-over effect of advertising should be self-explanatory but is sometimes forgotten. The effect of advertising is not immediate. After John watched a car commercial he does not go running to his local car dealership to buy a car and after Hana reads a PR article about a new shoe line she does not forget about it a day after. This is where the advertising carry-over comes to life and adstock theory helps us measure this. The simple way of measuring adstock is the following (Broadbent formula):

At = Tt + λAt-1 t=1,...., n

Tt is a measure of advertising weight (e.g. GRP) in time t. λ is a memory decay speed and At-1 is advertising weight in t-1. 

Decay speed is represented by a half-life of an ad which basically means the speed of an ad to go to 50% of its original impact. So, if an ad has 4 week half life it takes 4 weeks for half of the impact (e.g.awareness) to be gone. Every ad copy has its own unique half-life but it can be generalized with the following media channels (Average media half lives - MXM analytics):

TV - 2-6 weeks half life

Radio - 1-5 weeks half life

Newspaper - 2-3 weeks half life

Print magazines - 4-8 weeks half life

Online (upper funnel) - 2-4 weeks half life

Online (lower funnel) - 1-2 weeks half life

The different half-life for the same media is mostly due to different industry and creative.


As we build our campaign and invest more impressions the effectiveness falls. After a certain range of impressions, reach and awareness there is no point of investing more as incremental advertising is getting less and less effective. 

Reach build up related to GRPs - reach x axis and GRPs y axis 

Reach build up related to GRPs - reach x axis and GRPs y axis 

To give you and example here is data from one random TV campaign. In the first 100 GRPs the campaign achieved nearly 30 reach points and in the range from 500 to 600 GRPs the campaign achieved roughly 5 reach points. As 100 GRPs are equally expensive during the whole campaign every incremental reach point in the latter stages of a campaign is more and more expensive.


a)MEDIA PLANNING - brands should understand and use AAT in planning. Find a point where there is no point in investing more in TV advertising and then use online to get light TV users and get cheaper and more effective incremental reach.

b)DATA - adstock models have to be tested to find the best fit model. Simpler models are popular because they are easier to implement and less expensive, still interpretation of the data is as important as the formula.

c)SALE NUMBERS - Advertising sells, but this effect is not seen immediately. It can happen more slowly depending on the brand and industry context. This is especially true for industries where the consumer journey is long (car and insurance), as the advertising effect can be seen 8 months after the start of the search.