We need to rethink the concept of Adstock for Gen AI Ads
OpenAI AI and Perplexity have been experimenting placing of ads in their platforms. I think it is only a matter of few months when the ads will be fully rolled out.
With ad placements comes the question of efficacy of the ads and this is where Marketing measurement techniques like MMM comes into the picture.
However MMM can’t be directly used to model the effectiveness of ads in Gen AI.
Why?
Gen AI Ads break one of MMM’s Oldest Assumptions: Adstock
Almost every adstock function we use today is built on a assumption:
▪️ Advertising creates awareness from scratch, its impact decays over time and an over exposure creates a saturation effect.
This assumption made sense for TV, CTV and even digital channels. But Gen AI is different.
📌 Context Rich & Intent Rich
When an ad or a brand appears inside a Gen AI platform, the user is already there with intent, context and a purpose.
They come with intent rich questions like “It is my Dad’s birthday, he loves to run, which Nike shoes are the best for this purpose?”
I have worked in the B2B intent signal space before, it is one of the hardest thing to crack.
Who would have thought the best way to discern an intent is to create a chatbot 😅.
But anyways, back to our problem- in Gen AI, the “memory” of the ad is not being created from zero.
The ad is riding on existing cognitive state + task context.
The carryover is not just temporal, it is also contextual
📌 Why traditional adstock will struggle?
Classic adstock operates like below:
Exposure -> memory -> decay
or
Exposure -> memory -> saturation
But Gen AI exposure behaves differently. The user’s state persists, not the ad. Recall is triggered by task similarity not time.
In other words, Gen AI adstock doesn’t decay like TV or Digital channels.
📌 We may need a different notion of carryover and saturation effect for Gen AI.
For Gen AI ads, adstock may need to account for:
▪️Contextual half-life (how long a problem space persists)
▪️Intent alignment rather than raw frequency
▪️Non-monotonic decay (effects can spike again without new exposure of the ads !!)
▪️State dependent adstock (same exposure ≠ same impact across users, we may need custom adstock for each user !!)
Gen AI is not just a new channel. It is also a new interaction primitive. And new primitives demand new measurement constructs.
Adstock worked because it matched how media ‘kind of’ worked.
We may need to rethink how adstock work in Gen AI.
That’s what we at Aryma Labs are doing.
We are a futuristic company and we already have made some rapid strides on how to capture the adstock effect for Gen AI platforms.
More details soon 😎.
P.S: Image just for illustration purpose.
