1) When AI is good enough, predicting outcomes become obsolete because the outcome it self is created.
2) Upper funnel advertises don’t care about incremental gains like performance advertisers do. They are incentivized and motivated to not lose what they think they already have (loss aversion). It’s like inverse incremental if that makes sense where ROI can still be calculated/estimated.
I also agree about your points on Data Accuracy (aka Data Quality) x Data Scale, though I'm still quite wary of the Outcomes story. Even though there will be improvements here in identifying and resolving user ad exposures and purchases, there's always going to be pressure to overstate the contribution of exposures in these models even when they're not as incremental.
That's what I see as the danger of the Outcomes obsession and offerings, if they do not place strict guardrails to apply this to only a small fraction of the total marketing portfolio, for small businesses and campaigns that are almost exclusively focused on lower-funnel, DR performance from mostly in-market customers. For mid/upper and full funnel campaigns and marketers, there's a lot more risk here in over-relying on attributed "Outcomes", which will either be non-incremental, or over-prioritize the short-term at the expense of the long-term.
Thanks Erez. Two thoughts come to mind:
1) When AI is good enough, predicting outcomes become obsolete because the outcome it self is created.
2) Upper funnel advertises don’t care about incremental gains like performance advertisers do. They are incentivized and motivated to not lose what they think they already have (loss aversion). It’s like inverse incremental if that makes sense where ROI can still be calculated/estimated.
Thought this was great, giving it a shout in our newsletter tomorrow!
I like the Gravity Theory application here!
I also agree about your points on Data Accuracy (aka Data Quality) x Data Scale, though I'm still quite wary of the Outcomes story. Even though there will be improvements here in identifying and resolving user ad exposures and purchases, there's always going to be pressure to overstate the contribution of exposures in these models even when they're not as incremental.
That's what I see as the danger of the Outcomes obsession and offerings, if they do not place strict guardrails to apply this to only a small fraction of the total marketing portfolio, for small businesses and campaigns that are almost exclusively focused on lower-funnel, DR performance from mostly in-market customers. For mid/upper and full funnel campaigns and marketers, there's a lot more risk here in over-relying on attributed "Outcomes", which will either be non-incremental, or over-prioritize the short-term at the expense of the long-term.