Is AI Media Buying Efficient Or Just Opaque?
AI media buying does what it is asked. That is the strongest argument for it and the whole of the problem with it, because the asking is done in a metric and the business runs on something else.
Automated bidding, budget allocation and creative rotation genuinely outperform manual management on the measures they optimise. Nobody serious disputes that. The dispute is about what happens to everything the measure does not contain.
Where AI Media Buying Clearly Wins
Speed and scale. No human team can evaluate the number of combinations an optimiser evaluates, or react as quickly to a shift in cost. On a well-specified objective with clean feedback, it is not close.
The category where this works best is the one where the outcome is immediate, unambiguous and directly attributable. That describes a narrower slice of advertising than the tools are sold for.
What Opacity Actually Costs
The cost is not fraud, it is the inability to learn. A campaign that worked without anyone knowing why cannot inform the next one, so each cycle starts from the same place and the accumulated judgement that used to sit in a planning team does not accumulate anywhere.
An optimiser will find the cheapest way to hit your number. It has no opinion about whether that number was worth hitting.
The Question Worth Asking A Platform
Not how the model works, which no vendor will usefully answer. Ask instead what it was told to maximise, over what window, and what it is permitted to trade away to get there.
Those three answers explain most of the gap between an efficient campaign and an effective one, and they are answerable without disclosing anything proprietary. A platform that will not answer them is not protecting its model. It is protecting the fact that nobody set the objective carefully. More on Ad Tribe The AI Shift.
Advertisement
