[orangeoxidation]

> Advertisers are measuring the actual effectiveness of their ads, and paying for that only. Therefore falling effectiveness means falling prices and you have to increase the quantity of ads, to earn the same money (and achieve the same effect) as before.

Note there are significant uncertainties in measuring the effectiveness of ads. Measuring click-through rate is comparatively easy, measuring things like brand-recognition doable, measuring how much a specific ad or campaign influences purchasing decisions is quite a hurdle. Predicting it even harder.

There have been success stories, but online, especially "personalized", ads might have been somewhat of a bubble. Appearing better than they are?

[hooby]

It doesn't have to be super accurate or super certain, for advertisers to catch a general trend.

Even if the price adjustments lag significantly behind the actual decrease in effectiveness - the loss of effectiveness will still drive down prices.

As for personalized ads - I think the worst is still to come. Right now, most it is just about looking at a person's profile to decide which ad to show to them. But we are getting close to ML being used to actually create individual ads, optimized for each respective viewer.