...'cause _historically_ contextual advertising doesn't work better than targeted. And contextual targeting is hard, because determining context is hard and it can fail catastrophically (I recall the apocryphal story about a washing machine manufacturer's banner smack dab in the middle of a macabre news story about a toddler abandoned in a clothes dryer. Brand safety FTL.)
Your example on the "car hatches" is a really simple case and there's only so many niche publishers out there catering to specific audiences. That's how you end up with media companies like Vox and Gizmodo. And in terms of scale...well, go look at the Digital Media comscore rankings: https://www.comscore.com/Insights/Rankings
As for "advertisers get higher intent traffic", that simply isn't true. I'm not refuting your intuition 'cause it absolutely sounds right, but in terms of tracking and targeted ads knowing stuff like, "You left product in your checkout cart." is a waaay higher signal of intent.
> As for "advertisers get higher intent traffic", that simply isn't true. I'm not refuting your intuition 'cause it absolutely sounds right, but in terms of tracking and targeted ads knowing stuff like, "You left product in your checkout cart." is a waaay higher signal of intent.
For simple ecommerce flows, sure. But it isn't just intuition, having bought plenty of programmatic inventory over the years, there are perhaps as many non-online-selling advertisers using programmatic as there are straight "something in your cart" retargeters.
We used to get orders of magnitude higher conversions from websites where we did a direct buy aligned completely with the site's audience. The reason we didn't do more was scale - you can't do direct buys with every site, especially not when you already don't trust your buying agency. If we could, we'd have dropped programmatic entirely.
You also get agencies boosting their retargeting conversion numbers with BS like "view throughs". If your agency is trying to get retargeted view throughs in their hard conversion data, you're being scammed.
(In fact, I "accidentally" left the retargeting pixels switched off in GTM for a month. Made no difference to volumes or conversion rates - the reason they were turned back on? FOMO, plain and simple).
> cause _historically_
I'd argue that there isn't much "history" here to work on. The novelty of ad tracking is still very much there, and brands are still buying (and buying in huge scales, making up for precipitous drops in conversion rates), but when non-ecommerce brands stop being hoodwinked by buying agencies there will be a snap back.
Very few advertisers would want their content associated with a tragic news story like that. Not so directly as on a page with just that story anyway. During a national news cast? Maybe later in the program.
Tagging for the negatives is almost always more important than the positives.
Who tags though, publishers? I mean, you're right; I'm not being a contrarian for the sake of argument.
Contextual targeting probably will have a more interesting pitch in a post-third-party cookie Internet, but it's not going to be _the_ solution. There are already proposals for targeted advertising that may be more compelling/cheaper-to-implement, see TURTLEDOVE: https://github.com/WICG/turtledove.
Your example on the "car hatches" is a really simple case and there's only so many niche publishers out there catering to specific audiences. That's how you end up with media companies like Vox and Gizmodo. And in terms of scale...well, go look at the Digital Media comscore rankings: https://www.comscore.com/Insights/Rankings
As for "advertisers get higher intent traffic", that simply isn't true. I'm not refuting your intuition 'cause it absolutely sounds right, but in terms of tracking and targeted ads knowing stuff like, "You left product in your checkout cart." is a waaay higher signal of intent.