This does not surprise me. A few things that I know about the ad business:
- There are a lot of players in the market
- Most of them oversell their ability to actually target effectively; I actually know some sales people in this space. Bla bla, machine learning, bla bla bla algorithms, bla bla bla smoke and mirrors.
- Especially the smaller players tend to not have usable profiles on the vast majority of users for reasons of not having existed long enough or not having enough customers to have actually captured enough relevant data.
- Any new ad company has to fake it for quite some time until they actually have enough data. And with GDPR, that data is now a lot harder to come by legally.
- Some of ad companies are fraudulent in the sense that they overcharge their customers for clicks that never happened. E.g. bot traffic is a big revenue driver for ad providers and most of them conveniently can't tell the difference between a bot and a user they supposedly profiled.
So, what just happened is that the NYT cut off most of the worst offenders in this space and ended up with better quality ad providers with better conversions (even without profiling).
Profiling is actually only needed if you have lots of ads competing for the same space. If you reduce the number of ads, the need for profiling goes away. Also, you compensate for bad profiling this way since more (random) people will see your ad. So previously under-performing ads might actually benefit from being shown to random people as opposed to some silly algorithm that uses bad/incomplete profile data to take the wrong decisions.
So what the NYT figured out is that they are better served by a small number of high value ads shown to random people than a great many low quality ads from low quality providers targeted to a handful of their users.
Targeting still has a place in this market but it needs to be consensual; which is going to be a tough sell to end users.
- There are a lot of players in the market
- Most of them oversell their ability to actually target effectively; I actually know some sales people in this space. Bla bla, machine learning, bla bla bla algorithms, bla bla bla smoke and mirrors.
- Especially the smaller players tend to not have usable profiles on the vast majority of users for reasons of not having existed long enough or not having enough customers to have actually captured enough relevant data.
- Any new ad company has to fake it for quite some time until they actually have enough data. And with GDPR, that data is now a lot harder to come by legally.
- Some of ad companies are fraudulent in the sense that they overcharge their customers for clicks that never happened. E.g. bot traffic is a big revenue driver for ad providers and most of them conveniently can't tell the difference between a bot and a user they supposedly profiled.
So, what just happened is that the NYT cut off most of the worst offenders in this space and ended up with better quality ad providers with better conversions (even without profiling).
Profiling is actually only needed if you have lots of ads competing for the same space. If you reduce the number of ads, the need for profiling goes away. Also, you compensate for bad profiling this way since more (random) people will see your ad. So previously under-performing ads might actually benefit from being shown to random people as opposed to some silly algorithm that uses bad/incomplete profile data to take the wrong decisions.
So what the NYT figured out is that they are better served by a small number of high value ads shown to random people than a great many low quality ads from low quality providers targeted to a handful of their users.
Targeting still has a place in this market but it needs to be consensual; which is going to be a tough sell to end users.