You can disable interest based ads.

Regarding to your other message on the subject, contextual ads, "might" have lower CPC, but they "might" have higher CTR.

It's hard to predict consequences of disabling interest based ads. In one hand, showing ads more relevant to your content might be a good complement to what people are searching , when they reach your site. In the other hand, interest based ads are showing (supposedly) ads which are relevant to the center of interest of each visitor, which should also lead to more interaction with the ad.

It's hard to predict consequences of disabling interest based ads.

No it is no hard to predict. It very easy. RPM will fall off a cliff. Ads are shown on an auction basis. Reducing demand by blocking a large chunk of your demand can only have one effect and that is a drop in RPM and earnings in general. One may see an increase in CTR but even there I would have my doubts, and that increase in CTR will not make up for the overall loss in revenue.

Try it.

Just FYI, you can set it up as an experiment in Optimization > Experiments > Ad Serving > uncheck Interest-Based Ads

@LuckyD great idea I'm trying it just to see.

Be sure to give it enough time to work. At least two weeks, preferably a month or two.

Just FYI, you can set it up as an experiment in Optimization > Experiments > Ad Serving > uncheck Interest-Based Ads

Is that option only available for certain publishers? I went to optimization > experiments but did not see an option for "ad serving". I had three options available - "ad unit settings," "allow and block ads" and "ad balance."

Use Allow and Block Ads,from there you can select the option to adjust the targeting type.

BTW - After day one I'm down 37%.

I killed the experiment as it continued to be down at 30% and it was impacting about 50% of my total impression. The reason for the experiment was to prove my point, and it has at a minimum proved it to me.
Conclusion: Blocking interest based ads will result in lower earnings.

Killed mine too, down 17% for 50% of my traffic. Results were significant enough to terminate the test.

@Grapetimes, here you have two examples of what *can* happen, but I still suggest trying it yourself