Did you remove all manually placed ads at the same time? In my testing with Auto Ads, I've found it's best to keep 1 or 2 best performing (old) manual units and then add auto ads. In one case I removed all ads, put up Auto ads only then saw a substantial revenue drop. I put back the top ad and left auto ads up and things are OK, pretty much the same performance as before. But the ads are placed in places I can't put them.

Did you remove all manually placed ads at the same time?

No, I kept (and I'm still keeping) two manually-placed ad units per page, I only removed one above-the-fold unit from all pages.
Yet, the RPM of my "old" manual units dropped drastically after enabling AutoAds.

The uptake (and leaving) of Auto Ads has not reached sufficient numbers to actually see any trends. Auto is "too new" and the number of "disheartened leavers" are still yet to be determined.

The uptake (and leaving) of Auto Ads has not reached sufficient numbers to actually see any trends. Auto is "too new" and the number of "disheartened leavers" are still yet to be determined.

OK, then I'll add my two cents and try an experiment. I'll disable AutoAds for some weeks, use only my manually-placed units, and I'll report the results.
My suspicion (for what it worths) is that, if you carefully position your ad units into your content, things won't change much in terms of earnings. Yet, the user experience could possibly improve. We'll see...

There are a lot of threads discussing the pros and cons of Auto Ads. Sometimes it isn't the kind of ads but the choices we make that helps or hurts our success. I suggest using the site Search feature here and just looking through some of the recent discussions. It beats waiting and wondering if things could be better.

This example has a few pages of observations and ideas: [webmasterworld.com...]

There are a lot of threads discussing the pros and cons of Auto Ads.

Thank you not2easy. Yet, I'd seen those posts already. I know that there is a mixed feeling about AutoAds. Someone praised their simple approach while others criticized the effect on users' experience. I'm in the middle, I had been liking the concept behind AutoAds as soon as I realized that it wasn't an evolving environment as I expected it to be. Not an AI-based system capable to understand where to place ads by experience. To date, It has been just a piece of code that places as many ads as possible between paragraphs, period. You may like that or not. After trying it, someone likes it, someone doesn't.
Yet, my question is different. Not if people here like AutoAds or not. But, what happened to your AdSense revenues after you disabled AutoAds, reverting to manually-placed Ad units?
They went up, down, or just remained the same?

[edited by: riccarbi at 9:09 pm (utc) on Dec 30, 2019]

Yes that is a different question. My thought in making the suggestion was that since you hadn't yet disabled the AutoAds that there might be something in related discussions that you might try before disabling them. I do realize that one size will never fit all and I didn't mean to interrupt your goal here.

Any change will introduce its own variations and... while generalities might be possible, it is nearly impossible to correlate that "this" caused "that"... each observation is limited to the one datum only.

Sans a large enough sample all you get is

"well it went up for me" or "it went down for me" or "never saw any difference" and those must be compared to "never used it".

20 of each still won't give a useful answer...

But for g, several billion instances does have a value ... and for g auto is just another method of milking the ad machine... and doing so contrary of near 20 years of g advice on how to place ads on your content.

As far as I'm concerned the jury is still out on whether this auto ads is benefit, no benefit, or anything else. Not enough data.

When these questions are met the overall should come down to USER EXPERIENCE as it is the USER one aims to please ... else there's no traffic at all.

Some have opted out of the g ad system and do well. Others pile on and seem to do well. The vast majority have mixed results with enough getting SOME revenue to keep the system still working.

What should be noted is that g sought fit to offer this new method ... and that might be an indication that (from their side of things) the old was slipping/falling off and this is a way to raise it up.

I just look at g's past performance on inserting new things/guides into webmastering and having g (some 5-10 years later) say "Whoops! We won't be doing that any longer!"

Go from there.

Personally I am rarely a "first adopter" ... need a little time and grade before I make any serious moves that might affect my users and site presentation.

YMMV

.

Personally I am rarely a "first adopter" ... need a little time and grade before I make any serious moves that might affect my users and site presentation.

You are right and much wiser than me ;)
Indeed, after entering one of those g things (such as Auto Ads, AMP, and the like) is usually difficult to go back, mainly because it's not easy to evaluate the consequences beforehand.
I am not happy with AMPs either (and I'm an early adopter of them as well), but I'm hesitant to disable them fearing the negative effect this move can have on ranking and organic traffic. Similarly, I'm not happy with Auto Ads but I don't know if disabling them could be detrimental in the long term (I'm pretty sure it will be so in the short term, as it usually happens when you make a substantial change in your AdSense configuration).
The only thing I can do is to try an experiment, disable Auto Ads for two months and see what'll happen.
Another possibility is to create an A/B experiment, write some code and serve Auto Ads only on a selected set of pages; then compare the RPMs.
This will be not conclusive evidence, I know, since my case is arguably different from someone else's, though it would be better than nothing...

Happy Experimenting! Wishing good results ... or results that are good for YOU!