I think you may be right about taking a while to fix.

... the Adwords team got pissed off at Adsense's team for tanking half of their ads

Yeah. I was wondering about that.
Could this be an internal political issue?
To be decided by economics (shareholders)?
I'm staying tuned.
.

You set it at 20% fill and get 50% fill

I know, I've been told several time I was wrong, but I still believe that "fill rate" is inappropriate word used by Adsense.

A 50% fill rate, means ads are displayed one time in two.

But I do not think this is it. My understanding, is that "50%" means that Adsense displays only the 50% of the most paying/successful ads / advertisers for your account. So for a given request (page/country/time of day/profile of visitor/etc...), Adsense selects an ad to display. If this ad is in the pool of your most performing one, it's displayed, otherwise not.

... the Adwords team got pissed off at Adsense's team for tanking half of their ads

I disagree, what the ad-balancer does is reduce the supply, this in turn pushes up prices. This is beneficial to both AdWords and publishers, and the cost is born by the advertisers. But it is the large segment of advertisers that are essentially freeloading that will pay the biggest cost. By eliminating what are essentially spammy annoying ads even the legitimate advertisers win as their ads will get seen more and users will not be blinded by the spam.

A 50% fill rate, means ads are displayed one time in two.

But I do not think this is it. My understanding, is that "50%" means that Adsense displays only the 50% of the most paying/successful ads

It is both, a 50% fill rate, means exactly that, ads are displayed one time in two times. But the ads are not selected randomly, instead ads are ranked by those that pay the most and only the top 50% are shown (see revenue profile report in Adsense).

The problem is that you can't rank the ads by price until after the ads are displayed. So there is some sort of a prediction algo that is used to guess what ads will perform which will not. This prediction is what causes the variation around 50%. In my experience it has been pretty good and staying near the target number.

If it pushes prices too far up though, it risks losing advertisers to FBAN, Taboola and other major ad providers. That might be their issue.

Hi folks,

I have no timeline just yet on when this will be back up and running. I am aware that there is some fine tuning going on, and once the engineers feel confident in that update we would then send confirmations to publishers.

I will make a point of reaching out proactively if I learn of anything as well.

Sorry for the delays/trouble.

JB

Thank you JB !

@pubpolicycomms thanks for the info.
Are there any features being added, or was this simply maintenance?

Thanks for the update, JB, it's greatly appreciated.

@NickMNS, I believe the update is simply with the algorithm to better fine tune revenue estimations. I have no insight into additional features added at this point.

@pubpolicycomms thanks again.

One thing that has been missing from the roll-out of the ad-balancer is documentation explaining how the algorithm works (broadly speaking). It would be great if we could get information about how the percentage of suppressed ads is calculated and over what time period as it would allow us to better gauge and optimize our performance.

@pubpolicycomms - thanks for the update.

A suggestion that I'd like to make, is that the ad-balancer be made adjustable on a per-site rather than per-account basis.
I have one site that is very different to the rest of my sites and it would benefit from different settings in ad-balancer. I think that on a site with a high percentage of return visitors, this could be a powerful tool to combat ad-blindness, however I can't utilise it because it would negatively affect my other sites.

@glitterball - one word... channels.

Some feedback.

Of course it's impossible for me to correctly correlate because of so many factors involved (seasonality, economy, advertiser bidding, etc.) but, in my case, Ad balance DOWN, revenue UP, by some significant margin (>10%).

It's been a long time now, that the Ad balance is down. I wonder what is wrong with it. If I don't make mistake, before going live, it's been months that it was available in beta. It's too bad, because it looks like a great improvement.

Obviously it was not working as intended. Those that raved about it doing wonders may have been motivated by the placebo effect.

Those that raved about it doing wonders may have been motivated by the placebo effect.

I assure you there was no placebo affect. The ad balancer is the only thing that was able to eliminate all the spammy ads, such as download buttons, horoscopes (despite blocking the sensitive cat), cider vinegar ads and a whole lot of other crap. Whether or not it increased, reduced or changed nothing in terms of earnings is hard if not impossible to say, but it didn't hurt.

Since mid May when the feature stopped working my site is once again flooded with crap, and I am back to spending hours reviewing ads in the hope of mitigating any negative impact of showing my users this spam. Most of which comes from Ad-Words.

Now the question is when will it be back up and running? Or, is the question will it ever be back?

The other question is, is it worth it to set up a DFP account to use DFP to achieve the same goal as Ad-balancer?

Let me not leave out the other big change, that is ad-blocking in Chrome and the associated Ad Experience Report in GSC. The "poor experience" ads are the ads that were previously being blocked with the ad-balancer, now without the feature these are impossible to block. So what comes next, we get all are ads blocked by Chrome. Because A- Adwords allows the spam, B- Adsense provides no effective tools for blocking it. As you can probably tell I am little p'ved by all this.

The numbers never matched up. You set it at 20% fill and get 50% fill.

I wasn't seeing that. Then again, I wasn't sitting at every computer in the world. :-)

The Ad Balancer was working great for us. I had it set at 12% fill/100% income, with the remaining 88% of impressions being used for house ads that boosted affiliate bookings. The results were excellent.

My only complaint or reservation was that the ad balancer settings were for our entire account. The balancer would have been even more useful if it could have been set for each ad unit or channel.

Very annoying - i hope it comes back.
I was one of those raving about it - earnings up significantly.
Then at the end of May earnings were back to what they had been - i only recently saw the ad balance message and it looks pretty clear to me that this is the cause.

If the ad balance is just a matter of displaying ads XX % of the time, this is something you can simulate by showing adsense code XX % of the time, and displaying something else in the remaining YY %

@Peter_S it is not that simple. The algo ranks the ads based on potential earnings, then if you set it at 50% it will only display the top 50% of the ads by potential earnings. Simply blocking 50% of the ads without taking earnings potential into account should cause a 50% drop in earnings. Probably a little less as you are limiting supply, for the same demand. So you should see an increase in price, but not enough to make up for the 50% fewer impressions.

Addendum to what NickMNS just said:

By restricting ads selectively (i.e., displaying ads with the highest revenue potential for your pages), the Ad Balancer can also improve the quality of the ads and the overall user experience.

For example, if you've got a site about Biblical scholarship, the Ad Balancer (if used aggressively) should result in more ads for Bibles and fewer remnant-style ads for dating services, Windows downloads, and weight-loss products. That's good for the reader, and besides helping you to make the most productive use of your screen real estate, it improves the perceived quality of your site.

Although i'm a long time adsense publisher i've tended not to micromanage and there's a whole load of functionality of adsense that i may have missed. I think i have optimum ad types and placements on page. Ad balancer dramatically improved my income (ie +100%) which dropped around the time that the tool went offline. So my question is this - in what ways, if any, can we mimic the effects of the ad balancer using other settings?

@nickreynolds, with plain vanilla Adsense you can't mimic the ad-balancer. The only way that I can think of is using DFP (DoubleClick For Publishers) where you can set a price floor, under which no Adsense ads will be displayed. So you set the floor at a price of say $0.10, then only ads that win the auction for more that $0.10 will be displayed, otherwise some other back-up ad is shown. The danger is that you set the price too high and then no ads are shown or too low and no ads are blocked. And since demand changes a good floor price one day can be too high for the next day.

The service was created for publishers that sell ads direct, then you can show the direct ad for up to floor, and anything else goes to adsense.

The problem is, it seems complicated to set up. It requires creating new accounts and then changing ad code and so on. So as long as the ad-balance is not declared dead, I am just going to wait it out. But if it is killed I will then set up a DFP account. I was actually considering doing it in January, but then they announced ad-balancer.

I'm not sure Adsense realized what a great and powerful feature the ad-balancer was before they launched it.

Just wondering... Do you think the Ad-Balancer may also "look" at pages with ads that receive little to no clicks on the ads, and figure it's not worth displaying any ads on this/these pages as they never get clicked.

Anyone here still under the impression ad balance in coming back?

@Sal Collaziano I also have the feeling that it isn't coming back. BUT! We have had assurances from @pubpolicycomms (Adsense employee - WW-account) that their engineers are working on it, and in Adsense it still shows under maintenance so rationally there is no reason to believe that it wont be back. It sure would have been nice if they could provide some sort of a time line. If it is going to take another 4 to 6 month then it would be worth a looking at implementing a work around such as DFP. But not knowing anything leaves us in the learch, I could spend the time and money to set up a DFP account and then implement the code only to have ad balancer starting working tomorrow.

@No5needinput

Do you think the Ad-Balancer may also "look" at pages with ads that receive little to no clicks on the ads,

I don't think that the Ad-Balance takes that into account, but pages with little to no clicks would be the most likely to be impacted, in that advertisers are the least likely bid for ad space on those pages, thus the bids for those pages would be low and subject to being suppressed by the ad-balancer.

Yeah, I wish they'd tell us something. I liked the feature. I had a feeling, though, about two or three days after it went away that it wasn't coming back...

You can always send them feedback, the send feedback button still appears under the under maintenance message. I have, I would encourage everyone to the same. In a worst case it will do nothing, but it may get them to move a little faster or at least keep us informed.

The impression counter must had an impressive nosedive... and that, will surely raise some flags to someone.

Now, inside Google, there's a battle, such as the one that existed years ago, when AdSense was allowed and incentivised in parked domains.

In the end, the user experience won. Will it win now?

You see, I told you that they were testing new ads that look like matched content ads (Native ads).

Msg#:4851757
4:54 pm on May 31, 2017 (gmt 0)
My guess is that they are doing some testing with some strange ads that look like the matched content ads and they dont won't us to block them.