Adsense's new Ad Balance slider

[Evan Salamanca]

Anyone notice this yet? You can change the balance of ads that show kind of like FBAN's balancer on steroids. It even shows you the percentage of ads that show versus percentage of potential income. If it weren't for the fact that the ads units show as blank rather than collapsing entirely, this would sound like the best thing ever. Although this creates huge potential when combined with good backup ads.

[anefarious1]

I think a big chuck of the audience that will see any backup, as a result of the so-called slider, are folks not in English speaking countries.

[Evan Salamanca]

A nefarious: It's not a "so-called slider" it's a slider. And yes whenever the ad doesn't show, you will get the backup ad. First you have to go to the ad-unit editing area. For backup ads you can either choose show blank space, fill with a solid color or enter a backup ad URL. The URL should be a separate HTML page with either ad code from another network or an in-house ad on it. You need to manually create the HTML page before you can use it as a backup.

And yes, you will be able to go to your own website and see the backup ads.

[Dimitri]

Since my backup ads are serving different ads depending of visitor's countries, i can see that backup ads are served in rather equal proportion to English speaking countries and others. (My site's content targets USA/UK/Australia, and shows 3 adsense units per page).

I think there are plenty of factors which can explain this, and not only low inventory.

For example, let's take the example of someone who visits several sites on same theme in a row (which is possible, if you are seeking information about a given subject). If Google already served ads to him at these other sites, when he reaches "your" site, Google might consider it's not of use to serve these ads "again".

Also, with Google's tracking system, lot of ads are based on what a visitor uses to do online, which can have an impact about ads served.

Also, it was not rare to experiment seeing the same ad in two ad slots, with the ads balance slider, I am not longer seeing this.

etc...

[surfgatinho]

Does using the slider effect the "Active view viewable" figures?
I'm guessing not as this metric is only for served ads.

[IanTurner]

Yes it effects Active View Viewable figures as the served ads tend to be the ones higher up the page. My overall active view viewable has gone from 32% to 45% from implementing the Ad Balancer (I have pretty harsh settings as only showing 55% of possible ads at the moment)

[EditorialGuy]

I think a big chuck of the audience that will see any backup, as a result of the so-called slider, are folks not in English speaking countries.

That's certainly a possibility, and--to borrow a phrase from Seinfeld--"There's nothing wrong with that." We earn the same amount from an affiliate transaction whether the reader is in New York City or New Delhi. That can't be said for AdSense ads.

[EditorialGuy]

Follow-up:

We started using the "Ad Balance" slider just over a week ago. Today, AdSense is reporting a slight increase in total earnings for the past seven days, with significant increases in page RPM and impression RPM. That's with only 26 percent of potential ads displaying (26/100, according to the slider).

Affiliate revenue is also up, and I suspect that being able to run house ads in the other 74 percent of AdSense impressions is one of the reasons why.

As I said before, I think the "Ad Balance" slider may be one of the biggest improvements to AdSense that Google has made in a long time.

[NickMNS]

I am going to repeat myself... Do not rely on RPM or CTR to judge the effectiveness of the ad-balancer. These metrics are heavily biased by this change, specifically that the goal of the ad balancer is to eliminate impressions that do not earn any revenue. Thus without any change in revenue, the ad-balancer will cause the RPM and CTR to rise sharply.

To judge the effectiveness you should compare revenue directly. Or you can adjust the RPM to include the impressions that were suppressed by the ad balancer.

Example using Editorial Guy's 74% coverage rate:
Assume you had an impression RPM of $2.00 for the period when using the ad-balancer and the balancer only showed ads 74% of the time. That is 2$ for 1000 impressions, where 1000 impressions is equal to 74% of all impressions. Thus 1000/0.74 = 1351 on an adjusted basis you earned $2.00 on 1351 impression or an adjusted RPM of $1.48.

To the astute reader this is equal to $2.00 * 0.74% = $1.48. Now if you compare $1.48 to the previous period you will have a more robust measure.

Also remember that there are many other factors beyond the ad-balancer that can influence RPM, CTR and most importantly revenue.

[EditorialGuy]

NickMNS: Of course a change in the "Ad Balance" slider will influence RPM and CTR. And of course total revenue is what counts (as I've argued for years).

As I stated in my post, using the slider to discard bottom-feeding ads has

1) Boosted our AdSense metrics and total AdSense revenue.

2) Freed up wasted ad impressions for more profitable uses.

It's a twin win.

[NickMNS]

Don't get me wrong, I'm fully in agreement with you "win, win".

I just see it here and in other places too where people look at a single metric and use it to make all kind of outlandish and unsubstantiatable (is this even a word?) claims. This not to say that RPM or CTR are bad metrics, one simply needs to use them in the right context. One needs to understand how metrics are calculated and what their limitation are in order to use them to make informed decisions. Otherwise we can just go and use tarot cards.

BTW if you intend on using tarot readers, please go to my site there are plenty ads for them... Actually no, not anymore thanks to the ad-balancer.

[ember]

I look at all kinds of metrics. In the end, the one that matters is the revenue. Since the ad balancer, revenue is up. It is starting to tick down a little, so maybe there is an initial bounce and then it settles back down to where it was, especially is you get a lot of repeat visitors. Only time will tell. Even if revenue does decrease, I am very glad my readers do not have to see those ugly bottom feeder ads.

[romerome]

Speaking of metrics. It looks like before ad balancer my page views in adsense and analytics matched up. After turning on the ad balancer they still match up. Is that what other people are seeing?

So it seems page views are counted if ads fire or not.

[Evan Salamanca]

That's what it seems like on my end, but I run link units also which are not affected by the slider. So if you don't run link units, page RPM could be altered.

[IanTurner]

Have you set the Ad balance so it has to show less than one ad per page (i.e. below 33%if you have 3 ads per page)?

I just created a coverage report by Ad Unit which shows me which units are getting the best fill rates - so it gives a good indication of which ad units are your best performing units.

[Dimitri]

I am not sure the % fill rate, is related to the number of ad slots which are going to be filled. I have a 45% fill rate, and often I see my 3 ad spots filled by Adsense. I believe this is more complex than that, but I might be wrong :-)

[IanTurner]

I am not sure the % fill rate, is related to the number of ad slots which are going to be filled

Not on any particular page but overall it will be the percentage of the total number of ad slots requested - thus if a particular slot gets filled more it is probably a better ad slot and those that get filled less will be your weaker ad slots.

There will be pages when no ad slot gets filled I'm sure.

[surfgatinho]

How do they know which ad slots are under performing? On some of my sites I use a single responsive ad unit it in different places, Plus some of my ads are in great spots for mobile, but not so good for desktop (and vice versa).

I'm sure they have all this figured, but it would be quite involved to work it all out...

[magician]

I have 2 medium level US information websites (one in healthcare and one in financial services). I have set 75%ads-100%revenue on ad balance. After that, healthcare website which already had great RPM shown 70% coverage and its earnings has actually improved further (with much better RPM). The financial website is now witnessing about 94% coverage with no apparent change in earnings or RPM - it's RPM is about half of healthcare website. I was expecting it to work the other way round but this is what I have witnessed.

[Lagonda]

Now, there's a problem.

This isn't a "set and forget" feature.
The last time I tuned the slider (a week ago), it was 98%/51%.

Wellllllll... today I get 100% at 38% fill rate. And, I suppose, that ratio will change without a warning, all the time, right?
You get a big national programmatic campaign, where the advertiser wants all eyes for him, they get the CPM rate up, correct?

But, I didn't change the slider... am I missing revenue? Because I chose 38% fill rate, earning 100% revenue at that precise moment but, when the aforementioned advertiser boosts the CPM, those numbers will change, and I will be missing some serious cash...

Solutions? Check the slider everyday?
But in that case wouldn't I be a day later? With national campaigns, that's one day of good revenue missed...
Maybe set it to 90% revenue to allow some growth space?

[NickMNS]

@Lagonda I have noticed the same thing. This is my explanation and caveat.

The ad-balancer numbers are based on ones current earnings. When one reduces the coverage and revenue, say move to 75% coverage == 99% revenue, the next day's figures will reflect this change and if the change works as planned your revenue will be 99% of the previous day. But it is no longer the previous day, so now the number shown suggests that you earn 100% at 75% coverage, and it would seem like you could reduce the coverage further if you are willing to accept 99% of your earnings.

Now the caveat, the 99% earnings reported on the second day is not 99% it is 99% of 99%, or 98%. If you readjust the coverage the next time you will check the 99% will be to the power of 3 and so on.

today I get 100% at 38% fill rate

First if you have adjusted things incrementally as described above you may be getting a much lower percentage of your initial earnings than reported by the tool. Second if you have a very flat revenue profile curve then even large changes in coverage will have a limited impact on revenue.

In my opinion the goal of this tool is to control the quality of ads that appear on your site. So one needs to monitor that, if one notices an increase in low quality ads then reduce the coverage, if the ad quality seems good either leave it be or maybe boost the coverage a notch. I would only rely on the percent revenue figure as a rough guide.

[NickMNS]

Is it possible that your backup ads influence the non-backup ads that are being served on the page.

My back-up ad is an ad for another project/site I am working on and it is in a completely unrelated topic to the host site. The topic has wide appeal. I am starting to notice a greater number of ads related to the topic of my backup ads. I have no hard data, and it is entirely possible that these ads are appearing regardless of the backup ads. But it has me wondering.

[james_09]

NickMNS - Thanks for your previous detailed explanation.

I have now set mine at 100% / 75%. It doesn't drop below 100% until I drop to around 50% fill rate, but I am not brave enough to remove half the ads on my site until i understand it a little more.

My bar now has an added section at the bottom which wasn't there before. It is a range between 1% and 46% fill rate which google is calling "a better user experience range".

[NickMNS]

@James just checked my account. Your are right there is this a new bar showing. Very interesting... I'm seeing 0 to 59% which translates to 99% revenue.

I'm also a little nervous about moving things too far down. I think at this point it would be nice to get some direction from Adsense about this feature (ad-balance as a whole). An information video, or a hangout, some kind of explanation on how this tool is intended to be used, what everything really means what are the risk of using or not using the tool.

[magician]

Yeah, it seems G has recently added this new 'better user experience range" - it wasn't there yesterday. I am also seeing 0 to 58% (at 99% revenue) as the better UX range. I have been testing with 75% with 100% revenue since the launch of the feature (earlier 99% was happening at 65%). I am not using any backup ads so not very keen on reducing the ads any further at this point in time.

[IanTurner]

I've currently set mine to 40% (99% revenue) just inside the better user experience range.

My thinking is that if there is any kind of smart pricing going on - then this should be the cure for that. And that if Google has assessed a better user experience range - then it can do this for all sites and is probably also figuring this into the SERPs too.

My revenue has really fluctuated over the last 4 days of using the Ad Balancer - a couple of days really good, one average and one poor. Traffic to my sites has been really good though. I'm persevering with the harsh settings for the moment.

[ChanandlerBong]

Have set to 55/99, just outside the "better user experience" range. Too early to see results, obviously the epm figures are up, but happy to be able to send thousands of ad views to affiliate links and other parts of my sites.

[frankleeceo]

It appears that Better user range is anything from 1%~99% revenue.

It feels that this feature is introduced in part to combat other networks that tout certain CPM while only having a 10%~20% fill rate. Adsense is saying...hey if we only fill you 20% of the time our CPM is more than competitive.

It's pretty smart that it sets a system wide fill rate. Much easier to manage from a publisher's point of view. Rather than a few other networks that let you set % or floor on a tag basis. Those systems quickly get out of hand in complexity.

p.s. I personally do not plan on using this. My traffic is hard to monetize with affiliates so adsense is as good as it gets for me. I call it somewhat lazy on my part, but I find my time better spent on content generation.

If I use it, I would not push it to the absolutely minimum, but maybe 10%~20% above the threshold to take account of short term burst campaigns. Which happens more often than not. I would probably also look at the slider over 7 day period and use that average value. Than routinely check it every month or so to make sure that they're still current.

[IanTurner]

At the moment it looks like the Ad Balancer is updating every couple of days - and I am seeing a short boost in income each times it updates.

[surfgatinho]

Is there anyway to collapse empty ad slots? That would improve the user experience...

I found some reference to this in the Google GPT API but not sure if this is relevant to Adsense in general

[LuckyD]

Side note: just checked my ad balance slider, and, apparently, I can still earn 50% of my current revenue with just 3% of impressions.

Never thought that the top 3% of my ads would be that valuable