The more I do, the less it works

Sounds familiar to you? it's so obvious they could leave me a note

[explorador]

Hi there webmasters, let's talk specifics (yes I know, not those specifics). I have a network of sites, all white hat, CMS built by me, fast, all original content, all original pictures (lots of it, I'm also a photographer). For years my work has been well positioned, if not by text + pictures, then at least by one of those. Ok, let's talk about some clues, specific stuff.

• Not all the sites have/had Adsense
  
• The traffic of the websites WITH adsense is about 5,000 unique daily on low season
  
• That's about 150,000 monthly unique visitors
  
• Little bounce, many readers stay quite a while online

Responsive #1. A while ago analyzed things and thanks to a forum member saw lots of opportunities for improvement, first responsive adaptation launched and earnings showed only a bit, small increase but steady. Visitors on smartphones were seeing the right ads. Earnings drop. About a year ago earnings went down a bit. Deeper look pointed at needing more updates and full responsive on all sites.

More updates and earnings up. While traffic was steady, posting articles showed increase on earnings the same day and then diluting over the week, it was steady. Solid behavior for a few months (like 6 months) **Then, the earnings didn't show any change despite updates. Earnings frozen, then went down a bit.

- - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - -
The previous is the FIRST change I saw and I can remember: when I did something affecting positively the earnings, then went down. It's not as earnings going back to previous numbers, they went down a bit from there, let's say I was earning US$ 10 per day, the effects of change increased that to US$ 12, then after a period of about 24 hours to 72 hours, earnings went DOWN to a solid US$ 7 per day
- - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - -

As learned, did nothing, just waited for things to settle down, this is such a confusing ecosystem with many variables we have to give it time.

Responsive #2. Launch a full responsive fast update, sites load fast, right ads-size on the sites, full change on sites but managed to keep the urls, redirects, traffic juice, etc. Everything looked ok. There was an earnings increase, it didn't last**Then there comes the ghost again: earnings dropped not to where they were, they went down a bit more. Let's continue with the example, I was let's say on US$ 7 per day, it went up to US$ 9, then after 24 hours to 72 hours went down to US$ 5 per day. As learned, did nothing, just waited for things to settle down avoiding several changes at once so I could see some effects of A over B etc.

Extra website. Added an extra website to the network using adsense. This is a solid traffic website, the same deal: original articles, pictures, organic traffic, etc. Right after placing the ads the income went up, this time a bit more. Here comes the ghost again: earnings went up and then down. Continuing with the past example let's say I was on US$5 per day, it went up to US$6-7, then after 24 hours to 48 this time, it went right down to US$3 per day. SOLID.

Did I paint the picture well enough? Every change causing any increase on earnings is effective only for about 24-48 hours, then earnings go down but not where they were, it gets worse. Sounds too silly to be true? hear this.

Extra changes: blocking ads and networks.After checking the so many weird ads and non wanted stuff, I go blocking some ads and networks. When it seems the house is clean, earnings go up a bit, just a bit. And here comes the ghost again: continuing with the example I Was on US$3 per day, it goes up to US$4-5 per day, then back to US$3 and down to US$2.5, the "punishment" seems lower this time but it's there.

Extra changes: Networks OFF, blocking all networks.Tired of blocking silly ads and unknown weird named networks I go bersek and blocked ALL THE NETWORKS leaving only Google Adsense on. Earnings go up. The ghost comes again, after 24hours the earnings go back to where they were. The last stage had US$2.5 per day, goes to US$3-5 and then back to US$2.5 and stays there

Extra changes: Networks ON, allowing networks. This time I do the opposite: SAME RESULT. Earnings go up a bit and then back to where they were in exactly 24 hours. I decided to play with this, I go on blocking all networks one day, the next I allow all networks, then switch them off, then on, and play like this day after day. Well, magic, earnings go from US$2.5 to US$4 per day, seems solid. The ghost again, it goes down to US$2.5 per day. After just ONE WEEK, no matter what I do: on or off all the networks or blocking the ads... the earnings won't change.

After 15 days to 1 month earnings are even lower, it went down a bit from the (remember hypothetical) US$2.5 to US$ 2.00. Traffic has been up, in fact it goes up year after year, pure organic traffic. The page views on adsense (adding the total traffic) went up, adding an extra site increased this even more, still... while the traffic goes up the earnings go down. I understand this and have no problem dealing with it, but the evident effect of 24 to 72 hours sending the earnings not only back, but even down a bit from where they were seems fishy and silly to me. My sites are stronger, better traffic, more exposure and interaction than ever, the panorama of going direct ads looks better... while on the other side Adsense is working less for me.

This is not a complain, I won't loose my sleep on this, it's just solid to me, the behavior doesn't look good, it seems like constant penalization. Want to speculate? sounds familiar?

[explorador]

For the faint of heart, if you want to avoid the details it' simple: any change (white hat, legit) increasing the earnings from X to X+3, ends up in 24 to 72 hours in an earnings decrease to X-1. SOLID. If you think this doesn't make sense, then go back and read the explanation, it's quite solid.

[Leosghost]

<= not this ghost.. :(

[explorador]

HA HA not a chance, thanks for passing by

[Leosghost]

Question..are you running Ganalytics on these sites ..?

[londrum]

Ive seen similar things with my changes too. I always assume that its because some kind of ad data gets reset everytime you make a significant change. Google has lots of data on your original ad placement, or ad design, or whatever, and then you go and change something and google decides to chuck all that data away and start your ad from scratch again. So you no longer get as many higher paying ads. You have to build up all that 'trust' again, to show the ad can still convert, before you get all those higher paying ads back.

Maybe it takes their system a couple of days to filter through that you made changes, so that's why you get a little boost straight after

That might be a total load of nonsense, but that is what i think happens, based on what ive seen

[explorador]

Leosghost: nope, removed GA on the last launch of responsive designs and change to my latest CMS, but yes I have data from before it. I will.. add GA again, sorry it's something I don't like, giving away so much data to G (I know they still get it from any kind of script we insert from them in any or other way).

londrum: Been thinking about what you say too, I see no other way than waiting to see what happens. It's not about waiting and waiting, it's about finding the scope of effects after change, how long it takes Google to see things after their reaction (sandbox?)

[tangor]

Regardless how this is happening, there sure doesn't seem to be much incentive to keep hammering away! Do that and you might end up in truly negative numbers! (sigh)

[explorador]

Yes tangor, I agree and have better things to do, the good thing (the good outcome of this) is all the changes have been good for traffic, design and speed overall (regardless of Adsense). Adsense is the only thing not responding positive to the good changes.

I believe it could be a standard cautionary measure caused by an algo (stupid algos), G always avoiding human interaction. Now I wonder about the other scenario: doing everything at once instead of chapters but I will never now, the new modifications have proven to be good and I will not change anything my visitors are seeing just to find out about how G reacts to the other way around.

[Leosghost]

Nope ..I wasn't thinking of adding it ( for the same reasons as you state ) ..just curious if you were using it at the time..

I do know that G have access to data from other networks and ad suppliers, I mentioned it in a post here recently..I was on a site that uses ads from a Canadian ad network rvmedia..
Youtube started showing me videos suggestions in the "sidebar" which were obviously based upon the content I had been reading on the other site..Officially Google and rvmedia have no connections, however I do know that some ad networks sell user data between to other networks..add this to what londrum describes ( which I have also seen, and which I seem to remember has been mentioned here before, ad "targeting" needing "resetting" after "adjustments" ), could very well combine to be possibly why you appear to be hitting "ceilings"..

I would n't put Ganalytics back on ( although one can wonder if G do not occasionally comparatively "demote" sites that are not using it, but I think that would more possibly apply to ecom sites, although I would never put Ganalytics on an ecom site )..But I would wait longer between step changes, at least 15 days after any individual change on any individual site..Otherwise you would be dealing with too many variables in too short a time span..

Do your logs give you any indications of traffic patterns ( particularly GEO traffic pattern changes or time of day traffic pattern changes ) after changes made to date ? <= Might indicate "throttling" of likely to convert /click traffic..

Edit:I was typing when you made your post ( just above this one of mine ) and didn't see it...

[explorador]

Thanks Leosghost, exactly what I think and been seeing. I don't want to add GA, I rely on internal stats from the server (actually the real thing). Unfortunately thinking on direct ad sales in my region involves clients "needing GA reports" so I guess I have little space for denying to it. I also agree on leaving more space between changes, I'll have to wait and see.

Now that you mention GEO pattern and changes I did post something about it on another thread. My traffic has been solid on patters, but I've been seeing changes showing on G Adsense reports in the sense of: country #1 Austria, then some other day "México" some other day "EEUU" some other day "CANADA" and so on, unrelated countries, unrelated languages day after day (not just random) almost sounds as a rolling wheel.

I also believe on those "ceilings" while many deny it. My extra THING to keep believing on those things is... the ones denying it never post something worth believing, no specifics, no trends, just "no no no na na na", while other people living it do post data, explanations, they paint the picture. Will keep trying.

[Rasputin]

I have often run 'experiments' within adsense until they reach 90% confidence that results will improve, made the recommended change, seen income rise for a few days then seem to fall back to where it was before. Same with various other changes.
So this gives the impression a limit is imposed - but I think it might just be an 'impression' ie I have never done a statistically valid test of the type 'if I change X I think Y will result' on this subsequent reversal of income increases, and general impressions are very often wrong unless proven statistically. So for me the jury is still out...

[magician]

If everything else is working for you - traffic, repeat visitors, visitors' time on site, speed, then, adsense will catch up. You should see if the earning drop is due to lower CTR or lower CTC or combination of both.

If CTR is the major issue, experiment a little with the ad position, color, etc. Also, note that CTR is lower on mobile devices and may be causing overall lower CTR. May be, you should compare only desktop CTR prior and after the changes.

If CPC is the problem, then you may not do much about it. CPC on mobile devices is always lower than desktops. I suspect your traffic growth would mostly be from mobile and hence lower CPC.

As far as your changes are concerned - you are doing things in the right direction if your traffic and visitor experience is positively impacted.

[eek2121]

150,000 people from India will earn you pennies, 150,000 people from the US, UK, and other Euro countries will make you a (small) fortune. It's all about your traffic. Also, 500 sites that have 10 visitors each per day (from earlier stated audience) will earn you pennies, but 1 site with 5000 visitors per day (from earlier stated audience) will earn you a fortune.

I'll tell you like I've told clients in the past. Don't 'assume' anything. There are 3 things that will guarantee that you'll do 'okay' (subjective term...just look at my own rants) with adsense:

1) You have traffic coming from high paying audiences
2) You have quality, original content, not scraped, not written by 'joe' from india who can't speak 2 words of english, even if he did live in america his whole life, not copy/pasted, but high quality, original content BY YOU that is not available elsewhere. (Quality content meaning lots of quality text supplemented by a video, maybe an image or two)
3) You don't have large amounts of non organic traffic.

Live by these 3 rules, you'll make a fortune. Yep, there are exceptions. Yes someone will luck out and make a bazillion dollars by some stupid 'oversight' or 'mistake' or 'loophole'. This isn't about that someone, rather it's about you. You need to live and die by those rules. Think they don't apply to you? Think again. I'll bet you $100 that one of the 3 points i just made is being violated by your website.

[tangor]

@eek2121

As a generalized statement will freely grant your specifications, but there is one fatal flaw in your presentation:

An assumption that any given site is a user value site, one that has something the users seek and will stay long enough to even see an ad.

There are way too many sites on the web that get tons of traffic (for whatever reason) that have little to no value, resulting in an immediate bounce, by some odd fortune of keyword or perhaps niche.

If the site is krap, no user will stay long enough to even see an ad, much less click on one, and that includes all original, non-scraped sites. I'll grant the "high quality" aspect, but that, after all, is subjective.

The ugly truth is that a rather large percentage of websites have no value and (ahem) should dry up and go as dust on the wind (ie. DIE).

But hope springs ever eternal for webmasters clueless as long as the myth of getrichquik via adsense remains.

While some adsense income is iffy these days I can't respond to your $100 bet in kind. I will, however, up you with 1M productions of the byproduct of internally modified frijoles. Both have about the same value as regards commentary. Only difference is that you are not likely to part with $100 while I will part with the byproducts of internally modified frijoles on a daily basis! :)'

I tend to believe the OP there's a hit point in his mods. Why? G is always rearranging their payouts. Have been for years. And as long as they can stretch he game (that adblockers are about to bring to its knees) they will do so.

This will take a very long time, or at least a few more months at several billion transactions per day. When a market saturates, the "slice" of the "pie" gets thinner all the time and only the "house" wins.

Heresy of course for this forum, but that is also a call to NOT change what is CURRENTLY working for RWD, mobile, change G decrees because that will bring on a new "audit/accounting" and guess what? The numbers skew down, not up!

[netmeg]

My experience has not been the same. Most changes have been positive. Where it hasn't, I strongly suspect it's because of mobile.

[explorador]

Thanks rasputin, I'm still waiting before jumping to the next stage of changes.

eek, I do agree with tangor: fatal flaws on your analysis, so badly it makes me loose interest to reply. It's not a matter of liking or not the comment, a simple question of "are you on, on those 3 things?" is better rather than being so sure on "violating" them. Feels I would instead be giving you useful info in exchange of your bet on things being wrong, and that's free training and corrections I will not give.

tangor... thanks I agree, and you hit the nail: not changing what is currently working. I have been posting for quite a while about a ugly duck site, that best was 500 visitors per day in average for years and made me very good money until G decided to hate it. Pure original content, pure original thousand quality pictures, etc and every redesign affected the earnings so it stayed ugly for years. In this case my visitors, the traffic, layout and technology demanded a change, I'm seeing very good impact on every area except Adsense. Yes I agree with you but unfortunately I had to make a choice because Adsense is not the primary goal or the boss, and keeping a site as it is because it works ok with Adsense meant loosing other business. I have to keep up with changes, good, original content, user first, it's what has been keeping things evolving for me, adsense became a nonsense so it doesn't surprise me so far. I hope it will catch up, but we all have to decide at times changes with no going back despite of adsense says. I think you can relate to this, I'm explaining it because it expands things to readers.

lucky you netmeg, most changes have been positive. I will continue trying to find out what's happening and if things get better, if not it doesn't matter that much, the direct ad market is looking better now, people have been asking for my spaces, the line is where they pay more than adsense, so... there is space and options.

[netmeg]

AdSense isn't a good match for every site or every demographic or every niche and it can even change over time. People are pretty much never coming to our sites intending to click on ads; they're more likely if we happen to catch them in a buy cycle. And if they come in on a phone, maybe not even then.

[explorador]

True netmeg, totally agree. The thing here is not about a complain or "oh boy Adsense is not paying the rent", it's about an intermediary playing dirty because there are no humans... but algos in the middle feeding on good traffic, and the evidence of things not being functional, that's why I mention evidence because for every member saying he/she is making great amounts of money while others fail... I don't see a single quality post of anything, not that they should, but they really look like a diff kind of trolls.

[EditorialGuy]

For me, AdSense is pretty much "set and forget" (and has been since 2003). I give AdSense one 300 x 250 space on each page, and it's Google's job to supply ads that bring in revenue. So far, that approach has worked pretty well.

[trebuchet]

I'm the same as @EditorialGuy. Adsense works for me but as to why it works, I'm only half informed. Theoretically it shouldn't work on my sites, most of which contain 'dry' information, some visual, most of it text. My users aren't buyers, they're readers, researchers, students or people looking to fix something.

I will say though that I too have experienced the slump that follows major site changes. Going responsive, design updates, navigation overhauls, etc. have always caused short term earnings drops (usually 1-2 weeks). As to why this is I can only speculate. Blocking networks and rubbish ads has never achieved much for me so I don't waste time trying it now.

[glitterball]

I too have seen drops after every site-wide improvement.
I'd also add that I have had cause to change countries twice, and therefore had to change accounts, I also changed accounts from an individual to a company once.
All account switches have produced big changes in earnings, sometimes for the better and other times for the worse. The jumps or drops in earnings were immediate, so definitely caused by the account switch.

[tangor]

Taking nothing away from what I said earlier in this thread, I have come to believe that G has grown so large and compartmentalized that there are collisions regarding site changes at the website level, and filtering those through all the various indexes (there is not just one!).

However, in the adsense side, there is a general hit, hopefully short, but these days does not quite get back to where it was. THERE ARE EXCEPTIONS of course, but even those are not leaps and bounds better than previous.

The nature of the beast is being revealed in bits and pieces, ie. there's only so much it can do with all that computing power, to work those changes through.

Example:

I have three sites remaining with adsense. All in travel/accommodations.

One had a complete redesign to RWD, took a total 9% HIT and has not recovered after 9 months.

One changed from static to dynamic AND RWD and took a 33% hit initially and is now 40% down.

One I did nothing. It remains an earner with slight growth (about 0.9% monthly).

I have no answers. Can't explain it, can't seem to fix it (the second I poured another 40 hours of coding time and no love). But the bump down appears real. Seen it in a few client sites as well. We'd like to say "if it ain't broke don't fix it" but we can't do that any longer with the advent of adblockers, et al on the horizon.

THIS IS NOT DOOM AND GLOOM, kiddies, just reporting an experience in this day and time.

Note: the RWD changes were to gain mobile traffic. That did not pan out well.

[iamlost]

I've been fascinated by the long running examples of apparent Google mischief be it earnings cap, zombie traffic, et al. Yes, I have seen G referred traffic flow switch suddenly from one geo-location to another with an apparent artificial abruptness. I also remember wondering what all the fuss was about the 'sandbox' as during that period I was publishing a new site every six weeks without a problem ranking well within the week and continuing to do so.

While it would convenient to say that Google is setting various 'top' thresholds it is more likely, as tangor says: ...G has grown so large and compartmentalized that there are collisions regarding site changes at the website level, and filtering those through all the various indexes.... That would certainly be a more sensible, to my mind, barring further information, answer to blips in the continuum.

And that brings us to the long known common effect that major site/page changes seem to flip a Google reset switch. I know that as I switched (all sites evergreen info type) to responsive (with server assist) each site dropped in G traffic by 12% - 37%, which lasted 1 - 4 weeks before slowly climbing back up over the next year. Currently site G traffic is 90% - 246% of what it was prior to the change. That variance in both drop and recovery tells me that both are extremely variable effects. Just what the causes of the variance I haven't been able to come to a wholly satisfactory answer.

One point that I haven't seen mentioned (if I missed it please direct me) is the difference in traffic device concurrent and subsequent to a site change. I know that as I switched to responsive in 2012 - 2013 my G referred mobile traffic was nudging double digits and in the year following doubled. That traffic percentage is now 15% - 45% depending on the site. Which brings us to AdSense.

AdSense on mobile generally pays less per click (or per M). First think of the much smaller screen view available, then the amount of scroll between ad blocks (do you track how far your visitors scroll? do you know how many ad blocks they see?). About a year ago there was a huge fuss that spilled over into this year as advertisers realised they were paying on impressions (misleading term meaning the ad was served, not that it was seen). Ad networks including AdSense switched to the Interactive Advertising Bureau (IAB) defined viewable ad, which is minimum 50% viewable for minimum 1-second. One reason that AdSense earnings jump about (usually down) is that the initial is that the ad was served then it became IAB viewable and then (if applicable) it got clicked and finally it was determined not to be fraudulent. Quite the downhill slalom.

One reason that I am in the set the code and largely ignore AdSense. If the weekly projected changes by greater than 10% I do take a look in case something is mucked. Otherwise G can fufill it's agreement by filling the ad block(s). Granted AdSense is become the default fill revenue source, not the primary.

Check your AdSense performance reports for the 'Active Views' tab. You should see two interesting fields: Active View Measurable and Active View Viewable. Yup the names are stooopid. AVM is impressions, AVV is IAB viewable. Note: all ad views are impressions, not all impressions are ad views.
Note: if Active View Measurable is less than 100% it means that the ad is not being served properly (if consistently or if significantly <100% you need to find out why asap, your visitors can not click what isn't there).
Important afterthought Note: I believe that currently AdSense for content, for games, and for video are the only ad units stats shown; that DFP Small Business dynamic allocation and link ad units are NOT included.

If an ad block is being served, i.e. AVM is 100%, but AVV is low aka 0% or whatever less than your comfort threshold (mine is 80%) you may want to investigate how best (without getting banned from AdSense) to put visitors in front of it.

Besides all the above, we have difference in use: granted that mobile users visit your site/page are the ads appropriate and of interest? Unless 'wasting' time on games and SM mobile users are typically looking for specific information. Query, answer, go. If the ad is answering better than the page the page may ready for a fall, if the page the ad may be ignored.

Mobile is a pita. And a goldmine that typically requires more thoughtful effort than desktop. Simply 'going responsive' may not be sufficient.

[ken_b]

Thanks for that post iamlost.

I haven't quite understood the Active View stats and your post cleared that up for me.

[Edge]

The three most important things for any website... Traffic, traffic, and more traffic - however honestly you do that. One can play with technology, ad layout, ad size, banner width, mobile, https, or Google hyped up preference # 465 and you will never increase the sticky-ness and revenue awarded by increased unique or return traffic. Tweek to your hearts content - I'm working on traffic and all the good stuff will follow.

[explorador]

@trebuchet, what you describe sounds a lot like what I've been seeing.
@glitterball, exactly too.

@tangor, that, exactly. Your description is actually, by far very accurate and precise to what I'm seeing. Been researching and it seems most times whatever people do on redesigns it triggers a quarintine, G algo says "what?" and sites get hit. Yes, the redesigned sites got hit and didn't fully recover, what surprises me is: being careful on my decisions and doing the right thing, it brought an income increase right after the change, then drop.

It's like the visitors see the change, the algo reacts later, the spider? the bot? the temporary increase seems effective on the people who visits the site prior the next bot/spider check, sort of saying.

The secrecy of adsense/webmaster makes difficult to study this in full detail other than our own cases, but it seems is not 100% of cases, but there are very few exceptions.

[explorador]

Thanks Iamlost for the detailed post. Yes, been seing what you describe on sudden changes on geo location, my traffic has been for years very steady on that, and suddenly after such changes the primary source of visitors is #1-unrelated country, #2-another, etc, and each change triggered diff countries, then back to normal.

I also think on collisions, also on some kind of geo different bots and updates, that would explain the registry, perhaps due to constant spidering. Oh yes I agree on the reset switch too. Your description is very accurate and sounds as what i'm seeing.

I'm seeing now the AV and AVV tabs, I get your AVV is 80%, I have something like 75% and above, I will look into this, perhaps sometimes the "thing" is not visible due to responsive changes: pushing some ad to the bottom and rarely reaching the viewing area.

will keep researching and trying.

[explorador]

yes edge, I'm also focusing on traffic and content, adsense became a secondary concern, but sure catches my interest on trying to understand what's happening.

I survived the adsense anxiety (there are threads about we all checking the stats daily constantly etc) now... I don't care, but just as we talk about diversifying income sources (not seeing just one thing) I also take care of other stuff the same way: keeping an eye on several things just in case, adsense included.

[SnowLeppard]

I am currently reading this new book:

Content Inc.: How Entrepreneurs Use Content to Build Massive Audiences and Create Radically Successful Businesses
[amazon.com ]

You would probably find it useful too.