Forum Moderators: martinibuster
This stayed like this for a good 6 months, on a very even flow. Then, all of a sudden in mid-August, the CTR dropped from one day to the next by 40%. Right during that week of major Adwords changes too. (maybe it is just coincidental, but in this game...who knows...)
The lower CTR has stayed at that new low ever since, even though largely the same advertisers are on my site, showing the same pool of ads more or less, and traffic is still at the same quantity, and of the same targeted quality, coming from the same refferers as well. My site is PR5, and my Alexa numbers are good and stable too.
There could be some very obvious, rational, measurable (and common) causes for this. But unfortunately, none of them seem to fit to my case.
CTR in most cases is really likely to decline for reasons which (IMO) fall into three categories:
1) The ads themsleves, such as: change in ad inventory, ad quality/targeting, etc.
2) The quantity/quality of traffic (amount of visitors vs. targeted traffic, traffic quality, etc. as well as visitor's behavioural habits)
3) Site changes, ("spider food", keywords, content, colors, etc. changes in ad positioning and similar stuff..)
Another reason for declining CTR (which could fit into #2, under "visitor behaviour trends") is this phenomenon of "ad blindness". But this doesn't fit my case, as this would most likely occur gradually, over the course of several weeks, or even months -- not from one day to the next. Several thousand unique visitors, coming from the same search engines, are not likely to get ad blindness all on the same day with the same ads showing(!?)
So, what I am wondering is (while my tinfoil hat goes on...) could Google simply be discounting some of the clicks right off of the top,and adjusting the CTR to accomodate this, merely because some algo has determined a percieved "CTR -to-conversion" for my site, sitewide? Could it be that maybe this ubiquitous "smart pricing" could affect not only CPC/eCPM, but CTR as well somehow?
....because something real funny happened here....
Tell me it was, I think, August 12th...$hit happened that day to me too resulting in a massive re-presentation of my sites over the ensuing days.
Why has it taken you so long to find us here and wonder what had occured?
It may only be a couple of months ago but so much has transpired since then that trying to get you up to speed means you will have to read a lot of the posts here to understand fully just what has been changing!
I started reading as much as I could about smart pricing and all the variables, (ctr, cpc,etc) trying to understand it. I have read hundreds of threads here, and couldnt really find anything that stuck out.
The reason I posted this today is largely because of the big threads on smart pricing got me thinking, and enough time had transpired that I could be certain that this wasn't just some 'blip' or transient weirdness.
What got me thinking that CTR could be capped somehow, is some recent threads I have seen, suggesting that earnings could be capped off for sites. I bet that all of this is a part of one big, strange puzzle.
for me, that week of horror started the 12 of August. Same time too, that I see that a thread started in the Adwords forum about AW changes.
But my main assertion in all that above is that I bet that CTR is factored in some sort of relationship it has to conversion rate. And then, thatconversion rate is the main variable in the nefarious "smart pricing".
in this thread (msg. #6 + #7) there is some of the same suspicion that CTR is used somehow to measure conversion:
[webmasterworld.com...]
And msg #7 suggests that one thing that drags CTR down is targeting. In my data above, the targeting on my site has always more or less been bang-on right.
There are full of discussion on the SP matters, but that is simply a sign to diverify your income sources....do it eariler and you get less pain.
Do the ads that typically run have a seasonal component to them? Is it possible the mix of traffic sources changed suddenly (i.e. search engine reindexing)? Is there anything worth mentioning about where the traffic comes from, geographically or otherwise? Is it a forum/community site with a lot of repeat visitors in addition to the SE traffic?
Oh, and how confident are you that something really changed suddenly? For example, if you take your CTR per day and paste it into an Excel spreadsheet and make a chart of it, is the pattern unmistakeable? It looks like you're being quite rigorous so no offense intended, but a visual display of statistical data can yield a subtler intepretation of statistical patterns that simply looking at raw numbers.
Best we can figure out is that the adverts are less well written.
The pattern is definitely unmistakable, I can confidently say that all the "usual suspects" involved in CTR declining, aren't measureable in my case.
What I am thinking is that although CPC, (and overall income) is definitely determined my SP, CTR is possibly something that has the biggest connection to SP, as CTR is likely measured against conversion rate.
My fear is this: what if Google somehow crudely and algorithmically figures that a certain percentage of the CTR is worthless (ie. doesn't lead to conversions) and caps it somehow, or applies some sort of artificial downward pressure on it.
There are some other threads lately surfacing too, which are no the run-of-the-mill "CTR droppped" laments.
What gets me thinking like this is just that when I strip away the obvious, I am left with nothing but the mysterious.
Exactly the same thing you speak of for your site, has happened to our main site, in exactly the same time frame.
Our first thought was... the changes to the adwords system.
Our second thought was (2 weeks later), it's just the results of Katrina.
But we are still down nearly 40%, and it's almost November.
The only farfetched conclusion we have been able to come up with, is that there are many more sites now in our niche running adsense, and the ads are spread to more publishers, (whether legit or not) and they drain the higher paying ads quickly.
Our main site is not seasonal, and we know it converts well. Yet, we have a summer site on the same account. The summer site's traffic is down of course, so... big question... Is the summer site dragging down the main site's EPC because of smartpricing, even though the main site converts?
Plus, it seems many advertisers are turning off their ads every single weekend.
Neither site has lost rankings.
In some publishing niches, the increasing number of sites may mean that the amount of available ad inventory is seriously outstripping the supply of ads. The result is clearly one in which the quality and quantity of available ads may be deteriorating for publishers. Any decline in quality and quantity will clearly have an impact on CTR.
Other publishing niches, of course, may not be effected at all or may be effected later.
I guess it will all be a matter of supply and demand.
Only Google, of course, can tell us what is happening in the area of ad supply and demand in various niches.
On my main site, I have also noticed a serious deterioration of CTR starting in late August. This in spite of the fact that I have not made any changes to my AdSense ad layout (I am not inclined to spend a lot of time tinkering with AdSense). I regularly check some of my main pages; and, in my opinion, I have noticed a deterioration in the quality and extent of ads on many of them.
This raises another question:
is there a way to best keep track of, or index the percentage of new competitors (as adsense publishers) in your niche?
Apparently the advertisers for our niche are not very smart, otherwise they would leave them running at night, since our visitors are blue collar workers, and only do their browsing at night and on weekends.
On the other hand, if they are smart, and just running out of budget, then we can lay fault at the "made for adsense" sites that have enter the niche in droves since Sept.
Max, the only way we know how to track that is by how many new ads show up in the search for our main keywords, from unheard of wierd domains.
