|WMT: significant increase in impressions while decrease in clicks|
| 5:33 pm on Jan 29, 2013 (gmt 0)|
So, for the last two days I'm observing a rather unsettling trend for two of the sites I work on (not related, not linked and in two different WMT accounts). The amount of daily impressions on the graph in WMT->Traffic->Search Queries jumped to about twice the usual amount during the preceding month and yet the amount of clicks on the same graph is either unchanged or even slightly down. Even more worrying to me is that looking at the server access logs, the actual Google visits are definitely down (10% on average).
So, I'm bracing for a possibility of a steep drop after this impression frenzy is over because how long would they keep showing my pages in results if no one clicks on them?
Does anyone have experience with these impression spikes? Anything can be done about the sudden decrease in click-through rates? I'm still trying to find a sample of the title/description they would show for the listings not generating clicks but so far all I see are "normal" listings - nothing out of the ordinary, they look just like they've always been. Do you guys know to what degree they personalize the SERP listings in terms of changing the title and description and could this possibly lead to getting consistently lower CTR? If that has to do with personalizing the listings, do they have a way to automatically correct this issue, I wonder?
| 7:04 pm on Jan 29, 2013 (gmt 0)|
For the last days i'm seeing the same thing. I see a 300% increase in impressions, but no significant increase or decrease in clicks ...so traffic remains the same.
Deeper inspection shows that the site (which I believe was hit by both Panda & Penguin), is now ranking at the bottom of page 1 and on Page 2 for some low-volume keywords. That explains the increase of impressions while traffic remains about the same... at least for me.
| 9:37 pm on Jan 29, 2013 (gmt 0)|
I don't know if that explains the spread all that well - the bottom of the first page has always been a higher-CTR location than, say, #5-#9 on the first page. #1 on the second page would have also been a relatively high-CTR location for as long as they actually open the Page 2 (and, presumably, if they didn't open it, it would not count as an impression).
|Deeper inspection shows that the site (which I believe was hit by both Panda & Penguin), is now ranking at the bottom of page 1 and on Page 2 for some low-volume keywords. That explains the increase of impressions while traffic remains about the same... at least for me. |
Actually, come to think of it, this Page 2 thing may not be as straightforward - when someone searches for "blue widget" and only looks at Page 1 - do all those people that have sites in the result set, all the way down to page 100 (or whatever the cutoff is) - do they also see a +1 on the impressions at their relative positions?
It would have been ridiculous to do that for practical purposes on your own website but who knows how creative Google gets with their stats. For example, I have 200+ impressions for some KWs that show their average position as 880+(!) I seriously doubt that there may possibly be 200 people that got all the way up to Page 89 (or even 9 @ 100 per page) looking for that thing - this must have been either an error or an exercise in "creative statistics" by Google.
I have a nagging, not yet supported by any factual observations feeling that this 200% - 300% increase in the "Impressions" only means that there has been a fresh influx or pages from our sites that had entered the first 1000 results set (the main set as opposed to supplemental) on some (hopefully significant) number of new KWs. I don't know what to make of it yet. It does not sound like it should be a bad thing, and yet the total traffic is dropping.
| 11:10 am on Jan 30, 2013 (gmt 0)|
Something is wrong with that graph. For the last few days, it is showing that both impressions and clicks are down 30% to 50% for my website. My analytics shows no corresponding drop in search engine referrals.
| 4:18 pm on Jan 30, 2013 (gmt 0)|
As the graphs get updated in WMT, I am now seeing more sites starting to show this spike of impressions not followed (or going in opposite direction) by clicks. I now strongly believe that they've just messed with the math behind the impression calculation or simply changed the definition of what an impression (at least the one shown on the graph) is.
It's Google's version of cooking books - practically harmless and costs nothing but may make you feel a bit better about them because they show your links more often yet slightly embarrassed to complain about anything else (Google Images, for example) because your CTR is so dismal. I am sorry if this comes through as paranoid, I find myself to be unable to think positive about Google intention anymore these days.
| 4:33 pm on Jan 30, 2013 (gmt 0)|
The Google has been very active this month. Nearly daily SERP adjustments since the last Panda update.
We saw the same impression spikes across all our web properties. Course what goes up, must come down...
| 5:04 pm on Jan 30, 2013 (gmt 0)|
I got the same thing and my traffic has not changed a bit in the past week! Even with the latest panda...
looks like this and tells me nothing:
| 12:20 am on Jan 31, 2013 (gmt 0)|
Could it be that they're counting impressions differently... like counting every single word in a long tail query as an impression for that word? I'm only half-joking.
| 3:30 am on Jan 31, 2013 (gmt 0)|
I see the same with the now default "web" filter on. If I switch to "all" to include image searches the numbers are flat, no change.
I hope that a spike in the number of impressions for "web" searches means they are easing up on showing 1 site 87 times in the top 100.
| 4:34 am on Jan 31, 2013 (gmt 0)|
@Sgt_Kickaxe: Sarge, we cross paths on the matter of images once again, in another thread :) I am looking at the stats of a site that has no images whatsoever, none! Other than template of course. It's about very dry, specialized and unappealing accounting matters. If images would have existed, they would be images of text documents, so they don't - no need. Still showing the same impressions bell curve (left part of it so far, sigh...) with Web or with All filters enabled, so images did not play a role.
Nevertheless, thank you for the idea to look at the filters! I have noticed that in the "Location" filters a sector called "Russia" appeared. Since the site has nothing at all to do with Russia and never had any Russian traffic before, I though it's peculiar and looked at only filtered Russian traffic - impressions start from zero on the 16th, hover just above zero until the 25th and zoom all the way up to the unaccounted 50% of the total traffic on the 27th - an astounding 100,000% increase in Russian impressions in two days. Yandex must have crushed over the weekend and I didn't hear about it (no, seriously, did it?) . I have no content in Russian and the subject matter is very narrowly US-oriented, so, not surprisingly, all these thousands of impressions generated less than 10 clicks (in fact, I think it's just zero).
Anyway, this gets me closer to some kind of an explanation but does only explain one site. I have other sites showing the same impressions explosion and yet they are not as narrowly geographically focused and have had some valid traffic from Russia before. You would think that 100,000% increase in impressions would give me at least some sort of an increase in clicks, and yet it does not.
I am thinking this is probably a reporting error (or feature - however you look at it) which made an entire country (and perhaps several), impressions from which were hidden before and suddenly they started to get counted in total impressions. You would think with big changes like this an announcement from Google would follow...
I am off to checking for signs of other geographically misplaced impressions that may be skewing the stats on my other sites.
| 4:53 am on Jan 31, 2013 (gmt 0)|
Yup, that's Russian impressions all right - they can explain pretty much all of the spikes I'm seeing so far. What I don't understand though is that it's not just any additional Russian impressions - it's a hell of a lot of BAD impressions at google.ru, the ones that don't convert to clicks. I was paying particular attention to a couple of my sites that have always had traffic from google.ru, so I know that there should be *some* conversions, and yet I'm not seeing any additional traffic from there, despite the incredible increases in the number of reported impressions.
I thought that they may be experimenting with some badly converting new layout or something and took a stroll to google.ru - it does not look or operate any different from .com So, the sudden influx of impressions still looks like a reporting error to me.