| 10:56 pm on Nov 3, 2012 (gmt 0)|
Did you filter by location?
[edited by: Robert_Charlton at 11:06 pm (utc) on Nov 3, 2012]
[edit reason] added clickable link [/edit]
| 11:41 pm on Nov 3, 2012 (gmt 0)|
|Did you filter by location? |
Seriously? Of course I did ...
| 12:03 am on Nov 4, 2012 (gmt 0)|
|I thought it was a bug at first. |
Maybe it is. Can you think of a better explanation?
| 12:27 am on Nov 4, 2012 (gmt 0)|
|Seriously? Of course I did ... |
Seriously? I'm not a mind reader.
| 4:03 am on Nov 4, 2012 (gmt 0)|
I think it is a bug, I see the same stupid reports that tell me that every single page of my site is in position 1.0 and I wonder 1.0 of what? It certainly is not in position 1.0 of anything I can find. And I see the same useless information for several sites. I still wonder why they have us verify sites and select a targeted location only so they can feature your site on the other side of the world. When I apply filters for location: US to some traffic reports it drops from thousands of impressions to less than 10 with no indication of why. There is a lot of useless stuff in GWT. I still look there because occasionally I will find a useful gem of information. Almost never though.
| 4:11 am on Nov 4, 2012 (gmt 0)|
it is possible that is the average rank when your site does show up in search results.
perhaps it doesn't rank well for your personalized (or perhaps "depersonalized"?) results but ranks well for others' results.
| 4:41 am on Nov 4, 2012 (gmt 0)|
|it is possible that is the average rank when your site does show up in search results |
? It would never have occurred to me that it could mean anything else.
Now, if someone could figure out what they mean by the occasional "average position 1.0" which is manifest nonsense, since it implies that your page is either #1 or it doesn't show up at all. And --necessary disclaimer since this is my site I'm looking at-- these aren't searches so arcane, they only bring up one result ;)
| 5:46 am on Nov 4, 2012 (gmt 0)|
Filter by content type. Quite often you will have a highly ranked image that is top ranked despite the page not being in the top 100. GWT doesn't distinguish between the two unless you tell it to.
| 9:21 am on Nov 4, 2012 (gmt 0)|
|it implies that your page is either #1 or it doesn't show up at all. |
this could theoretically happen with personalized search.
that's why i pointed out that the times is doesn't rank don't get averaged in with the searches where it rank well.
|GWT doesn't distinguish between the two unless you tell it to. |
local/vertical search - another case in point.
| 10:01 am on Nov 4, 2012 (gmt 0)|
|I think it is a bug, I see the same stupid reports that tell me that every single page of my site is in position 1.0 and I wonder 1.0 of what? It certainly is not in position 1.0 of anything I can find. |
I thought this was a brand new bug but I just went back and looked at my saved file of September's numbers. I never noticed before that all those obscure pages at the bottom-- the ones that might show up in two searches during the whole month-- all claim to be #1. Total, about 2/3 of the site. Oh, come on. If those numbers are from personalized searches then I have to say g### isn't doing a very good job of it because I don't see a fat lot of "&cd=1" in my logs.
If I look at the top of the list-- sorting by impressions-- a couple of them are bona fide 1.0. But even those are fishy. In all possible searches that ever bring up such-and-such page, it's always at #1? Nuh-uh. Not even wikipedia can say that. Besides, I see those pages in logs. And they're not always #1.
:: noting with interest that through the miracles of Daylight Saving time, a large chunk of recent posts were made up to an hour in the future as I type this ::
| 10:48 am on Nov 4, 2012 (gmt 0)|
most of those personalized searches are now showing up as "(not provided)" =8)
| 11:35 am on Nov 4, 2012 (gmt 0)|
Yup, that's my #1 search term in piwik too. But someone noticed a while back that even when g### doesn't disclose the actual search terms, they will usually tell you where you ranked. It's in the logs. So you know that someone searched for, uhm, something -- and whatever it was, you were the first result. Or the 10th or the 75th.
| 10:13 pm on Nov 4, 2012 (gmt 0)|
Most of my spurious #1s come at the bottom of the list-- that's why I never noticed them until I actively looked. That means the ones where both "impressions" and "clicks" are given as simply "<10". But if I scroll down to the bottom of the bottom, I again meet a fistful of bigger numbers. (Bigger in this context means worse.)
g### says that <10 means some number greater than 1. I'm inclined to doubt the >1 part. In the case of impressions it obviously means >0 -- otherwise your list would go on forever :) -- but if your site is as small as mine, any given month will have pages with only one impression and zero clicks.
Here's what I think: When "<10" means 0 clicks, then the page position isn't calculated. It defaults to 1.0. It is up to you to figure out whether "1.0" really means #1 or whether it means, er, something lousy ;)
| 1:16 am on Nov 5, 2012 (gmt 0)|
> GWT doesn't distinguish between the two unless you tell it to.<
Wow! How do you tell it to? I've been complaining about this for ages because we have tons of photos that get high ranked.
Oops, never mind. Finally found it after posting. So THAT's what 'web' and 'image' filters mean.
I just realized though that that only applies to whether they SEARCHED for web vs images. For instance if an image comes up in that row of images under a text search which occasionally come up on the 3rd row of text results, and they click on it, WMT still shows the .jpg using the web filter. It also appears to report the avg rank from an image search.
| 3:29 am on Nov 5, 2012 (gmt 0)|
You also need to understand how Google "counts". For example, if the site is #1 and shows with large sitelinks, and someone clicks on one of sitelinks, then the position will be recorded as the position of the sitelink, not as #1. So bottom-right sitelink will be recorded as #7.
Secondly, if you are #2, that is, there is a competitor as #1, followed by a row of 4 images, then your site shows next, it will be recorded as #6 as each image counts as a position.
Local search including google maps also show as position, so if you are listed first after the 7-pack, you will be 8th.
And as Lucy pointed out, definitely filter for "web" if you do not want to be impacted by image search - which in fact not just impacts your average position but also the CTR shown.
| 1:37 pm on Nov 5, 2012 (gmt 0)|
I have had the same thing on a site of ours for a long time, maybe 2 years.
They say our site ranks at #5 for its main keyword when in fact it is nowhere to be found.
The site **was** on the first page then got hit with one of their so called UN-natural link warnings. We basically told them to piss of as Bing loves the site and we are building sites for **people** and not going to alter anything to try to please some search engine as per their guidelines.
What it looks like is their wonderful tools show the rankings without penalties, they are basically cloaking, showing the webmaster one thing and real visitors something else.
| 11:10 am on Nov 29, 2012 (gmt 0)|
"What it looks like is their wonderful tools show the rankings without penalties, they are basically cloaking, showing the webmaster one thing and real visitors something else."
That's an interesting concept. Any evidence to substantiate? Ours that report weird don't have any reported issues. I've been trying to figure this out for months now as they say some of our previously best pages are, say, still #8 (was once, but hasn't been in over 6 months) and I find it hovering around 425. While many others are right smack on. I've even tried proxy surfing to do the search, searching from a machine I've never been on, etc. But nowhere near 8 under any form of search.
It COULD be, as suggested, that WMT is accidentally reporting what it's rank WOULD be without some recently added Panda/Penguin penalty which they forgot to add into the WMT progam calculation. That would be a very interesting window into the algo...
| 12:48 pm on Nov 29, 2012 (gmt 0)|
- Image and web traffic are combined
- Average position is weighted by number of impressions
- Average reported rank sometimes includes multiple pages
Be sure to check image search to see if you image hasn't been replaced by the same image on a scrapper site.
| 10:04 pm on Nov 30, 2012 (gmt 0)|
"image hasn't been replaced by the same image on a scrapper site."
Yeah we had that happen a couple times.
And all the bullet items I've naturally checked and accounted for, but, OK, I think I have a new theory that seems to fit perfectly and that I've been testing the last few days.
Normally when testing to check your rank you go into G with your personal preferences turned off (to supposedly see what others see). Under this situation, you see that your page is say #500. Few people search that far to EVER see your page. Now say on the other hand you ACCIDENTALLY go to search for your page's term WITHOUT turning off the personal preferences once in a while, ta-da your page suddenly pops up at #8 for you. So you quickly turn off preferences, then proceed to see where you are without them, but that impression has already been recorded at #8 whether you click or not. OR, scenario 2, someone else, a non-SEO person, who has been to your page in the past by finding it when it was higher searches on the same term with THEIR personal preferences turned on and gets it at #8 or thereabouts as well, whether THEY click on it or not either.
So overall it indicates it in WMT at an average of 8, since no one EVER sees it WITHOUT their preferences turned on in order to lower it to around 500. Now if it's NON-preferences position is closer to the preference position, say 8 vs. 30, then WMT might show it's avg at more like 12 since some people might actually get to it at 30.
Simple when you think about it in that light.
| 10:35 pm on Dec 6, 2012 (gmt 0)|
Think I got it!
Mobile vs. Desktop.
It's entirely possible to show in the top 10 on mobile, but 100+ on laptop.