|Google Analytics Queries Report Inaccurate?|
| 8:01 pm on Apr 26, 2013 (gmt 0)|
For the past few months, I've been pulling the GA Queries report for every Monday, to get a birds-eye view of several different metrics, including average position. Now, I know average position doesn't mean the highest position a page is found out for a keyword, but an average of all the impressions for a keyword in a given time period. I get that, but it's currently one of the best ways to do what I want, and the trends are still pretty telling.
All seemed to be well (other than the discrepancy in numbers between it and Webmaster Tools' version) until the last month or so. Suddenly the GA report is showing a lot more keywords from page 5 and higher that have crazy amounts of impressions. However, they're short-tail, high-search keywords, so I figure they might get more impressions than your average keyword. I did note that these keywords don't seem to show up in the WMT report, and thought that was a little weird.
Then I saw something today that took the cake. Our brand keyword was being reported at as being on the third page. Panicking a little I looked it up, only to see that it was at position 1 today as it should be. The GA report also showed it on page 6 on Tuesday. So I went to WMT, and wouldn't you know, position 1 for every day this week the data was available. And that other keyword? Still not showing up.
Am I wrong to be using the GA data? Should I switch to WMT instead? I like GA because it has all the click data, but if GA is going to give me bad data, I'll use it instead and just make up a number for the >10 clicks.
| 10:20 pm on Apr 26, 2013 (gmt 0)|
I thought GA imported that data from Webmaster Tools? If that's the data you are talking about, I would put more faith in the WMT version... even though that data is notorious for being problematic, too.
| 4:48 pm on Apr 29, 2013 (gmt 0)|
The impression numbers, at least, differ by a very wide range, though the trend is more or less the same.
I guess I'll use WMT instead.
| 12:15 am on May 2, 2013 (gmt 0)|
|The impression numbers, at least, differ by a very wide range, though the trend is more or less the same. |
Are you comparing the right values? The default filter in WMT is "web" whereas the default filter in GA is "all" which includes image impressions, mobile impressions etc.
| 1:51 am on May 2, 2013 (gmt 0)|
I combined GA and Gwmt a few days ago and noticed along with other inconsistencies that GA also reports different keywords/phrases as getting more impressions than Gwmt.
| 2:15 am on May 2, 2013 (gmt 0)|
IME (In My Experience) WMT is buggy FUD.
I don't give half a second of though to the numbers it reports and haven't looked at it for the sites I'm involved with in nearly a year as far as I can remember, because acting on FUD and Erroneous Reporting is a dangerous thing to do, imo.
Get real stats. Make visitors happy. Forget the FUD and BS in WMT. You'll rank better and you'll have a site people like rather than a site a search engine gives you FUD and buggy (at best) data about.
Case-in-point (Apr. 26th, 2013):
Ignore That Last Google Webmaster Tools Notification: Googlebot CAN Access Your Site [seroundtable.com]
(To at least one of us who lives in the real world and doesn't like candy-coated BS, the preceding means: Oops! We didn't mean to send that e-mail out, but what do you expect for free, accuracy and reliability? Ha! You get what you pay for, even from Google, and you don't pay anything for WMT.)
[edited by: Robert_Charlton at 7:16 am (utc) on May 2, 2013]
[edit reason] fixed formatting [/edit]
| 2:50 pm on May 3, 2013 (gmt 0)|
aakk9999, thanks for that. When I change the filter, the numbers for individual keywords now match up. Total impressions are still off (higher in WMT than GA), but I'm not really concerned with them, anyway. Also, those keywords I couldn't find in WMT now show up.
Actually, it brings up something I need to look into more. One of the weirder keywords is showing up as image, even though we don't actually get much image traffic usually, as our images are hosted on a CDN.
| 4:00 pm on May 3, 2013 (gmt 0)|
Another problem I noticed with Ga/Gwmt is you can no longer click on the keyword in GA to see which pages were affected like you could in Gwmt and the two tools report different ranking on the same day (Gwmt is still stuck on Apr 25th ranking). Hmmmm, that's about when I combined them. I'm wondering if Gwmt is no longer recording anything since I combined the two!
| 4:42 pm on May 3, 2013 (gmt 0)|
I've found the data to be inaccurate too, but one other thing to check is the 'country'. You might rank better in a regional Google than you do in international Google for example, and you can filter the reports to show you that.