|GWT: Top Pages vs Top Search|
A long-overdue question:
In gwt we've got "Top Pages" and "Top Searches", each with numbers for Impressions and Clicks. Question: Why are the totals so different?
Say you've got a page all-about-widgets.html. It's the only page on your site that talks about widgets, and there's no plausible way a searcher would arrive there if they weren't looking for widgets. So you'd think that if you take all search strings containing "widget", they would add up to the number for "all-about-widgets.html".
In fact they're not even close. The number on the "pages" side is at least twice, up to 10 times as big as the total count of "searches". I went back and checked my figures from spring 2011 when I started keeping monthly records. No change. That means: before they added the "web" default, before the widespread transition from http to https.
The only way I can get it to work is if I postulate hundreds and hundred of long-tailed searches, each using a different phrase, each coming in with only one impression and maybe 0 clicks. But my site is small enough (in pagecount) that I'm pretty sure the table lists everything; it's not like the Keywords list that cuts off at an even 200 items.
What am I overlooking?
I think this blogpost by Jonathan Simon, Webmaster Trends Analyst (Google), can help you to understand this.
Is the explanation hidden in the comments somewhere? The post only explains the parts you can figure out for yourself :(
And why are these explanations always in blogspot dot in?
I (re)read that post, and I'm none the wiser either!
:: bump ::
:: pause to swear loudly because browser unexpectedly quit just as I finished composing post ::
I went by gwt to look up something else, and while there I took a closer look at the Top Pages list, particularly the dropdowns that show the individual queries that add up to the total. At least, uhm, I thought that's what that little arrow is for :(
Page #1, add up sub-queries: total cannot possibly be more than 1/6 of the number given for the page as a whole.
Page #2, same arithmetic: mmnn, this one could go either way.
Page #3 is a complete wtf: two of the listed searches-- but not the first two --are each, separately, much bigger than the number given as the page total.
Page #4, like #1, adds up to much less than the given total.
Page #5's sub-list goes on for about a mile and a half, but the sum of queries is definitely bigger than the total.
Page #6: the first three queries already add up to more than the total.
Now, if I try clicking on any random one of those sub-queries I always get a "No information available" graph. (Incidentally they also claim not to know anything about sites that link to mine-- which, ahem, do exist-- but let's assume this is an unrelated computer glitch.)
At this point it's awfully hard not to conclude that the computer in charge of Webmaster Tools is simply making up numbers out of the clear blue sky and hoping nobody will notice.