WMT Now Shows ALL Keywords - with CTR and Position

[mahidhar]

Not sure if this is the correct forum.

Just noticed new interface for top search queries in WMT. It displays all the key words (previously it used to display only top 100 queries), impressions and click throughs. Impressive, I like it.

Does any one else see this?

[alahamdan]

Looks like they added the ability to add more owners! any one see it?

Add or remove owners

near each website

[mark_roach]

This is really interesting data, but as with all WMT data the numbers provided shouldn't be taken as gospel.

The figures also illustrate how it can be misleading to use a high (or low) CTR value in isolation as an indicator of the quality of the search result.

The headline CTR value can appear to be significantly lower than it perhaps should if you have multiple listings for a given search.

For example on a misspelling of my domain name I have listings at both positions 1 & 2.

The Headline figure is 390 impressions 170 Clicks 44% CTR.

The detail figures are:

Position 1 - 210 imps 170 clicks 81% CTR
Position 2 - 210 imps 36 clicks 17% CTR

So ignoring the fact that the detail numbers don't tally with the headline figure you can estimate there were probably 210 searches and 206 clicks.

In this case the searcher is obviously looking for my site and would only ever click on one of the listings so the effective CTR should be taken at 98%.

Similarly in the case where you are searching for the answer to a question and you are presented with say 4 different forum threads from the same site. If every searcher tries all 4 of them before getting the answer they need (or not) that would give a CTR of 100%. If on a different search with another 4 results from the same site they all hit the right answer first time that would give a CTR of 25%. Which CTR is the better indicator of quality ?

Where it could be useful is in determining whether to show multiple listings from the same site and which ones to show. If in the above examples one of the 4 listings is getting a much lower CTR than the others it could then be replaced with an alternative from the site or removed all together.

[Angelis]

Position 2 sucks apparantly, must get P1 for all keywords!

[johnnie]

This should certainly help in tweaking for the optimal meta description!

[olias]

I know that a lot of webmasters have long speculated that position 11 is probably better than the bottom half of page one. I know it has always looked that way from my stats.

For one of my keyphrases the front page of my site usually sits at position 11 but will occassionally creep as high as 8. The figures on WMT do now show that I get a better CTR on page two than I do on page one. Nice to see that one confirmed from G themselves.

[idolw]

Looks great but:
1/ for one site - data shown in WMT are different from what I see in my Urchin stats. Lots of keywords are missing.
2/ for another site - data shown in WMT are way different from what is seen in Analytics.

Does anybody else see this or am I missing the point of this feature?

[alahamdan]

Looks great but:
1/ for one site - data shown in WMT are different from what I see in my Urchin stats. Lots of keywords are missing.
2/ for another site - data shown in WMT are way different from what is seen in Analytics.

Does anybody else see this or am I missing the point of this feature?

G WM center says:

The data displayed in Webmaster Tools may differ from the data displayed in other tools, such as Google Analytics. Possible reasons for this include:

* We may not have crawled your site since the changes were last made.
* Webmaster Tools does some additional data processing - for example, to eliminate duplicates and visits from robots - that may cause your stats to differ from stats listed in other sources.
* Some tools, such as Google Analytics, track traffic only from users who have enabled JavaScript in their browser.

[Reno]

I'm coming to the same conclusion -- be in the top 3 or die.

Question: Will this send even more webmasters to AdWords? If so, it explains why all this information may not be a "mistake". Crazy like a fox, as they say...

..............................

[Jessica]

where did the rank position info go?!

[Calculus]

So how is everyone using this expanded data?

Any ideas/strategies (the simpler the better)

[Gomvents]

data is way off for all 100+ of our clients including low, medium, and high traffic sites... hosted all over the world, owned by various companies and individuals so no common thread... buggy data... let's see in a few months... could be a step in the right direction.

[ffctas]

Hi,
Is there a way to compare adwords positions and click thru with the new webmaster tools report. Being able to compare this will help with efficient bidding on keywords.

[iNET_SEO]

It is nice to see a change that actually adds value. Very analytic-esque in its look and feel and in what it offers - about time :)

[Rugles]

This is fantastic! Thanks Google.

There goes my day ;-).

[Rick_M]

I suspected they were using ctr when Matt Cutts posted in his blog last month about improvements in someone's intentions when doing searches in various languages. The simplest way to determine a user's intention is to see what links they click on. The data Google can gain from ctr's is extremely powerful because it can uncover unanticipated user intentions. An additional suspicion I had was that serps would have to include some randomization (like adwords) to assess which links do best at different positions with the goal to fit the highest ctr at the top position. I believe some of the randomness we see in results now is part of this approach and over time as user intentions are better understood things will settle down. Of course occassional randomization has to occur to constantly assess if user intentions are chaning.

The problem with this is people often create misleading page titles which will lead to high click-thru's. I am hoping Google has sufficiently worked out ways to detect this (and fake clicks) before releasing this information / notifying webmasters that ctr is a part of a quality score.

I find the whole thing very fascinating if they are doing what I suspect. It is sort of a way of using user choices to better understand language, intentions, and human behavior.

[BillyS]

Do you guys also have your impressions capped off at 6,000/day?

1script - Our top term only runs to 1,600 impressions per day.

This is monthly or daily or what?!

alahamdan - Look at the chart, you can vary the timeline. It looks like the default is around one month.

I'd imagine the data is taken from all search engine variations (.com, .co.uk, .de, .ca) and across all data centers ... and then throw in advanced searches, logged in searches etc. etc. etc.

internetheaven - You can filter by date and location.

[Kufu]

I noticed this last night, and started drooling immediately!

[robho]

Do you guys also have your impressions capped off at 6,000/day?

No, I have terms on several sites with over 200,000 impressions a month, one with over 300,000 impressions (but only a 4% CTR otherwise I would have retired already!).

18,000+ impressions for the top term on a recent busy day.

The actual click totals recorded are very different from Anaytics, especially on sites with many search terms. Still, an interesting sample.

Does anybody know a way to download the expanded version of the table (with positions)? The link at the bottom left just downloads the summary for each term.

[ken_b]

OK, when a keyword gets a higher CTR in position two or three than it does in position one, I'd sure like to know what was above it.

This is great stuff!

[aakk9999]

Well, I have interesting data - narrowed down to one week and for one search term I have 220 impressions, 280 clicks and CTR of 127%

Now how is this possible - more clicks than impressions? Would this mean that the user clicked "back" to SERPS and then clicked on another of my listings in the SERPS?

The interesting test could be done with this data - if you have a few sites, add some phrase that is unique to all these sites (and also on a several pages on one or two sites) and wait for the page with this phrase to be indexed (on all sites). Then search for this phrase and hopefully your sites will be ranking for it in some order. Click on some and keep making note on which site, which position you clicked, including using <back> and then clicking again or clicking on another of your sites or another page of your site that ranks for the same term.

After a week or so, compare the results on impressions, CTR and position.

[hugh]

Am I missing something, where have the links to the SERPs gone?

Could one reason be that clicks and SERPS browsing from WMT have been skewing the SERPS if CTR is a factor?

[chewy]

quick look, this data is very different (and lesser in quantity and detail) than the data I am used to seeing and analyzing.

there are better tools out there.

This "offer" seems somewhat disingenuous because lesser experienced people might take it as gospel truth and it just isn't yet.

we know the big G can do better.

[HenryUK]

Interesting tool. On my v. large directory site there's good agreement between these stats and GA, at least on the key terms that I have checked to date.

I've been a little bit frustrated by the "<10" cut-off when looking at long tail terms, and I had hoped that time (that great healer) would fix this as more data accumulated.

However, I'm pretty sure that the earliest date available shifted from 11 March yesterday to 12 March today. Can anyone else verify this?

If Google are only providing one month's rolling data, well, I understand why, but it would be better if they let us have more data and left the interpretation to us.

In case anyone thinks I'm whinging, I should just say that I'm very grateful to have the tool at all!

[BradleyT]

One of the immediate benefits I saw was finding phrases where we are on page 2-3 with no SEO where it would be very worth it to do the SEO to get onto page 1.

[UserFriendly]

I'm also seeing numbers that are clearly approximations, with all of my "impressions" figures being rounded to the nearest hundred, and many terms having identical numbers of impressions.

So either Google aren't able to provide precise values due to the nature of their sprawling systems, or they're not willing to provide precise values.

[Bewenched]

I am SO loving this. I'm seeing in alot of our pages that we're running in positions 6-10 wonder what type of penalty that is?

[BillyS]

I was just looking at Analytics and WMT side by side. Same date range in each tool, filtering Analytics only for Google.

I would say there is some rounding but all comparisons were within 10%. For example, a term that might have returned 5,800 visitors in Analytics was showing 5,600 in WMT (less than 10% difference).

The one big gap I saw was total visitors. This probably has to do with low volume searches (<10). Understandably, a lot of long tail data is missing.

[aristotle]

Quote:
I'm seeing in alot of our pages that we're running in positions 6-10 wonder what type of penalty that is?

For many searches Google tends to reserve the first 5 slots for sites like Wikipedia, Amazon, and youtube. This could partly explain why a lot of your pages are 6-10.

[Atharva]

I find the data really strange.
for eg
My first top 20 queries are have impressions which are similar to other impressions. It is impossible. all 20 queries should have had 20 values of impressions. Also we are talking about 5K+ impressions, so data set is large.

For a specific keyword i am in either 1 or 2 positions in most google SE domains, yet it shows 95% of traffic from position 6 to 10

[filbiz]

I'm not seeing the positions for keywords anymore (keyword ranks).

What's going on?

Yeah I second the motion. I can't see the results either in the local or worldwide search.