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?

[Reno]

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

When you are viewing your full results in "Top Search Queries", click the little "+" sign next to each keyword phrase. You'll see the positions 1,2,3,4,5; then 6-10; then 2nd Page; finally 3rd Page+. Plus you'll see what pages are ranking for each phrase.

What's interesting about this for me is that the results seems to be a moving target -- for almost everyone of my individual keyword phrases, I have impressions in multiple positions, which if I read it correctly, means that if I check to see where I am at any given moment in time, I should not panic if it's moved a little lower. Next time I check, it may be up again.

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

[Eurydice]

One of my sites showed in Google for 3,025 keywords.

However, only 119 (3.9%)of those keywords actually got clicks. Does this finally end the talk of "long tail"? Input your own numbers here.

I took the 230 keywords for "fuzzy widgets", put them into Adwords, and waited four days. Only 15 KWs (6.5%) has Quality Score 7 or better.

The tool is good: it helped me to find a few more keywords.

On another note: There's a lot of cheering about the new version of this tool. Too bad. With the old tool, we were able to make charts of trends for keywords and visually show how they rose and fell over the year. No longer possible with the new version.

[robho]

One of my sites showed in Google for 3,025 keywords.

However, only 119 (3.9%)of those keywords actually got clicks. Does this finally end the talk of "long tail"? Input your own numbers here.

I don't see anything in the reports that would allow me to get those stats. Assuming you mean the "top search queries" (as the keywords report doesn't have clicks), queries with less than ten clicks have "<10". So there's no way to tell if that was zero clicks for the term, or 9.

FWIW, out of 6,175 top search queries for one of my sites, around a third got more than ten clicks, the rest got less than that (but not zero). 90% of the clicks were produced by the top one third of the terms, the remaining 10% of the total clicks by the other two-thirds.

Another of my sites with 3,600 top search queries has similar figures.

Further stat: the top 5% of terms (111 in the first example) produced 37% of the traffic. So the true long tale (the remaining 95% of terms) did produce a very useful 63% of all traffic...

[Eurydice]

If you sort by Clickthrough, you'll find that it will show down to 5 clicks. One KW has 28 impressions and 5 clicks. So "<10" doesn't mean 9. It apparently means "less than 5".

I put the minor KWs into Adwords. They got low Quality Scores. If you use those KWs, Google will punish your account by with fewer impressions and higher cost-per-click.

I compared WMT data against Google Analytics. I used March 14 to April 16 (to get whole days, not partial days) in both. In GA, I went to Traffic Sources | Search Engines | Google and selected Non-paid.

- WMT reports 3,298 KW phrases
- GA reports 9,409 KW phrases (2.85x more)

- WMT reports 8,100 clickthroughs (visits)
- GA reports 18,874 visits (2.3x more)

Forget the ballpark, it's not even in the same county. If you want to know KWs, use GA.

From GA, we find that of 9,409 KWs, 7,829 KWs (83%) had one click. These are not unique KWs; they are permutations of the top KWs.

The only interesting thing about WMT is ranking, but I don't see how that is useful. I see a KW bounces around quite a bit. We knew that "Avg. Pos." (Average Position) was an average; now we know the range. This means KW ranking tools are fairly useless. The statement "my KW is in position 4" is not an absolute fact.

All in all, it's interesting, but not useful.

[robho]

If you sort by Clickthrough, you'll find that it will show down to 5 clicks. One KW has 28 impressions and 5 clicks. So "<10" doesn't mean 9. It apparently means "less than 5".

You're right, I did see the 5's but forgot it was under 10. :-) Seems to be very approx at those levels anyway as it goes 5 clicks, 12 clicks, 16 clicks, etc in steps. So apparently just estimated and multiplied from a sample, all the way up.

As I mentioned further up the thread it does just seem like a sample compared to analytics, especially when you have a lot of search terms (as you do, and I have). And the factor they multiply the sample by is way off.

For example one of my sites on WMT has 6,175 terms and 110,000 clicks. Same period and site on Analytics is 415,000 keywords and 1.3 million visits from Google.

WMT missed 98.5% of the keywords, and 91.5% of the clicks!

So I'm coming to the same conclusion as you, sample size is so small (a couple of percent?) the results are near worthless, when there are a large number of search terms.

[Reno]

the results are near worthless, when there are a large number of search terms.

I do not use Analytics and my sites are quite small in comparison to robho & Eurdice and many others here. But even so, my hope is that the proportions from GWT are right, and if so, I'll be getting much of the information that I want.

For example, if in reality I have 1,000 terms and 5,000 clickthroughs and 50,000 impressions, but GWT is showing 500 terms and 2,500 CT and 25,000 impr, then the ratios are still right and it tells me I'm only getting 1 CT in 10 imp, which for my purposes is in and of itself valuable information (I need to improve my clickthrough rate!).

Having said that, if however the proportions/ratios are completely skewed, then yes, it's certainly more difficult to take meaningful corrective actions.

So we have credible testimony that the numbers are not exactly right -- is anyone noticing whether the GWT reported SERP positions seem accurate? That may be more difficult to gauge.

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

[dk82]

The SERP positions seem accurate. Doing my own search (with personalization off) has matched the reported position (+/- a couple) most of the time.

However, the reported clickthroughs are definately inaccurate. On most of my top search terms, GWT is reporting clicks 20-90% lower than Analytics. Given this large range, it even makes relative comparisons within the GWT data difficult to trust. The impressions are very low as well. Analytics shows nearly 1,000 click-throughs for a term that GWT shows only 130 IMPRESSIONS for!

I've confirmed with my logs that the Analytics data is pretty accurate, so perhaps the sample size from GWT is just too small for what they're attempting to extrapolate.

I was curious if certain geographic regions were being dropped in GWT (despite "all countries" selected), but US versus international search terms are both coming in low by varying degrees. The problem also appeared regardless of the selected period - 1 day, 1 week, and 1 month.

[wingslevel]

wmt shows about 1k clickthroughs per day for one site, but analytics is showing about 5k per day as organic google referals - quite a difference...

[freejung]

1script, I'm seeing the same thing. My impressions are exactly the same for every weekday: 27100. Overall clickthrough rate is almost always the same as well.

[freejung]

Something they appear not to be making clear: the reported position sometimes does and sometimes does not include image search results, and they don't give any indication which is which.

Evidence: there is one particular keyword that I completely own and have for years. I always show up at the top of the regular listings from any location, any proxy, any datacenter. However, the bulk of impressions for this term are rank 5 (and 6+, that's if I have a second page listed which I usually do). That is the position I occupy if you take image results into account, so clearly the image results are included, though they are not marked as such.

However, there are also some impressions for positions 1 and 2 (and a tiny handful for 3 and 4, presumably from personalized search) -- those are cases where no image results appeared, but again, they are not marked as such.

Thus if your keyword sometimes includes image results, your data for different positions may be sort of shuffled together, depending on whether image results were shown.

[Boulder90]

I can't find my position for search terms anymore. I just get CTR, %CTR and impressions.

[love2respectseo]

Really amazing to see the keywords listed over there in GWT and their position wise CTR and impressions. It is also showing your CTR according to se search engine position, i mean what CTR you got when your keyword was in 1st, 2nd, 3rd.... positions with how much amount of views it get from the search results.

The graph is a clear estimator of the rank and you can easily track out your progress there.

[Boulder90]

Ok, I found the postions. I still don't like the new interface. I prefer the other system for clarity. The less clicks, the better.

I'm getting like five different results in the serps that aren't matched up. For one search term I get:

4
5
6 to 10
2nd page
3rd page +

And then grouping of pages. It doesn't tell me what page is what result.

[canopenerheart]

Hugh - I couldn't find the SERP links at first either. You need to go into the "more" section of your top queries, and then you can click on them and it shows some nested info.

[Eurydice]

Boulder90 shows the reality of SERPs. Yes, it's more clear to say that your SERP is position 4. But the reality is a forest full of gray cats at dusk. Your page's position is a general smear across four or pages at Google.

It's more like the position of an electron in quantum mechanics: it's a probability.

Maybe this will finally get people to stop obsessing over position and PageRank value.

[aristotle]

On most searches, whatever ranks number 1 apparently gets the lion's share of the clicks.