Dan.com

Google Adwords Keyword Tool Changes??

NameSilo
In using the tool over the last couple days, I've noticed that the tool is no longer distinguishing between singular and plural words. Words that are NEVER used in the plural suddenly show as super high just because the singular is a popular, real word. For example:

mesotheliomas 165,000 shows the same as..
mesothelioma 165,000

and...

hockeys 823,000 is apparently just as good as
hockey 823,000

same goes for keyword phrases.. for example:

domain names
domain name
domains names
domains name

are all equal and super popular showing 246,000 Exact Match, when we all know that is certainly NOT the case. I don't see how this is a good thing as it is completely inaccurate! It gives a poor indication of how keywords and phrases are actually used and searched for and even allows for made up words to show just as popular as the real thing. Anyone else seeing the same thing? Is it a temporary issue or is this a permanent change?
 
2 0
•••
The views expressed on this page by users and staff are their own, not those of NamePros.
It would make sense given what keyword planner is for...estimating Adwords traffic volume...

In Adwords, even when you bid exact match it matches queries including "close variants", which includes sing/plural and typos. Been that way over a year now.
https://adwords.googleblog.com/2014/08/close-variant-matching-for-all-exact.html

They're rolling out some changes in Adwords this year, they might have updated AKP to match.

Bing keyword tool still differentiates - https://www.bing.com/webmaster/diagnostics/keyword/research

And not a bulk tool, but Google trends still shows relative popularity of different variants.
 
Last edited:
Impact
1,678
Enly's Google Trends tip is a good one. Also GAKT [sld.tld] results may indicate singular vs plural keyword relative strength:

[mesothelioma.com] = 50
[mesotheliomas.com] = 0
[hockey.com] = 590
[hockeys.com] = 0
 
Enly's Google Trends tip is a good one. Also GAKT [sld.tld] results may indicate singular vs plural keyword relative strength:

[mesothelioma.com] = 50
[mesotheliomas.com] = 0
[hockey.com] = 590
[hockeys.com] = 0

The brackets [ ] were one of the first things I tried followed by quotes " " with the keywords but it makes no difference in the tool. I get the same numbers you posted with or without brackets. My guess is that it still works with the .tld on the end because it is considered a new string that has not been manually lumped together by the adwords team. This is a horrible change for determining exact popularity of a given phrase and I'm sure will be abused to pass off worthless names as super valuable. :(
 
If they recognize the tld as a domain name, it would make sense not to group them together because it's a navigational query and they're different sites. But for highly searched brands, they probably are.

I tried dropping out the variants by using them as negative keywords - that didn't change the search estimates, which is weird because negative keywords WILL filter them out in an Adwords campaign.**

I'm sure there are sellers not reading this thread who haven't a clue. And there will always be naive buyers.

*most of the time.
** and there will be angry mobs of advertisers if they change that!
 
Found a workaround. :)! Will write up the details after I check some things out and make some pictures :)
-------- UPDATE------------------

OK - this works, and afaik seems to be the only thing that works at the moment:

akp-screen.jpg

1) Enter your keywords in either the first or second AKP option, set options (if any) and search.
2) Add your keywords to a plan - either "Add all" or individually with the Add to plan >> arrow
3) Download the plan as a CSV.Here are the ungrouped search numbers. Note that the "download" option ABOVE the list of results will give you the grouped values.

akp-ideas-download.png


(I didn't bother trying to see if I could get exact matches for this example, but this shows how the numbers can be split out to restore some sanity to the search.)
 
Last edited:

Donnyd

Established Member
Impact
105
I emailed letting them about how this is not correct. I will not go into how important this is for est traffic.

Only a handful of people know how to use this correctly but right now it is messed up. I hear bing has not screwed it up yet LOL... Anyway rare to see this talked about anywhere. But you really have to be careful on new extensions also they are grouping those together also. Here is an example. take the new extension .insurance. GK is showing auto.insurance and auto insurance as same results. Imagine as a buyer you count on those results as being accurate. Anyway great subject to talk about for people that see how important this is to domain buying.
 
Probably won't get a reply. You might have more luck posting in Google's Adwords product forums, but since Adwords automatically matches close variants there is no more "exact" in the Adwords world. Giving numbers that include close variants parallels the way these keywords would behave in an Adwords search campaign.

They don't care about people using it for other purposes - in fact, they are now showing a message if you don't have an active campaign in the account you're using. It goes away if you persist, but the fact that they have a special message makes me think it will become permanent very soon.
 
It seems the current tool is grouping like words and phrases together, for example:

auto insurance 301,000
automotive insurance 301,000
automobile insurance 301,000

all show 301,000 exacts.. when we know most of that is really just "auto insurance". I've seen it do this for completely different words too that somehow are getting grouped like this for related terms and showing completely wrong exact scores.
 
It seems the current tool is grouping like words and phrases together, for example:

auto insurance 301,000
automotive insurance 301,000
automobile insurance 301,000

all show 301,000 exacts.. when we know most of that is really just "auto insurance". I've seen it do this for completely different words too that somehow are getting grouped like this for related terms and showing completely wrong exact scores.

Exactly.Those are "close variants", it's not a true "exact score." Read my earlier posts in this thread for the workaround (which was still working the other day.)

If you're an Adwords advertiser and you bid on exact match [Auto Insurance], they will also show your ad for "automobile insurance", "automotive insurance", "autos insurance", "auto isurance" ... That's how things work in Adwords these days, the keyword planner now seems to reflect that.
 
Last edited:
Exactly.Those are "close variants", it's not a true "exact score." Read my earlier posts in this thread for the workaround (which was still working the other day.)

If you're an Adwords advertiser and you bid on exact match [Auto Insurance], they will also show your ad for "automobile insurance", "automotive insurance", "autos insurance", "auto isurance" ... That's how things work in Adwords these days, the keyword planner now seems to reflect that.

Workaround still does not work for all terms and is inconsistent. Exact match scores even by adding to a plan, marking as "exact", and then exporting still gives results that seem to incorporate variants all as one single term. For example try these and see that the variant actually gets the same exact match as the main keyword, term, or brand.. even with the workaround:

Keyword Type Exact Monthly Searches
spotify 16,600,000 18% 0.17
spotifys 16,600,000 18% 0.17

android 2,740,000 13% 0.50
androidal 2,740,000 13% 0.50

popcorn 1,220,000 3% 0.35
popocorn 1,220,000 3% 0.35
 
If my examples aren't working, they closed the loophole. Choosing match type did nothing even when I wrote that.

Moz new keyword tool is good. Or use bing and multiply results.
 
For those who didn't take my word for what's going on ;):
http://www.thesempost.com/googles-keyword-planner-now-combines-keywords-for-search-volume/

Wouldn't be so bad if adding negative keywords worked, (advertisers use negative keywords to keep ads from displaying for certain terms) but they're not being factored into the totals. So you can't use them to subtract out the phrases you don't want.

So be VERY careful if you use AKP or any tool that relies upon it - chances are search volumes shown are much higher than they actually are.
 
Last edited:
Forget google for exact search anymore. Using a combination of Bing (though you must verify at least one website first to use the tool) and Wordtracker, you can still get a relative idea of what the best keywords are. Still, its not as good as the google exact match used to be.