Seems I recall reading enough SEO stuff as well as wise postings on WebmasterWorld that say there is some justification for using all 3 match types if it suits an advertiser's needs.
Let's assume for this example that all other factors are equal:
For instance, I believed that if the searcher hit your exact match perfectly, that the exact would trump any other match type for that same keyword term. Again, all things being equal, you may get that 'hit' as a lower rate. You may outrank other advertisers, again assuming all things being equal (other advertiser does not have an exact match).
OTOH, since it's sometimes hard to predict exactly how a searcher will phrase their query, there was a benefit in including the same term(s) as a broad match.
******
If there is some validity to the above, why does the Adwords Editor give an error message if I make "advanced bid changes" and the Adgroup contains a scenario as the one described above?
After making the changes, the Adwords Editor claims an "Error" with a yellow warning flag:
Warning: The keyword is the same as another keyword in the ad group with a different match type.
******
I can still post the changes but come away thinking I've done something foolish!
What's up with that?
Israel
Personally I always setup a campaign with broad match only first. Then within the first day or 2 I use my tracking software to identify search terms that I definately do not want to advertise on. And simply add exact/phrase match negatives to my campaign. This allows me to encompass most of the relevant search terms I want to show up for and not come up for undesired terms. I find this process of elimination far easier than coming up with every possible exact phrase I can think of.
Occasionally I'll also track a really good longtail that I know will convert well, so I create a seperate adgroup with an exact match to the keyword, a seperate advert specifically targeting the longtail and a seperate landing page.
Hope my notions help.
I do this because I don't want to have mispellings in the Ad copy because the searcher misspelled the term which got matched to a Broad match term.
I won't quote you on that since you said not to ;)
You do see some strange things in your logs with broad matches, like what was Google thinking?
******
I would imagine this still holds true, but I did find a while back that some of the "popular" longtail matches were really phony - and this was on Google.com (forget the Search Network & Content for now).
It seemed implausible to me that hundreds of people were matching an exact complicated model type # to a tee when I included the exact manufacturer's name of the product as a broad match. You'd figure there would be variants. Who gets 7-8 word syntax exactly right?
Always impressions too, never clicks which would seriously hurt my CTR. Someone suggested including globally some of the Advanced Google Operators as negative keywords.
People were using:
allinanchor
allintext
allintitle
allinurl
to search the SERPs for competitors or whatever and of course, the pages returned contained some Adword ads. And that counted as an impression.
I don't know how long your good longtails are, but if they're too specific, too often, then adding those as negatives will eliminate a lot of useless impressions. There are other Advanced Google Operators, but the names were general and I'd someday end up shooting myself in the foot were I to include them. Like "related" for instance.
I've got to watch my scattered self so I don't spend hours figuring out why "hormone free widgets" never shows when I made "free" a campaign level negative! I'm not organized enough as you to do it really right... So where plausible, I'll just toss all 3 match types in and see what it looks like after a few days.
******
SanDiego Art,
One thing I believed about dynamic insertion is that the "insertion" shows your keyword that triggered the ad, rather than what the searcher literally typed to generate a "match" for that keyword. Even in the case of broad match, AFAIK.
As you have no control over how the searcher phrases their query, dynamic insertion could be far too long if it relied on the "sentence" the searcher may type and would end up showing your default text.
If your broad match for dynamic insertion was buy blue widgets and the searcher typed wear kan i buy blue widgets, I'm fairly sure your ad would insert your keyword, buy blue widgets -- whether it was broad or phrase. And as long as you can spell, it will read correctly ;)
I'm pretty sure I tested that some time ago since I wondered the same thing.
Did they change that too? I've been away for a while and it used to be so easy....
******
Appreciate all input, I learn from every posting!
Thanks,
Israel
----
yes there is.... I'll try to explain a little bit more on this later...be right back.
We're waiting on your insight too, Tsinoy!
IMO, this is an interesting topic.
You must have a life to lead or something like that ;)
Israel
Quick history - this is what my numbers are with one of my accounts when I started... that's the data for about 1 month.
clicks 1,751
impr 61,392
ctr 2.85%
ave cpc $0.29
cost $509.90
after using the strategy somewhat similar to t_media... this is the current number... about 1.5 months worth of data.
clicks 28,857
impr 193,150
ctr 14.94%
cpc $0.32
cost $9,104.75
the cpc is higher now since these are mostly converting keywords and we can afford to pay higher... anyways..
here's what I do... if you think about what matters to google you can use that to your advantage.. it matters to them how well your ctr is... so if you spread out impressions across different keywords your CTR would improve dramatically... causing your cpc to go down.. which is what matters to you bottom line...
using all 3 match types would spread out the impressions for sure.. and whenever you look at your logs and find new keywords that you don't have in your accounts include them in your account in 3 match types and use whatever you don't want as negative keywords as well...
Start with that...
blue widget
[blue widget]
"blue widget"
your logs could show
buy blue widget
bug with blue x widget
blue widget
your first iteration should contain... the original and more..
blue widget
[blue widget]
"blue widget"
buy blue widget
[buy blue widget]
"buy blue widget"
-bugs
and keep on adding these keywords.. and eventually... you'll have tons of keywords with high ctr with low cpc... and you can also increase bid for those that convert well..to get higher ctr which will lower your cpc even more...
hmmm.. interestingly there seems to be a lot pay per click mentoring opportunities... maybe I should create my own mentoring program... hehehe... I wonder if there would be anyone who's interested...
anyways... hope that helps..
I've set this up to see how it works out. I'm really concerned about a drop in traffic so I'll give it a week to see what a difference it makes.
Thanks for the correction. I looked it up here:
[adwords.google.com...]
I was always under the impression that it inserted the actual query (if it fit).
israel, so what did you decide to do?
I added all three match types to one of my campaigns yesterday, and so far, our impressions are down, but our clicks are up, and our CTR, CPC, and cost per conversion has improved. Might be too early to really tell how adding all three match types will affect this campaign, but I just wanted to share the results thus far...
My joke about when you were coming back with a further explanation for your strategy was just that -- a joke. Sorry! I only get to visit here every week or so anyway.
It's heartening to see that you, T_media and likely many others employ roughly the same strategy that I do. Use the different match types, see what turns up and build an ideal refined list.
I could only wish I had Tsinoy's converting CTR once you've decided on the optimum list. I'm happy with 1/2 that.
******
Except as I found out recently (and I wonder if everyone knows this), that I can quadruple your CTR on certain pages - Parked Pages!
I seem to recall reading that parked pages at some point became part of the Search Partner network) (since some may search from those pages. So at someone's advice, I added domainsponsor to my list of excluded sites. Although I got away from Content long ago, it was said that certain exclusions would apply to the Search Partners, from whom I do get good clicks, AOL, cable/dsl home pages, etc.
However, this was not excluding parked sites. I found this out by accident when I stumbled in to Site and Category ExclusionŠPage Types. There was a spend breakdown and a CTR for each Campaign that I left in Search.
IMO, I've been "donating" hundreds of dollars each month to these pages with impossibly high CTR's -- often 50%
Now when I compare that to what I've learned are certain exact matches that are now time-proven, but are doing as well as one could expect with a 10% CTR.
******
C'mon is it indeed possible that people searching, clicking on and buying blue widgets are less likely sales candidates than people who end up on some defunct page or typo page?
Seeing the CTR on about 30 parked/typo pages that I own and I just let the clicks fall where they may with no hanky-panky, they aren't getting that type of CTR.
******
However, it wasn't until I checked off the box on the "Page Types" tab that I saw spend go down, profits go way up, but CTR go down a bit. "Parked Sites" clicks didn't represent too much of my spend, but it was still several hundred dollars USD a month?
Does anyone think many of these were sales?
It doesn't seem they were. People type a highly refined term, they want that item. People who stumble onto "Parked Sites" are often 5X more likely to buy? Can't accept it!
Did everyone know this except dumb me?
This probably should have been a new topic, it's a little off-topic, but a little on-topic since we were talking match types and then got into CTR. I only want meaningful CTR.
End of rant!
Thanks for everyone's insights!
Israel
