| 1:35 am on Jan 2, 2004 (gmt 0)|
Yup, I had the exact same problem. Didn't use to work this way. It's actually a huge pain in the neck. No response from google yet.
| 1:53 am on Jan 2, 2004 (gmt 0)|
Is your daily budget set high enough to ensure that your ad will appear all the time?
Depending on your budget, your ad might show for only a portion of the possible searches as Google makes its best guess about how to ration your funds and spread your ad exposures throughout the day. Try raising your budget and see if that improves your coverage.
| 2:26 am on Jan 2, 2004 (gmt 0)|
I also have one questions, i have some keywords then use the exact match then its Impr is high, but when use the phrase match its Impr is zero. any one know how to resolve it?
| 2:33 pm on Jan 2, 2004 (gmt 0)|
I have noticed the exacxt same thing. No budget problems. When I complianed to Google I got this:
"I wanted to send you an update on your question about broad match
keywords. Anytime you have broad match keywords in your account, our
system will treat these broad match keywords as exact matches and then
based on performance, it will start expanding them to the keyword plus
another word. "
This is a change, a BIG change in how broadmatcing works! the word in question has an average CTR of 3-4%. If that is not 'good enough performance' to 'expand' I am in serious trouble. I spend $800-$1000 month and have been very happy with sales. But, I have about 500 keywords all with decent CTR's. It seems that my only option is to think of every iteration specifically and broad match those.
I don't have the time for this. It also seems to defeat the whol purpose of broad matching.
I Don't understand why they are making this change. As long as my broad matched terms are performing well why should I be penalized.
I plan on complaining loudly about this and I suggest that if this is important to you, you should do so as well.
| 2:48 pm on Jan 2, 2004 (gmt 0)|
gfguy - thanks.. wow! this is a major change and represents huge problems for me and I assume for many others. now I have the time consuming task of looking at all keyword iterations. but I suspect there is something else going on.. I've notices 2 ads on my keyword that ALWAYS show, even when the serach string consists of the keyword and a random gobboldygook word. Google can't say that they have "expanded" these ads to this random gobboldygook word.. almost seems like there is some kind og "wild card" matching that's available to some people! This whole thing is serious.
| 3:24 pm on Jan 2, 2004 (gmt 0)|
That is how broad-matching (as opposed to expanded broad matching) is supposed to work.
If I have a broad match phrase 'red widgets' I would like the ad to show in any search phrase which includes 'red widgets'. ESPECIALLY search strings I have not thought of.
If someone types in "I really want to buy a red widget" you better believe I want my ad to show. In the past, as long as my CTR was decent for 'red widget' and my budget was high enough my ad would show.
Now (if I understand correctly) there is a new threshold of performance necessary before my will ad will show on anytthing but 'red widget' .
According to google my only alternative is to think of EVERY itteration necessary.
So I am supposed to add:
I really want to buy a red widget
I really want to find a red widget
I would like to get a red widget
etc etc etc
| 4:32 pm on Jan 2, 2004 (gmt 0)|
|If I have a broad match phrase 'red widgets' I would like the ad to show in any search phrase which includes 'red widgets'. ESPECIALLY search strings I have not thought of. |
Yep, that's how it should work. But, I just searched for "blue widget price list blah", and my ad came up #1 based on 'widget'. And that particular keyword has a CTR of only 4.3%.
| 4:42 pm on Jan 2, 2004 (gmt 0)|
I thought at first that this was a glitch of some kind.
I have a CTR of 4% on 'red widgets' searching 'red widgets' always shows my ad...always.
Searching 'buy red widgets' NEVER shows my ad (now anyway. A few weeks ago and for a year previously it did)
Google said this:
"In your case, you have the keyword 'red widgets' which is first used as
an exact match in our system before being allowed to be used along with
other words like you mentioned in our phone conversation. Currently, your
broad match keyword ''red widgets" is not performing well enough to be
expanded to anything else than the exact variation of it. I would advise
you to add the different variations that you want to appear on, so you can
ensure that your ad display for variations like 'buy red widgets' or
| 4:57 pm on Jan 2, 2004 (gmt 0)|
Once again... not that it makes a difference...
Yet ANOTHER example of Google tinkering around and messing everything up. Can't they leave well enough alone? I advertised on Google for 1 1/2 years... never had a problem... rarely had to check on my account.
Ever since November... Ugh. Oh for the good old days when it was simple: you wanted to bid on 'red widget', you bid on 'red widget'. If you didn't want that keyword, you simply didn't bid on it. That simple. End of story.
Why, oh why, Google, do all of these new "features" only make things more difficult?
| 5:02 pm on Jan 2, 2004 (gmt 0)|
maybe trying to build profits in anticipation of the IPO. Sorry...that was too cynical of me.
I expect as the arena gets more crowded they are tweaking the logic to maximize revenues, and I am totally cool with that. BUT, I would rather they keep the logic simple and simply increase the CTR requirements and tighten the relevancy.
In my case (read posts above) There are (as there has always been) about 6 ads on 'red widgets' and there used to be 6 ads on 'buy red widgets' (and similar other combos) . Since they changed it there are NO adds on 'buy red widgets" or "get a red widget" etc etc.
Who is served by this?
| 5:56 pm on Jan 2, 2004 (gmt 0)|
I have also been noticing that keywords that used to get decent CTR's now don't even get impressions. I even pumped up my daily budget big time thinking that was the problem. Didn't help. They are still sitting there with 0 impressions. I even integrated the keywords into my ads (both the header and body), still nothing. I even tried using the keywords in common search phrases (I use overture's keyword tool which seems pretty good), and still got nothing. I just opened an Overture account and I'm going to experiment with that for a while. Even with the delays Overture has, if the results are what I want, and I feel that they will be, then Google can kiss my $1,000 a month goodbye.
| 6:03 pm on Jan 2, 2004 (gmt 0)|
Here's a test to see who is getting better ad placement than you (on broad or phrase match). Suppose you have a single keyword: widget or "widget"
Do a search on:
where the second word in the search phrase is just random gobboldygook. The ads that show up on this search are probably getting matching the way it USED to be before all the recent changes. And they are certainly getting displayed a lot more than yours! But, how do they do this? Do they know some "undocumented" matching option, or do they get special treatment from Google. Any ideas?
| 6:07 pm on Jan 2, 2004 (gmt 0)|
I would imagine it is based on CTR. Some of my words are still getting the broad match (confirmed with the gobledook test). The odd thing is that I am still getting the old-style broad match on words with a much lower CTR than some of the words I am no longer broad matching on.
It is very confusing.
| 6:39 pm on Jan 2, 2004 (gmt 0)|
CTR might have something to do with it, but I think it's more complex. I notice that many of my 2-word keywords do work well in the gobboldygook test (described in an earlier post).
So, Google seems to be much tougher on single word keywords. I think they are running some type of adaptive algorithm which looks at many factors and constantly "tunes" itself. Whatever they do I know this.. that on single word keywords on broad and phrase match a small number of other selected ads get much wider viewing than mine! How are these ads selected for this special treatment? Wish I knew!
Anybody else done research on this?
| 10:17 pm on Jan 5, 2004 (gmt 0)|
Curious to see if AWA can provide any further information about this new "broad match" acting as "exact match" until a certain CTR/performance level is acheived. True? False? or somewhere in between?
| 1:52 am on Jan 6, 2004 (gmt 0)|
toddski07, I'll see if I can get some particulars on this, and post again. Most likely on Tuesday afternoon.
| 8:26 pm on Jan 6, 2004 (gmt 0)|
I am having this same problem too - for the term "widget freezing" I have it set up as both broad and phrase matched.
I search for "widget freezing service" and find my ad does not show up.
Yep, this seems like it contradicts the broad matching methods as documented in the Google Adwords help sections.
The Google AdWords Keywords Suggestion system even shows "widget freezing service" as something that will show up for "widget freezing".
AWA, I am grateful to learn that you are working on this.
| 2:12 am on Jan 7, 2004 (gmt 0)|
OK, I'm back with some info. And a really long post.
I've been working on boiling down a very complex system into an accurate and coherent explanation. As I've said before, I'm not highly technical, nor do I have access to the complete algo involved.
With that said, I'll do my best. ;)
Yes, the algo has changed recently, but not substantially.
The changes affect 'new' keyword differently than 'established' keywords (meaning those with 1000+ impressions). And one might even picture the changes as an advantage.
For 'established' keywords (1000+ impressions), the change boils down to this:
Previously if a broad matched keyword worked well for exact searches, but very poorly in its broad variations, then it was highly likely that the keyword would simply be disabled, even though it worked well as an exact.
With the new algo, if the keyword works well for exact searches, but very poorly in its broad variations, then only the broad variations are disabled. The exact match continues to run. IMO, this is a real advantage to the advertiser.
Now, on to 'new' keywords. Most on this forum know that new keywords, since they have no history of their own, are started out by our system with an 'average' CTR.
With the new algo, our system recognizes that historically the 'average' CTR for some (but not all!) keywords has been poor. These 'risky' keywords will start out showing as an exact match only, as mentioned by gfguy. If they work well as exact matches, they can then start showing on broad variations. If these broad variations are successful, they continue to run, and if they are not, well, they are disabled.
In either case, if there are keywords for which your ad must appear, no questions asked, then by all means add those keywords as exact matches. If they meet the usual performance standard, measured at 1000 impressions, they'll continue to run.
Hope that makes sense.
Now, for a few dozen words about why the change was made. The intention is to improve the experience of the tens of millions of people who search on Google each day, by continually working to show them more relevant ads. One theme always running in the background here is that if we show relevant ads, our users will come to trust the ads over time and click with confidence. This is good for Google, but more important to you, it is also good for the advertiser.
| 2:21 am on Jan 7, 2004 (gmt 0)|
I understand the logic behind this - however, there are many people who know when they input a keyword how well it's going to preform to a fairly accurate degree. If an account averages 7-10% CTR, couldn't they be given the benefit of the doubt and start the KW as a broad match?
It seems quite subjective as to how an automated machine is going to read what humans think. We run a lot of small keywords in broadmatch which might get 1k impressions a month, and if for some reason these were exact matched, it might get 1k impressions in a year as the KWs are so niche.
In addition, I think there should be a flag on the KW so that the advertisers realize this is what is happening and even though they think the KW is broadmatched, it's actually exact matched.
Thanks for finding out what is causing these problems, good explantion.
| 4:11 pm on Jan 8, 2004 (gmt 0)|
Yikes...this new "broad match not really broad match but we are not telling you which ones" is troublesome.
My niche site sells all colors of particular widgets and gadgets. Therefore I broad match target the following terms:
I use a list of negative keywords to help reduce the ad impressions for searchs I know we are not interested in. I continually add to this list as new clickthrus occur for searchs not relevant.
Now with the recent changes to how broad match works...I have no idea if my ads are really showing for the different phrase searchs that are occuring. Google does not indicate in ANY WAY whether my specified BROAD MATCH term/phrase is really BROAD MATCHING or EXACT MATCHING! How the heck are we supposed to know? Instead, I am now having to consider adding all different phrase combinations using broad match to try and make sure the ads DO show for ones I am targeting.
For what used to be just...
Now I have to enter...
But then the same problem occurs on all these new broadmatch search phrases...what happens when someone searches for "buy red widgets"...well if my 'buy widgets' broadmatch phrase is not really showing as broadmatch then my ad will STILL NOT SHOW. This is a problem considering Google does not indicate which broadmatch phrase is out of the "trial exact match only" period. Plus...some widgets will take quite a bit of time to receive enough EXACT match impressions before they can even exit this trial period of Broadmatch. URgggghhh!
AWA...some sort of indicator needs to be placed next to the broadmatch phrase to indicate it is currently not at a full broadmatch status. For example just put an asterisk* after the keyword/phrase so we know what items are affected.
I completely understand that Google is trying to make the Adwords experience better for all parties involved...searchers, advertisers, investors, and Google itself. But please give us the tools to intelligently manage our accounts. At this point it is like shooting targets in the dark. (and the targets continue to change without Google telling us the rules on how play the game have changed!)
| 4:32 pm on Jan 8, 2004 (gmt 0)|
|AWA...some sort of indicator needs to be placed next to the broadmatch phrase to indicate it is currently not at a full broadmatch status. For example just put an asterisk* after the keyword/phrase so we know what items are affected. |
I've already informally shared the feedback in the posts above with others - and will do so again, more formally, on Friday.
I've gotten into the habit of sending along feedback, suggestions, gripes, and even compliments that I've heard on Forum 81 to the appropriate folks each Friday evening. And these folks have come to expect it (and dare I say, even look forward to it?).
I'm extremely pleased to say they actually read my notes!
So please know that your comments are heard.
| 5:24 pm on Jan 8, 2004 (gmt 0)|
Thank you for doing so!
Do your work for Google?
| 5:45 pm on Jan 8, 2004 (gmt 0)|
AWA's intro thread:
| 5:46 pm on Jan 8, 2004 (gmt 0)|
As always...thanks for listening to us and for passing on our comments. Hope to see the **improvements** continue. HINT HINT!