And if so does it take this into account when that particular search term appears again?
Does Google remember/record how your broad match phrase performed for each individual search term it appeared for?
And if so does it take this into account when that particular search term appears again?
TravelSite, the short answers to your questions are 'Yes' and 'Yes'. ;)
Essentially, each 'variation' of a broad match has its own CTR, which is part of what determines it's position for searches in which it is eligible to show.
AWA
Please stop me when I make an incorrect statement.
1) From some threads that I have read recently, all broad matches start out as exact matches and must get some impressions (or clicks?) before they'll be allowed to expand. How much of that is accurate?
It has been my impression of late that a broad match like:
widgets of the green variety
will initially not match a query of:
a widget of the green variety
If that impression of mine is correct than I require many more keyword phrase combo than the broad match was originally intended to require.
If these restrictions indeed exist, do they affect phrase matching too?
I would like to be able to introduce, say a phrase match from the beginning that says "want a widget". Then I don't need to worry about whether "she" wants a widget, "he", "they", "i", "i really", "today" and on and on.
2) Now if I have a phrase or exact match of widgets of the green variety, will that not take precedence over the broad match variation during search term selection if someone searches on that very phrase? If so, my broad match of widgets of the green variety will never get an impression.
3) It appears that later, when the broad match is allowed some leeway, I start to turn up on some weird queries. Things the computer simply can't interpret correctly as a human mind would.
Recent, albeit limited, testing of mine supports both points #2 and #3.
4) My account is still yelling at me every day about having too many keywords. How many is too many? What will happen when I reach that limit. I've been trying to pare, bit by bit, those keywords that aren't a good fit in each AdGroup, but I'm simply not going to jeopardize my businesses by deleting them en masse.
5) What is really needed here is some very specific guidelines as to how broad matching and expanded broad matching are currenty working. Specifics like how are articles, prepositions, pronouns, compound words and numbers for words vs. the numbers themselves, etc. Do I need a seperate keyword for "type 7 widgets" vs. "type seven widgets"? Again, thus far my limited testing suggests that I do.
Also, which match types do whatever distinctions such as those cited above hold true for.
I raised some of these questions in a thread a while back, but I don't believe I saw an answer. Contacting support tends to just provide me with a swatch pasted out of the Help. I keep trying to explain that I don't sell tennis shoes :)
This would assist in being able to prune my campaigns better. For that matter, it will help a lot more people than just me.
Thank You!
patient2all
2) Now if I have a phrase or exact match of widgets of the green variety, will that not take precedence over the broad match variation during search term selection if someone searches on that very phrase? If so, my broad match of widgets of the green variety will never get an impression.
I have started my campaign a few days ago. I used every term as exact, phrase and broad matches.
The results are this:
Let's consider that the number of impressions is 100%, if the term is used as an exact match. Then my results show that the number of impressions for the phrase match is only 8-40% of impressions for exact match and less than 1% for broad match!
In the other words - my
[great widget] got 1000 impressions,
"great widget" got 80-400 impressions
great widget got 3-7 impressions
Everything is of course the same time period and the numbers are true for many words I have used.
So, the more targeted the phrase is, the more impressions it receives.
It is contradictory to all I have learned about AdWords :-(
Indeed it is contrary and it's upside down world there of late. I too often try using all 3 match types and notice that the exact match gets the most impressions and clicks.
Sometimes even when it is like a 6 word phrase which I find unlikely someone will type exactly in its entirety. It seems more likely they'd come "close" and qualify for the broad match.
As you suggest, this runs contrary to what AdWords help claims, you'll get the fewest impressions and clicks on the most targeted of the matching types, the exact match. It even makes sense too.
However, the stats indicate otherwise for both of us. Are the stats telling the wrong story or do we not quite understand human nature. I would think for a 6 word exact match, every article, pronoun has to be entered in the exact order. I would also think more people would mix that up than hit it on the nose. We're usually talking about titles in my case.
Now here is what we really need clarified. And the answer to this question would dictate whether it is sensible to include all 3 match types for the same phrase.
If all broad matches start out behaving as exact (which appears to be the case) and it takes a certain CTR and # of impressions before Google "trusts" the broad match enough to let it show expanded or truly broad. Now if the same version as an exact match gets all the impressions and clicks, will the broad version ever truly become broad?
Is the phrase itself taken into consideration with regards to when AdWords allows it to match more broadly regardless of the match type it is getting used in? That would justify using all 3 types for some keywords/phrases. Otherwise, one should perhaps only use broad match for a phrase they eventually want to match for additional terms.
Now, Dave, if you only used the broad match green widgets would you have picked up every single impression and click the combo of all 3 match types did? Or -- does the exact match take on a special precedence if it occurs in your keyword list and NOT in someone else who only used a broad match for green widget?
In my oldest campaign where I only knew about broad match, I find that keywords that have over 2000 impressions and a CTR > 5% will match to anything at all.
In a brand new campaign, neither the broad match or phrase match will show if the article "a" or "the", for instance, is used in the search query.
I am trying an AdGroup now with only broad match to see if the obvious candidate for exact match picks up a plausibly high enough # of impressions/clicks as it would have as an exact match.
Though based on your observations, at this point, I tend to concur with your results, that the more exact and targeted, the more impressions. The only problem with that is it makes too much sense for a AdWords algo to behave so properly given recent history :)
Thanks,
patient2all
Now here is what we really need clarified. And the answer to this question would dictate whether it is sensible to include all 3 match types for the same phrase.If all broad matches start out behaving as exact (which appears to be the case) and it takes a certain CTR and # of impressions before Google "trusts" the broad match enough to let it show expanded or truly broad. Now if the same version as an exact match gets all the impressions and clicks, will the broad version ever truly become broad?
I would say no, but maybe AWA can say more about it ...
I used every term as exact, phrase and broad matches.The results are this:
Let's consider that the number of impressions is 100%, if the term is used as an exact match. Then my results show that the number of impressions for the phrase match is only 8-40% of impressions for exact match and less than 1% for broad match!In the other words - my
[great widget] got 1000 impressions,
"great widget" got 80-400 impressions
great widget got 3-7 impressionsEverything is of course the same time period and the numbers are true for many words I have used.
So, the more targeted the phrase is, the more impressions it receives.
Dave,
I know this is an old one, but I can't get it out of my head. Your stats tell you about the hits your terms received, but what about the ones that it didn't.
I too have a habit of trying the same phrase with all 3 matching types and now wonder if I am doing myself a disservice.
We've been led to believe that the broad match behaves as an exact match until it gets 'x' number of impressions. In fact, I can see that is true when I test newer campaigns vs. older ones.
My older broad matches turn up for any permutation of the phrase and then some. A new broad match won't even turn up if preceded by an article or other throwaway word.
So, while your exact match [great widget] got the bulk of the impressions/clicks, your broad match of the same term, great widget hardly got any.
Therefore, if one searches on really great widget, will you show at all (assuming you don't have really great widget among your keywords?
I'm starting to believe not, because it won't show on the exact match. Of late, it appears that the phrase match too behaves as an exact until 'n' impressions and you don't have that many.
So my questions boils down to:
This will haunt me to the end of my days until I learn the answer. I'm slowly experimenting, but am doing so with money! Also, conditions aren't always the same between one AdGroup and the next.
patient2all
I have the same feeling as yours. A little while ago, when I tried to change match type I lost all the history of the keywords. Adwords considers [great widget] and great widget two different ones. I stoped the test. And now even if I feel like to change them from broad match to exact match I'm hesitating...
Just curious, when you say you "change" the match type, what is the reason that you change rather than add an additional match type for the same keyword? Is it to outwit the possible problem suggested above where the broad match may be more desirable than a more targeted match?
----
Also, for what it's worth, keep in mind that AdWord's differentation between matching types on the same phrase can be a two-edged sword.
If a keyword is in "Hold", "Trial" or "Disabled", in theory then, you get another chance at using the term with a different matching option.
Whether that helps or not is open to debate.
patient2all
1) It happened to me once, when I had one phrase with 2 different match types (I did it by oversight) both of them went on hold very quickly. Then I deleted both and added one as broad match. It lasts till now. Thus, I assumed that Adwords considered two different match types as one same keyword. It gets 'double impression' which shows only once when someone makes a search. This phrase goes very quickly to 'trial', 'hold' because of the poor CTR (half of actual CTR because of the doubled impression). This understanding may be wrong if your phrases with different match types are doing well?
2) Coincidence of what you said at the end. Last night I changed most of my ‘in trial’ phrases from broad match to exact match. Two of them went to ‘normal’ and rest of them went further to ‘on hold’ immediately. Ten minutes later, the two ‘normal’ ones went to ‘on hold’ as well. This morning, most of the changed phrases are back to ‘normal’. I need to keep one eye on those phrases to see how long it will last if the CTR is good enough. I'll let you know if something new happens. If you or any one has this kind of experience and/or testing, advice is welcome.
It gets 'double impression' which shows only once when someone makes a search. This phrase goes very quickly to 'trial', 'hold' because of the poor CTR (half of actual CTR because of the doubled impression).
Clynyi,
Yeah, that's yet another edge of that strange sword. One argument for using all 3 match types is that it prolongs the amount of time it *should* take for the words to go to Trial --> Hold --> Disabled.
Yet, we hear many instances of campaigns getting massacred right out of the gate by falling into "Hold" ostensibly based on earlier advertisers' poor experiences. Not fair, almost everyone agrees. That's generated one of the longer threads currently running on this forum -- "On Trial - On Hold MESS ". There's more commentary there than in the "Hat" thread which I haven't read yet.
The current "On Hold" policies are ruining campaigns for many advertisers, but that's another issue waiting on a response from the returning AWA.
----
Two of them went to ‘normal’ and rest of them went further to ‘on hold’ immediately. Ten minutes later, the two ‘normal’ ones went to ‘on hold’ as well. This morning, most of the changed phrases are back to ‘normal’.
The Official Explanation for how a keyword enters & exits "Hold" status:
Your in trial keyword limit has been reached; these keywords may enter "in trial" status only when one or more "in trial" keywords move to another state
Not everyone buys this, but I have seen an entire AdGroup's words in "Hold" at night, the next day, they are all "Normal". Now I have few "Trial" keywords at any given time, so it's not likely that enough of them went to "Normal" or got deleted (especially when I took no action) to explain 75 keywords jumping out of "Hold" suddenly.
The general consensus in the other thread is that the trial/hold status is not working properly, so I wouldn't try to make too much reason out of its behavior.
While many are having their best keywords denied them, in my particular niches the keywords that go to "Hold" turn out to be things I can live without. Upon relfection, they seem to be the less targeted ones so I just delete them and move on.
Oddly, my "Hold" keywords still show for searches. Is that normal?
Nonwithstanding my situation, I fully understand the nightmare people are experiencing when the keywords that best describe their business become unusable. And they become so for all advertisers and searchers, if we accept the stated premise that immediate or near immediate "Hold" reflects earlier, perhaps unrelated, campaign failures.
And Clynyi, if I were you, I'd try to come up with a new group of keywords, if possible. Try more longer phrases that people might be likely to search on, using all possible variants. It works for me sometimes. By searching referrer logs yesterday, I found about a dozen turns or phrase that I never would have thought of using, but people don't all think like me - I think that's a good thing :)
-------
Awaiting AWA' return, am anxious to hear responses to all this speculation, especially post # 8. I realize I've made that some point in prior posts, but never saw an AWA response.
Good Luck,
patient2all
Oddly, my "Hold" keywords still show for searches. Is that normal?
I've heard other people say the same thing, in this thread: [webmasterworld.com ]
I think they're still waiting to hear from AWA, too.
I heard someone said Adwords is like a book. I fully agree. The more time you spend on it and learn it, the more you will understand it and benefit from it
I suddenly have a feeling that Adwords is also like a woman. ;) Sometime you cannot explain why she's angry (disabling your kws) or upset (holding your kws). It may not be something ostensible that triggers her capricious mood, but something even herself may not be able to explain clearly either. You just need to spend a lot of time with her, getting to know her personality, talking with experienced people, with a lot of patience. Then your reward is that she will be very nice and tender to you (delivering quality leads with less cost). I’ve noticed that ‘disabled’ keywords turned to ‘normal’. I’ve noticed that the keywords easily to be on hold and trial are those have less advertisers but with good amount of quality leads. Most likely lots of other advertisers had these keywords once but gave them up after they were disabled, but those are the ones that convert the best. I cannot figure out why (if someone can, please advice), and I cannot figure out by which way I can definitely get the results that I want --- But I will, it needs time and a lot of tests/experiments. That’s why I like this board, we can exchange our experience, to make her and ourselves happier. :)
