Welcome to WebmasterWorld Guest from

Forum Moderators: buckworks & eWhisper & skibum

Message Too Old, No Replies

broad/phrase/exact - together or seperate



10:37 pm on Jan 19, 2009 (gmt 0)

5+ Year Member

I know this has been covered quite a lot here in this forum but I just can't get on the bandwagon of creating seperate groups for match types.

My method has always been to put all three match types in an adgroup and adjust bids or pause/delete match types where necessary. I want to see each keyword's performance and bids side by side, which would be difficult with seperate groups.

Also, for me exact match is not always the best performer. In many cases the broad is the best performer, so I don't feel that by seperating them out I am then grouping the best keywords together for improved ctr, quality score, etc. (I do understand that perhaps the stats are not perfectly accurate because of the matches being lumped together.)

Thirdly, I manage thousands of keywords and keep my adgroups very highly focused apart from match type. I can't imagine managing my account with 3 times the number of adgroups and adcopy. For the scale I am working on it just seems impracticle.

I just checked the adwords help pages and Google recommends starting with just broad and narrowing from there based on info gained from the search query performance report. I am building a new campaign for a client and I think I am going to try this method out, refining and adding phrase/exact matches later.

Is anyone else still struggling with this? Anyone made changes to the way they use match types that can share their experience?



6:28 pm on Jan 21, 2009 (gmt 0)

5+ Year Member

Sorry to bump this but it's really a big concern for me as I'm building new campaigns for clients.

Everyone here puts match types into their own adgroups?


9:01 pm on Jan 21, 2009 (gmt 0)

5+ Year Member

My ad groups consist of broad, phrase, exact and negative. I don't create specific ad groups depending on which match type i'm using because, to my knowledge, my ad groups might compete against eachother for search queries and i do not want that.

I think the best plan of attack is to create ad groups that consist of all match types, but the amount of keywords is narrowed down so that pretty much every keyword is present there in your ad copy.


10:08 pm on Jan 21, 2009 (gmt 0)

5+ Year Member

I've experimented with a) broad match only (+ negatives) in all my ad groups, b) broad/exact/phrase match for each keyword (+ negatives) in all my ad groups, and c) phrase match only (+ negatives) in all my ad groups.

When I switched to c) phrase only + negatives, my CTR went up 50% in one day. Has stayed at this level for a month now. Have switched all my clients to this approach, with similar results.

I used G Analytics (to see what keywords drove traffic and find negatives), G Reports (Search Query Performance to see what search queries triggered impressions and clicks and identify negatives), and G Keyword Tool (same idea).

It took 1 day to come up with a master spreadsheet of keywords (phrase + negatives), and 1/2 day to create new adgroups and keywords, using G Adwords Editor.

Hope this helps...


2:21 am on Jan 22, 2009 (gmt 0)

5+ Year Member

thanks bottomline, that does help. Many on this forum suggest seperate adgroups for each matchtype but it's not a managable format for me personally.

Based on your experience I will give the "phrase only" model a shot.


8:19 am on Jan 22, 2009 (gmt 0)

WebmasterWorld Senior Member 5+ Year Member

bottomline are you talking search or search and content?


8:28 am on Jan 22, 2009 (gmt 0)

5+ Year Member

What is your reason for using broad match keywords?

I use only exact, and keep a separate group of broad match terms (with all of the other terms in the account -ve matched) which is used as a keyword research tool.

The problem with using broad match in the way that you have been, is that you are triggering the same keyword twice, those two versions of the same term could have different CTR's and so a different QS, this means that a click on your broad match term could cost you a lot more then a click on your exact match term, when you couple that with clicks that are not actually relevant, that's a lot of money you don't need to spend.

ou really shouldn't need every keyword on all three match types, after all 1 broad match term is about equal to 100+ exact match terms, so you don't need to duplicate the entire account, take 3-4 (or more) ad groups that have a similar theme, and create 1 broad match group, and one exact match group, select the 1 or 2 broad match terms that represent each adgroup and put them in their own groups, then -ve match the exact match terms you have already established.

Of course what works best for each campaign is individual, but having all of the match types together isn't helping you do anything other then increase your costs.


12:27 pm on Jan 22, 2009 (gmt 0)

5+ Year Member

Khensu, I'm talking G Search, G Search Partners, and G Content Network. They are all in one campaign, and there is more than one campaign per account.

Search only CTR for the 15 days before the change averaged 3.3%. The same CTR averaged 6.4% for the 15 days after the change. Total CTR averaged 0.6% before, 1.4% after.

I gave a wrong piece of info to koncept. I said "CTR went up 50% in one day" when I should have said "CTR went up 94% in one day". Shouldn't matter, as it errs in the right direction.


3:02 pm on Jan 22, 2009 (gmt 0)

5+ Year Member

My reason for using broad match keywords is that they are converting well for me. I use the search query reports to add negatives where appropriate.


3:12 pm on Jan 22, 2009 (gmt 0)

5+ Year Member

exact match queries will convert just as well if you are using the correct search terms, and you will be able to control it better;


Broad match term = widgets - overall conversion rate= 15%- CPC 1.00
Search queries - Conversion rate
red widgets - 10%
blue widgets - 1%
green widgets - 20%

but if you had the three terms above on exact match as individual terms, you would find that

Red widgets - cpc 0.20
Blue widgets - cpc 1.50
green widgets - cpc 0.10

so both your conversion rate and costs could be reduced by using exact match terms correctly. You can establish what your effective exact match terms are by using the search query report.


3:33 pm on Jan 22, 2009 (gmt 0)

5+ Year Member

Thanks Yoshimi for outlining an example for me. That does make sense.

Lots to think about here. I'm definitely going to do some experimenting with this using the feedback I've received.


3:50 pm on Jan 22, 2009 (gmt 0)

5+ Year Member

Yoshimi, as I said your example really helped. My question now is how you would proceed when building a new campaign from scratch. Would you start with broad and then exchange for exact as you collect the necessary data to do so?


3:55 pm on Jan 22, 2009 (gmt 0)

5+ Year Member

I would brainstorm every term I could think of that would possibly relate to my product and put it as exact match, I usually end up filling a few hundred rows of a spreadsheet this way, using keyword tools, and my own knowledge and experience. I group these into the relevant ad groups and upload them, all on exact match.

I would then upload a broad match only adgroup, with no more than 3-4 broad match terms, and every other term in the account negatively matched. This way I know that I have all of the relevant terms, and a broad match mop up group to show me what keywords I have missed.


4:52 pm on Jan 22, 2009 (gmt 0)

5+ Year Member

when you add every other term in the account negatively matched to the broad adgroups (I will probably need more than one because of range of keywords I am targetting), do you add them just as the exact? ie: -[keyword] ?


6:23 pm on Jan 22, 2009 (gmt 0)

5+ Year Member

I set everything up as exact with the broad keywords in their own adgroup. However, upon thinking things over I realized that my client's service is one that people will likely be putting a city name in from of. ie: "toronto widget service" with many different city names possible and acceptable. Looks like I am going to have to start with phrase match for this reason, and then tighten things up from there.


7:53 pm on Jan 22, 2009 (gmt 0)

WebmasterWorld Senior Member 5+ Year Member


let me get this right.

you are saying

obama widget
"obama widget"
[obama widget]

"obama widget" is the best CTR? What about volume? in what comparison to broad?

Does that hold true for


I'll give you a closely guarded technique in exchange. (which you might already know but others may not)

If a site is opted out of placement targeting you can still access the ads on that site by putting the site in as a keyword.

Ergo you can put any site in as a keyword and garner extra impressions from that site.


8:12 pm on Jan 22, 2009 (gmt 0)

5+ Year Member

Ok, one last question on this topic... for now.

I have created all my adgroups as exact only. I want to duplicate all these for phrase match as well. Do you forsee any problem with creating a new campaign for the phrase match adgroups, rather than just adding them as more adgroups to the existing campaign? I ask, because when adding the [exact negatives] to the phrase match adgroups it would be easier to do that on the campaign level rather than adding the negs to each phrase match adgroup.

Is there any kind of truth to camaigns within the same account "competing" with each other?

Hope that was clear!


9:28 pm on Jan 22, 2009 (gmt 0)

WebmasterWorld Senior Member 5+ Year Member

I know you will get a dithering effect between them but I don't know it that can be viewed as an adverse event. (cut down your views)


10:02 pm on Jan 22, 2009 (gmt 0)

5+ Year Member


On your example of:
obama widget
"obama widget"
[obama widget]

"obama widget" would get the best CTR, with appropriate negatives like "hate". E.g. "I hate the obama widget" would not trigger the ad.

The same holds for:

keyword phrase
"keyword phrase"
[keyword phrase]

I don't understand your question about volume. Since CTR is clicks/impressions, I will assume volume means a) clicks and b) impressions.

Using Analytics source data for one client:

Before change date: Dec 14-29, 2008 (15 days)
Total of search-only clicks: 476
Total of search-only impressions: 17,847
CTR = 2.67%

Change date = Dec 30, 2008

After change date: Dec 31/08 - Jan 14/08 (15 days)
Total of search-only clicks: 1,644
Total of search-only impressions: 26,027
CTR = 6.32%

Note that there was a budget change for this client on Jan 1. All things being equal, budget change means increased impressions, and more clicks, but not a better CTR. I didn't change bidding strategy (manual CPC) or position preference (1-8) or ad scheduling or anything else.

Clients with small budgets improve as well as clients with large budgets.

Thanks for the tip on using site as a keyword. Another tool for the arsenal! However, as my username suggests, I really don't care about impressions. Or clicks for that matter.

Much more important is conversions. In business terms, they're called leads, and businesses live or die by the number of leads they get, the quality of the leads, the cost per lead, how many leads it takes to make a sale, and what is the ROI or ROAS.

My key metrics are # of conversions, conversion rate, cost per conversion, and ROAS. I optimize accounts to obtain the most conversions at the lowest cost. This is the bottom line in business; my customers care most about this, and therefore, so do I.


1:18 am on Jan 23, 2009 (gmt 0)

5+ Year Member


With your phrase-only technique, if you saw through the reports that a particular term such as blue widgets was getting considerable traffic and conversions would you then add it as an exact, or would you just be content to leave it as phrase match only?

I see suggestions on the forum that exact terms will recieve a lower cpc, better qs, etc. Have your experiments led you to phrase only with no further addition of exact when it seems appropriate?


3:07 am on Jan 23, 2009 (gmt 0)

5+ Year Member


If it ain't broke, why fix it? Or KISS (Keep It Simple, Silly).

Smart people, especially with an IT background, love to complicate things that are simple. Been there, done that.

If "blue widgets" works, why change to [blue widgets]? Exact match leaves on the table, for your competitors, these great search phrases:
- buy blue widgets
- buy blue widgets online
- buy blue widgets online now
- where can I buy blue widgets?
- wonderful blue widgets
- great blue widgets
- gift blue widgets
- blue widgets as a gift
- blue widgets for men
- blue widgets for women
- blue widgets for kids
- blue widgets for grandparents
- blue widgets in [name any city in the world]
- blue widgets stores in [name any city in the world]

If I sold blue widgets, I'd *want* that traffic. These are search queries of people in "buy" mode. They practically have their credit card ready.

Is exact match CPC cheaper? Don't know. However, a) show me real world stats proving this to be the case, and b) so what? CPC is the wrong metric. *And you are reaching the right customers.*

Measure the conversions, the cost per conversion, the conversion rate, and Return on Investment/Ad Spend. Write compelling ad copy, use Display URLs that are associated with ad/keyword combos of QS=10, create dedicated destination URLs (landing pages) with a compelling offer that are hyper-optimized for the keyword, and you'll get a QS of 8-10 pronto.

Then sell lots of blue widgets and make money.


5:26 am on Jan 23, 2009 (gmt 0)

WebmasterWorld Senior Member 5+ Year Member

OK you sold me!

Going to editor now.

Don't scoff at the URL as a keyword trick, it probably was responsible for $100K in profit last year and contributed greatly to my bottomline! ;). Dang you made me talk numbers now I'm going to get a Google spanking.

Good thing MB is not moding this, I think I can say Googlespank here?

Yes I meant volume of impressions.


6:38 pm on Jan 23, 2009 (gmt 0)

WebmasterWorld Senior Member 5+ Year Member

Well that worked and it didn't.

It worked in the US/CA campaign that much higher bid with a higher search focus but the content was also improved as a result of switching.

In the Global, CN/JP/IN, top middle east oil nations and western EU, it hurt volume of clicks by taking away the wild card factor of the broad. That has a lower bid more content focus with a small amount of search volume.


8:00 pm on Jan 23, 2009 (gmt 0)

5+ Year Member


Are you saying that for Global/CN/JP/IN/Middle East/Western EU, typical user behavior is to use web sites more than search? And what languages do you use to market in the these countries?

Just curious...


4:19 am on Jan 24, 2009 (gmt 0)

WebmasterWorld Senior Member 5+ Year Member

I have a clipart download site, all original art. 12 year old, Pre Google site with lots of page one organic search results.

The ROI on the Adwords "Global" is better on content while the US/CA is split pretty equal.

English, English and English.


12:32 pm on Jan 24, 2009 (gmt 0)

5+ Year Member

Interesting, and thanks. I think I may have mislead you, in error, for which I apologize.

This whole "phrase match" discussion assumed boosting *search CTR* was the goal, as this CTR affects QS, which affects position (among other things). Higher QS and position lowers CPC (if bid is unchanged). Lower CPC means more clicks. More clicks times existing conversion rate means more conversions, at a lower cost per conversion.

Content CTR is not affected by search CTR and does not affect QS. Content ads need a different strategy.

Google recommends splitting campaigns into Search and Content campaigns. I tried it in November for one client, one campaign, but implemented it poorly, so results were inconclusive. It was a maintenance nightmare.

Now I'm an AdWords Editor expert :). Should be easier to make mass changes. I'm going to experiment again.


11:35 am on Feb 18, 2009 (gmt 0)

5+ Year Member

Hi guys,

I couldn't find the thread which suggested using the same keyword as a negative in the same adgroup but a different match. For example:

blue widget
-"blue widget"

I tried this, but at the moment I am getting a message in Google Adwords telling me my ad isn't showing as a negative keyword is preventing it from showing. There were some conversions last week however.
I wonder does Google just pick up that the negative keyword is the same in the adgroup and flags this? Could it really affect my performance?


Featured Threads

Hot Threads This Week

Hot Threads This Month