Brad Geddes / PPC Geek
Official Google Ads Seminar Leader.
Author of Advanced Google AdWords.
Co-Founder, Adalysis.
(703) 828-5811‬
Brad Geddes's Theories on Marketing Your Broad Match Keywords Are Not Converting Higher than Your...

Your Broad Match Keywords Are Not Converting Higher than Your Exact Match Keywords

I was recently looking at an AdWords account where:

  • The broad match conversion rates were higher than the phrase or exact match keywords
  • The broad match cost per conversion was much lower than the phrase or exact match keywords

The company had therefore bid the broad match much higher than the other match types.

Broad match does not convert higher than phrase or exact match.

There are variations of other search queries that are triggering the broad match keyword to be shown that are actually converting. (As with any blanket statement, there are a few rare exceptions).

Just a reminder, with broad match keywords, you ad could show for related search queries. Just because a keyword is related does not mean it will convert or you even offer that product or service. Broad match is useful to find new keywords because: “20% of the queries Google receives each day are ones we haven’t seen in at least 90 days, if at all?” – source.

For example, if your ad group had these statistics:

Keyword Match Type Conversion Rate Cost per Conversion
coffee mug broad 5% $5
coffee mugs exact 3% $7
coffee mugs phrase 2% $12
blue coffee mug exact 4% $12
red coffee mug exact 4% $6

What is actually happening is that if you were to look into the broad match keyword ‘coffee mug’ and see the variations of searches that is being triggered for this keyword,  you will have a chart like this:

Actual searches triggering the broad matched word ‘Coffee Mug’ Conversion Rate Cost per Conversion
Starbucks coffee mug 1% $9
travel coffee mug 6% $3
tea cup 0% N/A
car coffee mug 4% $6
coffee mug set 2% $8

If you were to see that data, you should add the keywords (and maybe a dedicated landing page) for travel coffee mug, and car coffee mug. There will be decisions to be made about keeping coffee mug set and Starbucks coffee mug. And you probably want to add the negative keyword –tea (and maybe –cup).

How do you learn the search queries for which your broad match keyword is being Displayed?

If you see your broad match converting higher than your other match types (and this is a good exercise to run regardless of broad match converting higher than other keywords):

  • Run a search query report. This report will show you the actual search queries which are triggering your ad to be displayed.
  • Look for the search query variations which are converting
    • Add these variations as keywords (either to that ad group or possibly a new one depending on how related the terms are)
  • Look for variations which are not converting
    • Add them as negative keywords
  • Change your bids so that your broad match is lower than your other keywords. If you bid broad match the highest, you are giving Google the ad serving control. If you have a phrase or exact match keyword – you know more about the user intent and cost per conversion – you should want the more specific variation to show.

Take Control of Your Ad Serving

Its very easy to give Google a significant amount of control for serving your ad. A liberal use of broad match keywords, or bidding broad match keywords higher than your more specific match types – essentially takes the control out of your hands which keyword will be shown for a given query.

You know your business better than Google.

You should know that when someone searches for a given query:

  • Which keyword you want matched
  • What ad copy you want shown
  • Which landing page the visitor will see after the click

If you are not controlling those elements, you have lost some control of your account. Take back the control and you should find yourself with a more successful AdWords campaign.


  1. Nathan
    April 3, 2009 at 11:38 am · Reply

    I agree on every point here – it’s just sometimes a little frustrating that a search query report is always incomplete. In niches that experience a great deal of disparate search, exact match is simply not an option. Not that your higher CPC exact match terms wouldn’t convert better, it’s just sometimes a case of not being able to uncover the terms as there may be so many variations.

  2. Tom Demers
    April 3, 2009 at 4:43 pm · Reply

    Great post. I think the value of differentiating a search query from a keyword is starting to really get some traction. I’m starting to see it productized (my company focuses on search queries, Click Equations also places a lot of emphasis on it, I’m sure there are others) and you even see Google tinkering with more resolution here (within the new AdWords interface and in the search query report itself: It seems like sort of a natural reaction to the “expansion” of broad match…anyway very interesting stuff, thanks!

  3. ouch
    April 24, 2010 at 10:53 am · Reply

    wish i knew this, like forever ago. makes total sense *slaps forehead*

  4. Sterling silver jewelry
    May 4, 2010 at 2:45 am · Reply

    Would you know how to use broad match?

    • Brad Geddes aka eWhisper
      May 7, 2010 at 9:43 am · Reply

      Use it in what method?

      It’s useful for catching long tail terms and more obscure search queries. It’s not useful when you can put more specific keywords into your account and then bid on those keywords based upon returns.

  5. gravymatt
    March 9, 2012 at 9:26 am · Reply

    “Look for the search query variations which are converting
    Add these variations as keywords (either to that ad group or possibly a new one depending on how related the terms are)”

    So do you advocate adding the variations strictly as exact or do you add them as phrase?

  6. brad
    March 9, 2012 at 9:29 am · Reply

    That depends on how much volume the variations have.

    If they are high volume words, usually exact. If they are low volume then phrase/modified broad match (or both exact and modified).

  7. david_suarez
    September 19, 2012 at 5:50 pm · Reply

    hey brad, I actually have this same exact scenario where my broad match terms are doing much better then most of my phrase and exact terms.

    Once I have enough data on the broad match terms, I can dig in and find the actual terms to add into an “alpha” campaign where it’s just one adgroup / KW.

    Would you justify this adding of a KW to its own adgroup by data that shows just 1 conversion, or would you at least need a few conversions before making the addition to its own adgroup?

    Also, would it be ok by going off just 1 conversion?

    Hope to get a reply back soon. thanks!

    P.S will be seeing you at your SMX workshop

  8. brad
    September 20, 2012 at 7:11 am · Reply

    I have a hard time justifying doing work based upon one conversion.

    Of course, the total number of conversions matters; but I like to see a pattern before really digging into a keyword and giving it its own ad group. The exception is if there are lots of clicks, and the word should convert, but its not – then giving it its own ad group/ad/landing page could be a good idea.

    But I’d rather see that a keyword has at least 2-5 conversions in the past 3 months and has received conversions in at least 2 of those 3 months or something like that.

    That is advice for a small account. For a large one, often I just skip one conversion click stuff as that’s so random, and focus on the words with either lots of spend and few conversions or consistent conversions over time and then add them.

    Hope that helps.

    See you at SMX.

  9. david_suarez
    September 20, 2012 at 4:01 pm · Reply

    you are awesome brad! thanks. im helping my father with his PPC account. i get a lot of 1 imp, 1 click, 1 conversions.

    Would you say these are one hit wonders or is there actual meaning behind this data.

    thanks again for the quick response!

  10. brad
    September 20, 2012 at 4:50 pm · Reply

    They are one hit wonders. 1 impression really doesn’t mean anything. Now if you go to 3 impression and 3 conversions I’d start to wonder, but usually until you have a few hundred impressions that data is correlational at best.

Leave a Reply