The Phrase Match Keyword is Not Dead

phrase match keywords cute kids

Should We Ditch the Phrase Match Keyword from Adwords and Bing?

phrase match keywords cute kids

In our most recent 2014 Keyword Bidding PPCChat on twitter, I noticed a surprising number of people discussing the fact that they have switched entirely to Broad Match Modified (BMM or MBM, whichever you prefer!) and Exact (E) Match only in their strategies. I’ve heard this before, but I’m increasingly surprised by the number of people ditching the Phrase (P) Match. From what I can tell of the P Match, it’s been nothing personal. He’s still allowed to show up at parties and eat cake, but for the most part he sits awkwardly alone in the corner.

Edit 4/7/14: The inestimable Bryant Garvin has some great thoughts on his post in answer to this: Phrase Match is Dead. Go check it out! I still think I’m right, 🙂 but he of course makes some great points on account optimization and simplicity!

Initially, I was hesitant to admit that I still utilize Phrase Match since I feel the PPCChat’ers are about as advanced as they come in the PPC world. However, after I tiptoed into the conversation, I began to realize there are others out there like me. There are those of us still willing to befriend the P Match Keyword and utilize it in our accounts. I’m hoping this post encourages discussion so we can all learn more about Match Types and grow in our own personal strategies.

3 Ways in Which the Phrase Match Keyword is Still Essential

baby out with the bathwater in ppc marketing

In beginning my argument for the continued use of the Phrase Match Keyword, let me initially say that I agree with a lot of the BMM/E Only position. What I’ve heard so far about better account organization and optimization with less keywords makes a lot of sense and I think that is a “generally” good rule to follow. Therefore, please hear me in that I’m not saying “ignore that argument”.

What I am saying is that I believe never using the Phrase Match Keyword is to throw the baby out with the bathwater. Is a BMM/E match strategy good? Yes. Are there times where you still need P Match in your accounts? I would argue, absolutely. Here are three instances in which I believe a P Match Keyword is still useful.

Multi-Word Brand/Product Names

There is a substantial difference between the headphone maker: “skull candy” and the query “halloween candy shaped like a skull”. (I realize Skullcandy is really only one word, but the spaced variation is one I would add into my account to allow for better optimization).

Ok, so why not just utilize negatives for the BMM term +skull +candy? Boom, problem solved.

Here’s the problem I have with that thought.

You can’t distinguish, for bidding and optimization purposes, between the more “branded” P Match queries you receive, and the multitude of impressions you would get for the BMM Match queries. You are forced to treat all queries (both clicked on and non-clicked) the same, with your BMM query. Since you are dealing with possible quality clicks/impressions and possible completely unrelated clicks/impressions, you are putting yourself at a severe disadvantage for further optimization. You will end up getting far worse CTR for P Match queries within your single BMM Keyword than you would if you split out the P Match and were able to optimize and SQR them separately.

For those reasons, I utilize all three main match types (BMM, P, E) using tiered bidding or separate ad groups with mirrored negatives to eliminate this main concern (see more on mirrored negatives in point 3).

Some Multi-Word Keyword Phrases

This logic is similar to the Brand Name above. For Multi-Word Brand Names, I will go out on a limb and say I always think P Match should be involved in the strategy. The non-branded terms are slightly less obvious and require common sense. The same rule applies here though. Let me give an example:

Let’s say you sell children’s playhouses. You bid on the term “play house” to capture those actually intending to include the P match term “play house” in rather than any variation of keyword including the word +play and the word +house. I.e., “What are games for my children to play in the house” “rainy day play games for in the house” blah blah, the long tail terms are nearly endless for BMM query variations of +play and +house that are completely unrelated to an actual “play house”.

Again, for non-branded terms like this with a high propensity for non-quality traffic, I tend to utilize all three main match types (BMM, P, E) using tiered bidding or separate ad groups with mirrored negatives to eliminate the 2 concerns I brought up above (see more on mirrored negatives in point 3).

Mirrored Negatives for Top Terms

The final example I will offer in suggesting that the P Match term is still useful is the proper utilization of mirrored negatives between the BMM and P terms (Exact as well, but that’s a given). In the two examples I gave above, I believe utilizing mirrored negatives to your advantage will offer significant ongoing optimization help. Now, in my example below I will break the Play House ad groups into 3 separate ad groups for each match Type. I’m not suggesting you do this for every term as that would unnecessarily complicate your campaign. I am suggesting that, for those high traffic terms with greater opportunity for optimization, you consider breaking them into individual Match Types.
old abandoned tree house

Overly Simplified Example:

Ad Group: Play House (E)
[play house] [play houses]

Ad Group: Play House (P)
“play house”
“play houses”

-[play house] -[play houses]

Ad Group: Play House (BMM)
+play +house
+play +houses

-[play house] -[play houses] -“play house”
-“play houses”

What this does is to keep your more qualified visits in the top two ad groups.

  1. It allows you to bid separately and generally higher for the P Match terms which tend to be more focused on your target audience.
  2. It allows you to better identify and exclude low-quality SQR traffic since your P match query traffic is pulled out of the BMM Keyword.
  3. It allows your P Match CTR to be higher than if it was all lumped into the BMM term.

My (P) Match SQRs are generally far more qualified than my (BMM) to begin with, so this allows me to keep BMM at a lower bid and better manage it in its proper place… as a source of potentially new terms in random/various word order without infringing on my more-often-qualified-because-of-word-order P Match term.


In closing, I will reiterate that I think generally the BMM/E Only argument is a solid one (from what I have heard so far). However, I’m still not willing to give up my Phrase Match keyword in these 3 specific instances. What do you think? Agree or disagree? Let me know if you have any other examples you would add to these three or if you have reasons you don’t think these three hold weight against the BMM/E Only argument.

Tweet me your thoughts at @PPCKirk and I’d love to continue the discussion!