Completely agree – (Yes I know I am bringing back on old thread here).

Separating match types into different campaigns allows you to control budget per each match type.

It is also quite easy to contol canniblization of exact into phrase/broad – especially when using adwords editor and excel and quickly uploading neg exact and neg phrase versions of your keywords into the broad campaign, and neg exact into your phrase.

After you gather enough historical data, you can see at the campaign level that Product XYZ never converted in exact, for example, but did in phrase and thus act accordinly.

Also, Google loves to spend your $$ on broad (yes, I know they serve the most restrictive match-type first). Having said that, you can do some interesting things with the data via campaign-level targeting:

different dayparts per match type
accelerated serving for Exact
etc

Lastly, as it has been stated here, match types and their respective CTRs vary greatly. In general, broad match CTRs will be much lower. Separating match types by campaigns (or ad groups), lets you identify which match types need ad tweaking/testing.

Already, a CTR of an ad doesn’t tell you the full story by virtue of having multiple keywords pointing to the same ad. Your CTR could be skewed by, for example, a low-volume keyword with a high ctr in the same adgroup as high volume keywords with average CTRs. Now, imagine throwing in different match types into this. You dont see the whole picture when optimizing ads.

Is it more complicated to run this way? Yes. But if you really want to know how your keyword Blue Widgets performs across all match types, just dump the data it into Excel and Pivot and you will see the data blended.

Also, Google will decide which keyword to serve based on ad rank (QS * max cpc).
I read that having the same keyword across three match types in the same ad group means that Google will first serve the most restrictive match type that matches the
query.

By contrast, having them sepereated by either ad group or campaign is when Google relies moreso on the ad rank formula than their “restrictive match type bot”. Thus, you can better control which match types are served by bidding greater for one match type vs another.

A broad match keyword can compete with exact when Google gets fancy with broad matching. Workaround is adding exact match negatives of every broad match keyword in the broad campaign. This forces the broad campaign to ONLY be used for broad matches, just as it should. You can’t do this with the proposed technique above.

Nice post. I completely agree that splitting out match types into their own ad groups (or campaigns) is a bad idea. As long as the structure of the account is well though-out so that keywords in each ad group are very tightly themed – as Kenny pointed out – then having all three match types in the same ad group shouldn’t present a problem.

Obviously some irrelevant searches might get broad-matched to your keywords, so you could argue that they might bring down the Quality Score of the whole ad group, but I think this is out-weighed by the benefits of a simpler account structure and an increased amount of ad group data made available for ad text analysis, insight and optimisation.

What’s more, now that we can run detailed search query reports at an ad group level, it’s relatively easy to identify broad-matched searches that are being matched to the ‘wrong’ ad groups (see my ’10% Clicks Rule’ post). The search query report also makes it relatively easy to add new keywords and negatives, so that only searches which are highly-relevant to that ad group’s keywords are triggering that ad group’s keywords.

So as long as the account structure is granular and closely themed, I’m all in favour of keeping broad, phrase and exact match types together.

Maybe you can help me with that: Google write´s “Listing the same keyword more than once in a single campaign works against you (our system interprets this as increased competition for this keyword, resulting in a higher CPC for it).”

How should I use matchtypes from that point of view. Separated in campaigns or use different mt in separate adgroups of the same campaign?

Thanks,
M.

I’ve used the campaign match type methodology, and I agree with you 100%

In theory it seemed very clever, but as you described, optimization was actually more difficult, not to mention the extra time downloading into Editor.

That being said, I do have a best practice suggestion that spans the fence, which I will post to my blog soon.

@trebuchet: I think testing different ad copy for different match types can be beneficial. This is why I think that separating match types into different ad groups can help improve performance in certain case, but I think separating them out into different campaigns has no direct benefits. The complication of reporting comes from having to pull cobble together a clear picture of a certain keyword’s match types by pulling them from completely different campaigns.

@Kevin Adams: Thanks! Yes, the term ‘embedded match’ is the same idea as Negative Exact, Negative Phrase, and Negative Broad. Didn’t mean to confuse anyone. You are so right about the search query report, but at least now Google is providing deeper data for this report (we wrote on the expansion of this report previously). Yes, exact match is certainly the least trafficked of all the match types. It is the most tightly focused, therefore having the shortest scope. I guess it is a good idea to cleanse your campaigns of exact match keywords that aren’t generating any impressions. Good point!

@Kenny: I agree with you. I think campaigns and ad groups are structured properly, then all 3 match types can reside in the same ad group. You can then adjust your bids at the keyword/match type level. Thank you!