Do you use broad match or exact match on Adwords Keyword Tool when doing keyword research?

[debi_zyx]

I wasn't sure whether to classify this as a discussion or a question.  I'd love "the right answer" but I'm not sure if we're going to get one... Let's try.

When you use the Adwords Keyword Tool for doing keyword research, do you use the numbers from "broad match" or "exact match" when comparing relative search volume of keywords?  (And yes, I know the numbers need to be taken with a grain of salt, but when it comes down to it, you're using the numbers to compare and come to conclusions regarding the best keyword to use - so which match type gives you the data you're looking for?)

To be a little more specific - when you select "exact match" for, let's say the keyword "baking supplies", is that telling you how many people searched for that phrase within quotes <"baking supplies">, or how many people searched for only <baking supplies="">, as opposed to that word within a phrase <baking supplies="" stores="">or with the words reversed <supplies baking="">?</supplies></baking></baking>

Based on some keyword research we had done a year ago where any phrase reversals like <water bottles="">and <bottles water="">were coming out with the exact same numbers, even when it wasn't so intuitive that there would be the same search volume, we came to the conclusion - with the tentative suggestion of the SEOMoz staff on the old Q & A - that broad match would include all instances of the keywords in reverse order, so if you wanted to know how many people were searching for <water bottles="">only, you needed to use exact match.  </water></bottles></water>

That's what we did for about a year (I also think I saw Rand mention that somewhere in a presentation slide recently, although I could be mistaken and I don't recall exactly where it was to check it up) and then based on a recent forum discussion I had where someone was questioning that premise, I did another check with two KW reversals and while <water bottles="">and <bottles water="">still give the same number, <baking supplies="">and <supplies baking="">do not.  </supplies></baking></bottles></water>

So I'm left with a big question here as to what the best policy is.  Google Adwords Help is very vague on what the match type means in the tool (it seems to be talking about only your settings for your campaigns).  So - any input after this long saga?

Thanks!

[debi_zyx]

Thanks, Stubby.  Infomation appreciated.

All the best,

Aviva

[stubby]

I honestly don't remember the exact source.  I have done massive reading in the last few months, and I am sure it was in multiple sources.

One of the better books on Adwords is Advanced Google Adwords (although outdated, but an Updated Version of Advanced Google Adwords is coming in April)

If you waited this long, I would recommend waiting for the updated version, and then getting that.  It is always good to know something about any area where you are spending significant time and money.  I also would recommend getting it soon, as changes happen fast, and even the new book could be outdated in less than a year.

Surprisingly, your local library may have books on Adwords available to you as well depending on if you are near a medium to larger city.

[debi_zyx]

Thanks, Stubby!

That's what I've been assuming, but hadn't had any actual verification on.  Can I ask how you came to that conclusion?  Is it just what makes sense (100% agreed over here), or have you done analysis, or an official answer from someone who know at Google... or maybe you get visitations from the world beyond?  (I'm sure they know these things over there.  

Thanks,

Aviva

[stubby]

Hopefully you got your response long before now.

Just in case;

exact match is exact match.

So for

[baking supplies] it will show you how many people searched for [baking supplies].  No more, no less, no add ons, and in the same order.  It will not show you how many searched for "good baking supplies" nor will it show you how many searched for "supplies baking" or anything else.

phrase match is essentially exact match plus.

"baking supplies" in phrase match will show you how many people searched for "baking supplies", plus how many searched for "good baking supplies" and "baking supplies for chocolate chip muffins", etc.  It will not show you how many searched for "baking cheap supplies" or "supplies for baking"

broad match will show you everything regardless of the order, regardless of words, etc.  It is more like match soup.

so a broad search for baking supplies brings back every variation that includes those words and anything like it.  It is hard to know just how far this goes, but it could include "cooking ingredients for big weddings" as far as I can tell.

[debi_zyx]

Thanks, David.

Why would you think Google doesn't have answers to Google search query figures?  Technology is not fool-proof, so their tracking system might not be 100%, but why wouldn't they have or be able to have a record of what happens on their domain?

[DaveGerecht]

My PPC answer was focused on using the keywords for PPC. Google is not set up to focus on keywords for organic SEO purposes. I use the SEOMoz keyword tool BUT we generally have few direct competitors for our high end and niche products so they may be too general for your purposes.

To answer the question - I do not believe (with no stats or info to back me up) that Google has the specific answers as to how many people have looked for 'water bottle' and 'bottle water'. Wherever they do not have the accurate number they fill in what they 'guesstimate' the answer to be.

[debi_zyx]

Thanks, Chris.  It would be great if you could clarify a little - what is choosing to look at the metrics for "phrase match" of a keyword in the Google Adwords tool telling you, do you think? How many people are searching for that keyword as part of a phrase? If that's the case, then any number for keyword phrase <x y="">in phrase match should contain all the figures for <x y="">+ <x y="" z="">+ <w x="" y="">+ every other possible combination where you have an <x +="" y="">.  Bearing in mind that obviously the figures aren't 100%, does that seem to carry through on most searches?</x></w></x></x></x>

[debi_zyx]

Thanks, James.  Why would you add the different match types in?  What do you think each one of them is telling you?

Also, "I don't think any one can give a 100% answer to this question because every one has their own methods to what works and what does not work, people will find different methods which work better for them."  That's definitely true on a lot of things, but the question here is: what metric is Google giving you?  I assume they're probably giving the same metric to everyone, so while you might want to use different metrics depending on what your goals are, the question is why are you using them - what do you think they mean?

[debi_zyx]

Thanks for the thoughts, David.  Just wanted to clarify:

"there is a percentage of users who will put words into Google in an 'illogical' sequence." - Good point, and I definitely hear that a percentage do, even a good percentage, but - the exact same number? Really the same number of people look for <bottles water="">as they do <water bottles="">?</water></bottles>

" it's PPC - no click, no pay, so why worry too much" - I'm not quite sure what the point is here - meaning Google doesn't worry too much about giving accurate data? Or meaning one shouldn't care too much if the data is accurate?  That might be true for PPC, but I was talking about using the data for organic SEO planning purposes.

[debi_zyx]

Thanks all who answered, but maybe I should repeat the key element in the question, and if we could restrict the discussion to this question, how you deal with it and what your basis is for doing so, that would be much appreciated.

To be a little more specific - when you select "exact match" for, let's say the keyword "baking supplies", is that telling you how many people searched for that phrase within quotes <"baking supplies">, or how many people searched for only <baking supplies="">, as opposed to that word within a phrase <baking supplies="" stores="">or with the words reversed <supplies baking="">?</supplies></baking></baking>

Thanks!

[reilly3000]

So I believe it is really critical to use phrase match for getting to the truth for SEO campaigns specifically. Exact match can remove a lot of good long tail traffic, but broad match is always deceptively high. Consider the scenario where broad match may substitute "Wholesale Cooking Supplier" for the broad match keyword "Baking Supplies" - that is the kind of junk you get with broad match figures all the time. Phrase match is also what we use for checking search competition. Rankings of course always have to be checked using broad match- that is all clients care about isn't it?

[JamesNorquay]

I would advise on using exact match in one line of your research then also add in other match types to supplement your data and research, but then again I agree that we see huge problems with the data provided by Google you really need to mix up your keywords with your own research you can not rely 100% on what Google gives you.

I don't think any one can give a 100% answer to this question because every one has their own methods to what works and what does not work, people will find different methods which work better for them.

[DaveGerecht]

Hi Debi,

I was using broad search for a long time for a fashion based customer, but I too found that when I mixed up the words within a phrase I got the same numbers.

This kind of surprised me until I realised that there is a percentage of users who will put words into Google in an 'illogical' sequence.

And it's PPC - no click, no pay, so why worry too much?