Geo Targeting Questions

[jogrady]

I want to lower my CPC by targeting specific geographic areas with my high traffic keywords, but I have a few questions.

I use the broad match generic term Lawyers on a national campaign. I want to geographically target the same word using the following phrases in a national campaign.

Dallas Lawyers
Dallas, Texas Lawyers
Dallas, TX Lawyers
Texas Lawyers
TX Lawyers

and the reverse

Lawyers Dallas
Lawyers Dallas Texas
Lawyers Dallas TX
Lawyers Texas
Lawyers TX

(above with ALL CAPS and lower caps.)

My plan is to create a campaign for each state and an ad groups for each word category. (i.e. lawyers, Paralegal, Courts, etc.)

Questions:

How will the GEO terms compete with the national words?

Do I need to create phrases for TX and Texas, or is Google smart enough to know that they are the same keywords. Should I do all lower caps?

Do you know of any tool that creates all the possible word combinations for every city and state?

Is this approach flawed? I am open to suggestions.

This would be much easier if I could just create local campaigns for each of my high traffic keywords. My understanding of local campaigns is that I have to be in that specific area. If I searched Lawyers Dallas Texas from Detroit, Michigan my local campaign would not show up.

Thanks for any and all feedback. This forum has been an extremely valuable resource to me. I appreciate all the help I have received.

[jogrady]

I've done so more reading and here is what I have found. Let me know if you agree with my conclusions.

1.) Texas and TX would be considered two separate keywords.

2.) Broad matched terms Texas Lawyers and Lawyers Texas are considered the same word on Google Adwords, but different on organic searches.

3.) If Texas Lawyer (or some other variation) is searched the Texas Lawyer ad will display instead of broad match Lawyers. The caveat is that broad matched lawyers may display higher relevancy depending on the search phrase used.

Conclusion: I will create national campaigns for the following variations for every keyword, state and city.

Dallas Texas Lawyers
Dallas TX Lawyers
Dallas Lawyers
Texas Lawyers
TX Lawyers

It will be a lot of work, but I do not see any other way around it.

Let me know if you have any feedback.

[seobook]

thats one way to do it. also you can use targeting (as in state or city level) to target some of the ads.

[AussieWebmaster]

Actually depending if you are matching the terms broad or exact they will deliver differently. Texas Lawyers and Layers Texas are different unless local search has made this a different method... that we need AWR to answer.
But for general terms reversing terms are listed separately.
Best way to test is to run two different creatives for reversed terms and see or run one reverse and see if it is there for the other.

[orion]

Hi, jogrady.

I hope this help

1. "How will the GEO terms compete with the national words?"

Compare the following in Google and draw conclusions

texas lawyers
american lawyers

We have a geo targeting forum initiated by sfk poster at the Search Technology and Relevancy Section, but is a bit technical.

2. "Do I need to create phrases for TX and Texas, or is Google smart enough to know that they are the same keywords.? Should I do all lower caps?"

Texas and TX are different terms. Most users search in lower case using the default mode which is case-insensitive.

3. "Do you know of any tool that creates all the possible word combinations for every city and state?"

Google's default mode is FINDALL. This is a search without regard for order and proximity. Thus, term transposition is not that critical. It can be demonstrated that the absolute relative error due to transposition is very small for short queries and increases with long queries (> 3 terms).

Absolute relative error = ( | 2,270,000 - 2,290,000 | / 2,270,000 )*100 = 0.88%

Thus, the idea of permuting keywords seems to me a bit unnecessary for short queries. Besides, most users use short queries consisting of 2 or 3 terms.

Based on the above, I found keyword permutation tools unnecessary, but others are welcome to disagree with me.

If you still want to insist in both sequence "versions", try something like

...texas lawyers | texas...
...texas lawyers - texas..
...texas lawyers. Texas...

or derivative of these. In this way you cover both "versions" of the sequence.

The main idea is to disguise from users term repetitions by mixing these with delimiters. The optimization technique here consist in identifying which delimiters are ignored by the parser of the target engine. Try with document titles and in ads. With ads you are a bit out of luck since you need to limit your text to few lines.

4. "Is this approach flawed? I am open to suggestions. This would be much easier if I could just create local campaigns for each of my high traffic keywords. My understanding of local campaigns is that I have to be in that specific area. If I searched Lawyers Dallas Texas from Detroit, Michigan my local campaign would not show up."

Flawled is not the right term to describe your approach and you are welcome to post more often and learn from all the fine members around.

You may want to ponder the following:

1. Average users tend to query using natural language sequences. thus, texas lawyers sounds more natural than lawyers texas, but lawyers in texas is even more natural. However the "in" term is a stop word and will be ignored by the parser.

2. I don't live in Texas but in the Caribbean. When I query Google for texas lawyers and for lawyers texas I can see ads relevant to these terms without any problem. They are all Texas-specific.


Orion

[srini]

1. How will geo terms compete with the national words?

According to my knowledge, there is no competition element between geo terms and national words. It completely depends on the way you target the keywords.

Solution: Target your GEO keywords like tx lawyers, texas lawyer, dallas texas lawyer etc. and broad national keywords like
lawyers, lawyer, attorney etc. using Regional targeting (Texas Area) so that you will cover searchers who are searching with GEO keywords and also broad keywords (You are Maximizing Reach plus increasing traffic potential)

2. Orion has answered that question.

3. Do you know of any tool that creates all the possible word combinations for every city and state?

There is no ready made tool that gives you this. Try a few things mentioned below which proved effective for our clients.

a. There is a tool called "1-Click themes" by WEB CEO. This tool will give you all your competitor keywords upon specifying their
URL in the tool.
* This will give you all the competitor keywords by looking in their website title and meta tags.

Link: www.webceo.com

b. Download software called Keyword tumbler which will generate all the possible permutations and combinations possible.

Link: www.keywordtumbler.com

Example:

In the tool you need to enter words like lawyer, attorney, texas, dallas, austin....etc. then it will concatenate and generates combinations like:

texas lawyer
lawyer texas
texas dallas lawyer
dallas lawyer texas

* Very simple tool for generating combinations (You will save time & effort)

c. Visit Crossword Compiler and download their demo software.
Plug in your keywords and discover a multitude of additional words.

Link: www.crossword-compiler.com

4. Is this approach flawed? I am open to suggestions.

You need to try out some of the proven methods used by experts in the industry. I suggest forums are the best place where you would learn and disseminate information. I am sure you would do well and succeed.

All the best.

Hope my suggestions proved useful.

[orion]

1. Oops. Sorry, in my previous post I forget to indicate which terms goes with which results in the absolute error calculations.

As computed last night.

texas lawyers, 2,270,000 results
lawyers texas, 2,290,000 results

then,

Absolute relative error = ( |2,270,000 - 2,290,000|/2,270,000 )*100 = 0.88%

As computed today

texas lawyers, 2,300,000
lawyers texas, 2,310,000

then,

Absolute relative error = ( |2,300,000 - 2,310,000|/2,300,000)*100 = 0.43%

Results change over time.

2. As mentioned, if I query google.com for the above keywords from the Caribbean, I can see Texas-specific results and sponsored results (ads). However, if I query google.com.pr I still can see Texas-specific results but the ads changes. Still these ads are relevant to lawyers but no occurrence for the term texas is found in them.

If I conduct a local campaign in the Caribbean I would pay more attention to user's behaviors than to the ads themselves. Why? It is not a secret that here in the Caribbean users prefer to query google.com rather than google.com.pr It would be interesting to know if the same behavior is observed in other regional "googles".

Savvy users can avoid redirection to the regional google by disabling cookies and using any of these

http://www.google.com/search
http://www.google.com/webhp

3. Keyword permutation tools or any seo tool can be reviewed in the Beta Test section of the Forum. Developers and forum members are welcome.


Orion

[srini]

As computed last night.

texas lawyers, 2,270,000 results
lawyers texas, 2,290,000 results

then,

Absolute relative error = ( |2,270,000 - 2,290,000|/2,270,000 )*100 = 0.88%

As computed today

texas lawyers, 2,300,000
lawyers texas, 2,310,000

then,

Absolute relative error = ( |2,300,000 - 2,310,000|/2,300,000)*100 = 0.43%

Results change over time.

Hi Orion, this is lil technical, but could you explain me as to what is this absolute relative error and its relation to keyword combinations? How this can be utilized in the process of keyword expansion or selection?

[jogrady]

[QUOTE=AussieWebmaster]Actually depending if you are matching the terms broad or exact they will deliver differently. Texas Lawyers and Layers Texas are different unless local search has made this a different method... that we need AWR to answer.QUOTE]

Here is the official response from Google.

...Google doesn't recognize the difference between Dallas lawyers and lawyers Dallas - as long as we leave those words on broad match, you will capture all traffic for those words going either way. The same applies to the caps and lower case letters - there is no differentiation between the two, so we will be able to work with all lower case and know we would capture the traffic for both.

Creating a state campaign for each state and the ad groups based on category is a great strategy that should work well, since we can only geo target at the campaign level. We also want to be sure that you are not running the same keywords nationally as you are running geo targeted. For example, we would want to run lawyers as a broad match, but only target specific cities and states...

PS: Once again I am amazed at that professionalism of the posters in this forum. I find the SEO community surprisingly cooperative.

[orion]

Hi, srini

Hi Orion, this is lil technical, but could you explain me as to what is this absolute relative error and its relation to keyword combinations? How this can be utilized in the process of keyword expansion or selection?

This is related with the default query mode of Google, which is FINDALL. In this mode the retrieved documents should contain the queried terms without regard for sequence or proximity. Thus, term permutation should produce almost similar total number of results. Any deviation can be quantified in terms of the absolute relative error. The fact that the relative error is very small confirms indeed that ordering and sequence does not matter.

This is why jogrady got this response from Google


Here is the official response from Google.

...Google doesn't recognize the difference between Dallas lawyers and lawyers Dallas - as long as we leave those words on broad match, you will capture all traffic for those words going either way. The same applies to the caps and lower case letters - there is no differentiation between the two, so we will be able to work with all lower case and know we would capture the traffic for both.

However, if a user search in EXACT mode, this is a search with regard for ordering and sequence!!! That is, the results are a subset of the results retrieved in FINDALL. In such case, permutations do matter, but again, most users search in FINDALL mode.

All this is well-discussed in the Keywords Co-Occurrence and Semantic Connectivity thread.

I hope this help.

Orion

[AussieWebmaster]

[QUOTE=AussieWebmaster]Actually depending if you are matching the terms broad or exact they will deliver differently. Texas Lawyers and Layers Texas are different unless local search has made this a different method... that we need AWR to answer.QUOTE]

Here is the official response from Google.

...Google doesn't recognize the difference between Dallas lawyers and lawyers Dallas - as long as we leave those words on broad match, you will capture all traffic for those words going either way. The same applies to the caps and lower case letters - there is no differentiation between the two, so we will be able to work with all lower case and know we would capture the traffic for both.

Creating a state campaign for each state and the ad groups based on category is a great strategy that should work well, since we can only geo target at the campaign level. We also want to be sure that you are not running the same keywords nationally as you are running geo targeted. For example, we would want to run lawyers as a broad match, but only target specific cities and states...

PS: Once again I am amazed at that professionalism of the posters in this forum. I find the SEO community surprisingly cooperative. Broad match will get you both... though if you are not careful you can either get disabled for poor CTR if there are matches that do not apply to you... or you get traffic that is not convertible and thus are paying too much.

[srini]

Thanks orion, now I got it. For further reference I will also read the semantic thread in the forum. :)