This is a step by step guide to setting up Brand and Non Brand keyword segments for a complex brand – when you have multiple brand keywords and where your brand keywords mirror your non brand keywords.
I have used the example of a client with a number of sub brands to demonstrate. Pay attention to the difference between AND and OR statements in the examples. As a non-developer, I find them to be tricky little weasels. There are four parts to the process:
- Choose the right brand keywords
- Set up a custom segment for brand keywords
- Set up a custom segment for non-brand keywords
- Check that your segment numbers add up
1. Choose the right brand keywords
The process of setting up brand keywords is straightforward when you have a single obvious brand name, but there is more subtlety required in choosing keywords in more complex situations. For a client such as alh.hr with a range of hotels in Dubrovnik, Croatia, where hotels have names such as "Hotel Dubrovnik Palace" which mirror the non-brand search term "hotel Dubrovnik", we have to extract just the unique part of the brand name and use it as the brand keyword. So from their Global Brand and list of Hotel names
- Adriatic Luxury Hotels (ALH)
- Hotel Excelsior Dubrovnik
- Hotel Dubrovnik Palace
- Hotel Bellevue Dubrovnik
- Hotel Kompas Dubrovnik
- Grand Hotel Bonavia Rijeka
- Villa Agave
I extract the terms Excelsior, Palace, Bellevue, Kompas, Bonavia and Agave. I add to this list their global brand name ALH and its full written name Adriatic Luxury Hotels. Sometimes you need to make trade-offs based on searcher intent. You will note that Adriatic Luxury Hotels can be both a brand and non-brand keyword and we need to make a decision on how to classify this. I believe that keeping it as a brand keyword is the alternative most reflective of searcher intent and we will have to live with the few who use it as a non-brand keyword. We now have eight brand keywords to set up in our segments.
2. Setting up a custom segment for Brand keywords
- In Google Analytics, click on the "All visits" drop down to access advanced segments
- Click "create a new advanced segment"
- Search for "keyword" in the box
- Drag and drop the green keyword field from the left into the right hand side box
- Set condition to contains
- Add the brand keyword as a value
- Add all brand keywords similarly with OR statements
- Add an AND statement
- Search for "medium" in the box
- Drag and drop the green medium field from the left into the right hand side box
- Set condition to matches exactly
- Add organic as the value
- Add and OR statement
- Search for "medium" in the box
- Drag and drop the green medium field from the left into the right hand side box
- Set condition to matches exactly
- Add cpc as the value
- Name your segment – Brand Keywords
- Test your segment by clicking the test segment button
It should look like something like this:
3. Setting up a custom segment for Non Brand keywords:
- In Google Analytics, click on the "All visits" drop down to access advanced segments
- Click "create a new advanced segment"
- Search for "keyword" in the box
- Drag and drop the green keyword field from the left into the right hand side box
- Set condition to does not contain
- Add the brand keyword as a value
- Add all brand keywords similarly with AND statements
- Add an AND statement
- Search for "medium" in the box
- Drag and drop the green medium field from the left into the right hand side box
- Set condition to matches exactly
- Add organic as the value
- Add and OR statement
- Search for "medium" in the box
- Drag and drop the green medium field from the left into the right hand side box
- Set condition to matches exactly
- Add cpc as the value
It should look something like this
4. Check, double check and sense check your segments.
Experience has taught me to make sure my segments are correct before using them.
If you have created your segment as above, you should see this in you Custom Segments box
Do a sense check.
Number of brand keywords + number of non-brand keywords should = all keywords.
For ALH I found that keyword numbers didn't add up because people were searching in Croatian for the website and I hadn't included Croatian versions of the brand terms! I had to go back and put in Croatian language brand terms.
Check that values on a graph add up
Put both segments on and go to Traffic sources > Search engines. Scan a number of data points and make sure they add up. Simple.
Check that your actual keywords fall in the correct bucket.
Put both segments on and go to Traffic sources > Keywords. Scan a number of data points and make sure each keywords falls in the expected bucket.
These three very simple checks will save you a lot of headaches when working with advanced segments.
---------------------------------
This is my process when working with Brand segments, if you have other solutions please share them in the comments. I find this tricky to get my head around and would love some more tips
Note: You can only have a max of 20 AND or OR statements per segments and you can only have a max of 100 advanced segments per profile. You have the option in keywords "matches regular expressions" which can help you work around these limits.
Awesome tips. I’ll be passing this post on for sure
Great stuff! :) I do this a little bit differently, but same concept.
Export all organic search terms into Excel.
Use the following formula:
=IF(ISNUMBER(FIND("beginning of brand keyword",A2)),"Brand", "Acquisition")
For example, if the brand is Walgreens, we would use something like "Walg" to capture any misspellings. Isn't perfect by any means, but works alright.
Then I filter from z --> a in that column. Go down to where acquisition starts and insert a row above. Do a new formula starting in acquisition with your next brand phrase. Filter and then continue this process until you've gone through all brand phrases.
Then select the brand v. acquisition, visits, revenue, and any other desired columns. Create a pivot table and you can see what drove all of those metrics (whether brand or acquisition).
A good alternative for those who hate custom reports in Analytics or are using another analytics platform.
Using regular expressions when setting up the filter is critical IMHO. If you are fitering out branded terms then most likely you are working with a brand that has a strong presence. And in my experience this requires at least 20+ variations plus misspellings of the brand name. So I use something like, brand|brnd|brnnd... and so on. I keep this regular expression saved in a .txt file too, just in case you need to add to it over time.
Great post Stephen! Very well structured and explained. I also agree with Miguel for using .txt with all variations/misspellings in RegEx – quite convenient. I’ll definitely start using that.
Brother can throw down a how-to article. Love this. Between this and your great link building article, you're putting together a solid reputation of producing awesome info. Thanks again, Stephen.
Yes, I agree on both counts. Great work! Thank you.
Segmenting is a good tool and can save you plenty of time in the future! Nice tips for sure - thanks.
Great post. I love how there are so many untouched nuggets of gold left in Googla Analytics still be discovered!
Great post. Definitely a time saver and good for new ideas
Nice post & thanks :)
Unfortunately this guide is no longer very useful, because google hides the keywords in queries for users that are logged in to Google, leaving only the label "(not provided)" to work with instead of the actual keyword.
I've linked my site to Webmaster Tools, which should help, but I can't see any effect of this in Analytics (I can get some fairly good statistics in webmaster tools, but then I lose all of the filtering and segmenting powers of Analytics).
What's the easiest way to do this nowadays ?
Great post thanks for the helpful information I am on my way to update analytics right now
Nice post, Stephen!
For the German users out there: a couple of weeks ago I created an article about how to distinguish between brand traffic and non-brand traffic. There's also a German tutorial for creating those Advanced Segments in Google Analytics: >Web Analytics: Brand Traffic von Non-Brand Traffic unterscheiden.
Stephen,
Wow, thank you for posting this. This is a fantastic trail of how-to and I look forward to learning from this as I have an upcoming client campaign that needs this level of detail. Thank you so much for the post!!
Thanks for the walk through, it's important to understand where this traffic is coming from in order to make calculated decisions.
Interesting, I just implemented last week that on our company blog!
Nice post, I think I will document my own process on my blog :)
Realy Great post. I love how there are so many untouched nuggets of gold left in Googla Analytics still be discovered!
Do you know how many keywords you can exclude? I find I get error messages if I'm segmenting out more than 10, and I have some clients with literally hundreds of brand names (all lawyers at a single law firm and each lawyer is considered a brand). Any tips for filtering about 300 names?
Hi Bekka, you can use regular expressions in filters, have a look at https://www.lunametrics.com/regex-book/Regular-Expressions-Google-Analytics.pdf or https://www.blueglass.com/blog/regular-expressions-dont-use-ga-without-them/.
In them, you learn about the pipe character |, which might turn useful in cramming that much more names in your fields?
I will be using this as a resource for my marketing team. What a great how-to!
Another AWESOME post! Thanks!
Fantastic post and very useful. Because the company I work for is very regional in nature I used your techniques and setup a segment using local city names and got very interesting results.
for step 3, couldn't you just duplicate (copy) the segment you made for step 2 and just change the condition to "does not contain" for each of the branded keywords you entered? this way if you have alot of branded terms, such as mispellings, product names or specific services, you don't have to duplicate or remember what terms you used for the branded segment.
guess i should have tested this concept out first :)
it does not appear to work in that each keyword/term needs to be within its own individual "and" statement for the custom segment when you are wanting to use the "does not contain" filter for each phrase...unless i am missing something i could not get the concept i initially explained above to work.
your original instructions work just fine, i was just seeing if there was a way to eliminate some duplicate work ;)
Hey Stephen, thanks for the step by step. Keeping track in Advanced segments sure beats the hell out of excel filtering! I have no idea why we haven't been doing it this way, thanks!!
Thanks for the article. It's always good to see how to extract data from Analytics.
GREAT STUFF. Filtering is one of the most under-utilized functions in GA. There is so much more fine data that is out there for us to consume and react to. Thanks for the thorough post Stephen!
Thanks for the post Stephen! I feel like everyone can improve in segmentation, so thanks for the tips!
+1, another great post stephen...
Great stuff. I am thankful there doesn't seem to be a limit to the number of filters that can be added to a segment, because it seems most of my clients have TONS of variations of their names, including mispellings.
FYI this kind of reporting is great to show the impact of TV and online display advertising- brand searches should increase when offline media is in action.
very well explained, great post :)
I use a similar non-brand search traffic segment with the following settings:
Keyword - Does not contain - brand name (or regex if it's a complex name)
Source - Matches regular expression - google|bing|yahoo
Medium - Matches exactly - organic
What do you think?
When you have a simple or unique brand name, then its quite easy to filter. It looks good from here, so if the checking checks out, then id say yes ;)
Thanks for such a useful tip with Analytics, Stephen! Appriciate it.
I've never even thought that the possibility existed to segment it like that. The good about this is that it's usable in so many nisches.
I just finished setting this up on one of my sites. WOW I cannot say thank you enough. I was missing so much information before this I really was wasting alot of time and effort before. Thanks again for this great post .