How To Group Your Keywords; Plus: Q&A With WordStream’s Larry Kim
Organizing your paid search keywords into the right ad groups can be a difficult task for even the best search engine marketers. Many times, a keyword may seem to naturally fall into multiple ad groups. The engines tell us the best practice is to not have the same keyword in different groups, because then the platform determines when each should run so you lose that control. Control is key to great paid search management because if there are too many unknown variables, the data might be unclear and your optimizations may not yield the expected results. Thus, proper keyword grouping is an integral part to paid search success.
One of the main reasons why grouping upwards of hundreds of thousands of keywords for an account (sometimes millions) is so difficult is because there are many different schools of thought on which direction to guide segmentation. First, there’s the ad copy assigned to the group which can dictate grouping decisions.
We all know that the best copy for the job is usually the most relevant. So, if you find a keyword you’re questioning in a group, the most effective tip I usually recommend is to try to get into the mind of the user. Does a person searching on that term match the message? If they’re using terms such as “more info” or “research” and the ad speaks to “buy now”, then it’s off target for the user’s current place in the buying cycle and the keyword should be moved.
You also have to consider pure logistics when it comes to keyword grouping. Having too many campaigns and ad groups can really make the reporting, managing, and optimization of an account a weighty task. With your keyword landscape constantly changing, you have to be flexible when it comes to targeting, pacing, and bidding, but you can’t let this completely dictate your grouping strategy. Don’t shy away from a bigger structure just for the sake of ease.
On a recent client campaign, we made the hard decision to split their initial single national campaign into forty separate geo-focused campaigns. Yes, frankly, it was a pain in the butt — especially because the client didn’t care either way when we brought up the issue. But we knew that by targeting at the metro level gave us the ability to personalize the creative to be locally focused (including the address of the store locations) and would be crucial to the success of the account.
And don’t forget Quality Score. A bad apple can ruin the batch. We generally take out poor QS performers and move them to their own campaigns so that they don’t negatively affect the others in their group.
Once you’re gone through the process of segmenting by message, management, and Quality Score, then you still have the chore of segmenting by branded terms, top of funnel/bottom of funnel terms, where in the buying cycle they seem, competitor terms, product terms, etc. It’s a huge ordeal. Luckily, there’s help. WordStream (which I wrote about in a previous column) is a tool specifically designed to help search marketers segment their keywords. This week they released two free tools that utilize their grouping technology.
I spoke with WordStream’s founder and VP of Product, Larry Kim, the likable, humble brainiac behind the scenes about these new tools and to pick his brain on keyword grouping; something he is very passionate about.
Q. You’ve suggested that if PPC marketers have time for only one change to optimize their SEM, they should “turn their attention to keyword grouping and organization.” Why do you feel so strongly on this particular optimization?
Larry: Google says that click-through rate is the biggest component of Quality Score. I have found the best way to improve CTR is through a structured approach to keyword research. If you ensure that the keywords in your ad groups are closely relevant to each other – which in turn enables you to match them up with highly relevant ad text and landing pages, your CTR goes up because people tend to click on the ads that are most relevant to their searches.
With higher CTR and Quality Scores, you’ll realize better ad positioning, a reduction in actual cost per click, and even better conversion rates (by sending specific searches to specific landing pages, your potential customers are more likely to find what they’re looking for.) Furthermore, when you bucket similar keywords together into themed groups, routine tasks like bidding and reporting become easier and more effective. For example, if you decide to stop selling a particular product line, just turn off those ad groups rather than hunting for individual keywords.
My view is that keyword organization is a central point of leverage that affects everything in your PPC account. Do a good job here and everything else works much better.
Q. Keyword grouping can be hard. From a top level perspective, what’s your process for grouping?
Larry: Here’s a high-level overview of my strategy:
- Start with broad keyword research. Assemble and aggregate keyword data from as many sources as possible, including Web analytics and keyword tools. Starting with plenty of keyword data allows you to test various avenues and determine which pockets of keywords perform best—it’s difficult to predict which keywords will actually end up working the hardest.
- Form top-level keyword groups. These should be fairly broad and large. A good rule of thumb is to form top-level groups around your main offerings. So if you sell baked goods, top-level groups might include “cake,” “pie,” and “cookies.”
- Create smaller subgroups. Each top-level group will break down into more specific groups of keywords—usually with a modifier tacked onto the offering. Subgroups under the cake group might include “chocolate cake,” “birthday cake,” “cupcakes,” etc. Do this with all your top-level groups and you’ll start to see a taxonomy, or hierarchy, emerge.
- Keep going until your low-level groups are very small and tightly related. Pay particular attention to verbs and modifiers within a search query, which can reveal different layers of intent on the part of the searcher. The inclusion of transaction-oriented terms like “buy” and “compare prices” indicates someone late in the buying cycle. These clusters make great, highly targeted ad groups.
Q. Your company has just released two free tools to help with keyword grouping. Can you share some information on those?
Larry: Sure. We’ve released two patent-pending free keyword tools to showcase some of the power of keyword grouping and organization. The first is the Keyword Niche Finder. This tool is basically a keyword suggestion tool and a keyword grouping tool in one. You enter a keyword, and rather than getting back a list of single keywords, you get back structured keyword suggestions, grouped by relevance into potentially profitable keyword buckets. The idea is that once you have a keyword list, you need to start organizing it before you can do much with it, and the Keyword Niche Finder helps you get started with that organization. So for example, in the following screenshot, I did a search for “business cards” – the niche finder organizes the keywords into different niches like custom business cards, color business cards, etc.
The second tool is the Free Keyword Grouper. This one is similar but instead of starting with a keyword, you input a full list. So if you’ve already done some research using keyword tools, or if you have a keyword list from your analytics, you can just drop that in (up to 10,000 keywords) and it quickly groups them for you.
Using these tools instead of or in addition to standard keyword suggestion tools makes your keyword research much more actionable, by giving you tightly related groups that are ready to become PPC ad groups or themed content ideas for your website.
Q. So what if I have a keyword that could be in more than one ad group? That happens to me all of the time.
Larry: Different people have different approaches to duplicate keywords. My recommendation is to analyze the search query in question and try to infer the intent of the searcher. For example, consider the query “buy chocolate cake online”. Suppose you were selling food products online – you could possibly match this keyword in a group with other “cake” terms, or in a group with other “chocolate” terms. But which is more discriminating? What is the user actually looking for? In this case, I’d put it in the cake group. But you can always try putting the keyword in multiple groups and seeing where it performs better.
Q. How do match types enter into this conversation? Should you create separate groups for them?
Larry: Match types can kill all of your campaign organization efforts, because a non-specific keyword set on broad match can soak up millions of different search queries, even if you’ve gone through the effort of properly segmenting your keywords and creating targeted ads and landing pages. What we need is a keyword strategy for grouping and bidding so that AdWords will actually display the ads that we want to show.
I tend to group and subgroup keywords by intent, starting with non-specific categories and going towards more specific stuff, as described earlier.
If you’ve set up your keyword groups this way: bid slightly less on the less specific terms, and slightly more on the longer tail groupings. This way, you’re far less likely to have the more general keyword groupings “overshadow” the more relevant long-tail keywords. Also, consider putting the less-specific keyword groupings on more restrictive match types, like exact match, so they don’t soak up irrelevant impressions.
Q. Can you share any stories you’ve had where keyword grouping has been utilized for success?
Larry: Restructuring your ad groups based on your actual search query report data is like running your PPC campaign through a car wash. The before/after difference is staggering, particularly for accounts that have just a small number of ad groups with very general search terms (which is terrible.) We recently published a few case studies that show how much improvement is possible, including 1-800-Bakery and 1-800-Mattress. (Our strategy is to cornering the market on e-tailers with 1-800 in their company name… kidding!)
Q. Any last keyword grouping tips or tricks that you’d like to share?
Larry: Here are a few more tips:
- Prioritize! A keyword grouping strategy can inform your PPC and SEO workflow. Simply look at the keyword groups that are driving the most traffic and conversions and direct your optimization efforts there. Ad groups that are soaking up tons of impressions/clicks/conversions should be analyzed and broken up into smaller, more specific ad groups. You can also prioritize landing page creation based on those ads.
- Remember that keyword grouping is applicable for both PPC and SEO – it helps with topic selection, copywriting, on-page SEO, planning your information architecture, etc.
- Think “scalable”. This is one of the biggest mistakes I see people make (and a lot of the impetus behind our Keyword Management solutions.) Lots of people use Excel or AdWords Editor and build groups that are short sighted and aren’t designed to scale with new keyword ideas, new products, etc. Be sure to focus your tools and your strategy on long-term organization.
I’m sure this all sounds like a lot of work (and believe me, it is) but if you care about profits, these are the tasks that will actually improve your ROI and give you a competitive edge in search in the long term.
Sign up for weekly recaps of the ever-changing search marketing landscape.