Finding Your Broad Match Key Phrases

INTRO: In the most recent Whiteboard Friday, a good friend of mine, Tom Critchlow (I can call you a good friend, can’t I Tom?), went over some great PPC basic tips. I noticed a lot of people in the comments wanted to hear more about PPC. Well, I’ve been doing the paid side of SEM for a few years (I remember Overture before Yahoo bought them) and thought I might take a stab at answering a few questions people had about PPC. I don’t know everything, so ask questions. If I can’t answer, I’ll research it … and maybe Tom can help out too. 🙂

SEOmoz member k0k1man asked, “Which report should I run to get the key phrases my broad matches are actually returning?”

Google AdWords has a report called “Search Query Performance” – and they are the only ones to even offer this! (You’ll notice that in the PPC world, Google AdWords is ahead of everyone in almost every aspect. That’s a whole other post … coming soon.)

You can drill down pretty far in this report:

• Level of Detail (Ad, Ad Group, Campaign, Account)
• Unit of Time (Summary, Day, Day of Week, Weekly, Monthly Quarterly, Yearly)
• Date Range (Set Standards or Custom)
• Campaigns/Ad Groups
• Advanced Settings (Outputs and Filters)

Now the irritation here is that many times Google will return a result of “x other unique queries” and that holds most of your long tail keywords. It’s really annoying when you just got all happy about this great new report.

Rule number one in any search marketing efforts: Have a good analytics program. You’re dead in the water without one. It doesn’t help with those long tail keywords giving impressions but not clicks. But that’s a small amount. I’m sure. and it’s not costing you $ right now, so stop worrying. You can compare your analytics report and Google Search Query Performance; the difference should be the words you’re missing.

Big note: If you are using Google Analytics, the below instructions will not work. You need a log-based file analyzer – Google Analytics doesn’t do what is mentioned below. 

How to Do It

Any good analytics reporter will allow you to get a CSV output of your referrer keywords. Understand this: none of this is perfect, sometimes things won’t register, and your list might be a little off Google’s numbers – this is normal.

1. Get a list of keywords that Google gave you traffic for a small time period (it will be natural and paid if you rank well). Start with a day like yesterday. We use Hitslink.com for our analytics, so I can pull that list of referrer keywords and take out just the ones from Google (Search Terms by Search Engine report).  Note: I’m not going into other engines like AOL here since that spans so many different sites, but you can get the general idea and scale it out for your needs.
2. Take your Search Query Report from Google for your chosen time period and put it in the same workbook, different sheet. Note: It’s important here that your campaigns are split out well. You can only pull all words from Google in Analytics, not ad groups, but if you have them sectioned right, it should be easy to spot which terms came from what ad group.
3. Compare the two using COUNTIF
   1. In a new column from your analytics sheet take the first row and punch in =COUNTIF(
   2. Use the column from the paid sheet with your keywords (ex. Sheet1 Column C)
   3. Set term to the keyword on that row (C2)
   4. Example: =COUNTIF(‘Sheet1’!C:C,C2)
4. Then Filter out the ones in that new column with a “1” as a result
5. What you are left with are words that you are showing for naturally and paid that Google AdWords isn’t listing. If you really want to know if it’s paid or natural, test the keyword in your own search.

What you are looking for here are not good words (you could, but if they GOT to your site, you’re already coming up for the word somehow), but the key phrases that come up and make NO sense. If they don’t make sense, set them as negative.

Now, those negative keywords are tricky. They come just like regular keywords broad, phrase, and exact. AND you can set them at the campaign or ad group level. Think real hard about what you are setting negative. For example, auto loans is not what my company does, but as a broad match keyword for auto loan refinancing, sometimes we get traffic for auto loans. But I can’t negative the broad term auto loan … that takes out my auto loan refinancing terms. Make sense? If you wanna know more, here’s the Google Lesson on it. 

This is NOT perfect, but it’s closer than what you had before. Always keep track of the keywords people are coming in on through your analytics program. It can help you understand things like who your visitors are and what they are thinking about. Analytics is not just for SEO. 🙂

Using for Conversions

In my company, we rely on number of applications filled out, that’s our conversion. As Tom said, you should always have a goal in mind; without that, PPC is worthless. Well, HitsLink also lets me see those words that are bringing in clicks and how many have converted to an application.

Goal conversion is more important than bringing in traffic. If you’re in a small niche industry like mine, there are many keywords LIKE what you do, but not what you do. Traffic coming in from related areas may not convert as well as that coming from key phrases that include exactly what you do. 

For this reason, I use an in-house tool to track conversions using barcodes (a coded tracker placed in the destination URL of the paid keyword) and the original referrer string. Each day/week/month (whatever) I can see a list of new applications, their barcode that points me to the campaign, ad group, and keyword that it came from, and the exact query string. Sometimes the whole query string doesn’t show, but about 80% of the time it does.

When I first got here I noticed a lot of conversions, but the quality was poor. I checked the phrases that people filling out apps were using. Because of that research we had to remove keywords that were bringing in tons of traffic and apps, but not ones we could use. So taking out that traffic reduced our number of conversions, but the end conversion to dollars has gone up. And that is what pays the bills.

Conclusion

Takeaways are (in case you breezed through all that):

• Always have a goal in mind and track conversions (and if you can, the original referring URL).
• Get and keep a good analytics program.
• Keep up with the keywords bringing traffic to your site, it’s your heartbeat.
• For commerce type sites, conversions are the key. You want good traffic, not just a lot of traffic.
• Get to know and love Excel (or something like it, like OpenOffice).

Thoughts for Google:

1. Fix Google Analytics please. There are some major features that SHOULD be there … maybe a paid upgrade??
2. Fix the Search Query report … come on, even on drilled down reports, you get one keyword and then another column with “20 other unique queries.”
3. Every other PPC engine … work on even getting us this report!!!

Kate Morris is the In-House SEM for RateGenius. You can find her on her blog and on Twitter.