I don’t know about you, but I’ve recently been hearing more and more buzz about the value of query mining to optimize your search campaigns. Take the following two reports, for example. Over 500 different, unique queries triggered a visit from an ad for “dog remedy” on broad match. Now, look at all of the queries – “itchy”, “smelly”, “worms” come out in more specific queries. Those are all different problems and should be in their own ad group. Then, there’s actual diseases and conditions such as “scabies” and “cysts”,  each of which would certainly will require different landing pages. Also, notice different dog breeds are mentioned. Wouldn’t dog owners most certainly be more engaged with an ad that speaks to their breed directly?

If these are core keywords to your business, knowing those other terms may help you build our your long tail list for higher clicks on cheaper (and more engaged) terms. You’ll know if you need to build out more sections of your site to accommodate a variety of content to explain the nuances of that product or service. You may discover new ad group and campaign ideas and new ways to build ads that speak directly to a specific consumer need.

I can’t wait to get started query mining right now…

However, with so many other optimizations that are tried and true “must-do” best practices, query mining might not be on the search engine marketer’s front burner. Certainly portfolio bid management, landing page testing, creative message testing, quality score optimization, match-type testing, adding negative terms, position strategies, conversion optimization, and so forth – cannot be ignored, right?

Well, Craig Danuloff, the President & Founder of ClickEquations is someone who is certainly is not ignoring the value of this optimization. I’m not going to call this man the King of Query Mining–but he just might be the Duke. Craig has written numerous query posts and recently hosted a webinar on the subject.  But I think what really proves his passion for this topic is that he made sure his search management platform has a query analyzer built in.

I asked Craig for some good insight into query mining and how to take it to the next level.

Q. Where can an SEM pro go to find search query data?

Craig: In AdWords, search queries reporting has been added in the ‘See Search Terms’ button under the Keyword tab. This shows you the queries for a select keyword or all the queries for an ad group or campaign. But if you’re not using it on a keyword-by-keyword basis, it doesn’t show you which queries were matched to which keywords, which is a little limiting. Neither Yahoo nor Microsoft provide any access to search queries within their management interfaces.

In ClickEquations, we included a ‘Search Query’ tab where each search query is listed, along with the keyword and match type it was matched to, and full statistics on the query performance. Queries are captured and reported for Google, Yahoo, and MSN and can be viewed at the engine, campaign, ad group, or keyword level. We also allow you to pull search query data into excel using our ClickEquations Analyst tool, to build a bulk sheet for keyword or negative expansion, or for all kinds of analysis or reporting.

Q. Why did you invest so much time into building query reports into ClickEquations?

Craig: The traditional view of paid search has been that it’s about keywords and bids. And a lot of PPC management time and attention gets spent on keywords – expanding them, bidding on them, organizing them, etcetera. But the truth is that keywords are just a means to an end; they’re little magnets sent out there to attract search queries. And if you’re only able to review reports and make decisions at a keyword level, you’re not getting a very accurate or informative picture of what’s really happening in your account – so you’re almost certainly making bad decisions and not optimizing your results.

We call the ‘keywords and bids’ view ‘Low-Resolution PPC’, and we preach “High-Resolution PPC”. Take one simple example – suppose you have a keyword in your account that gets 100 clicks, results in 20 conversions, costs you $100 and generates $800 in sales. Sounds pretty good right? On that information alone, you’d probably increase the bid if your position wasn’t already maxed, and be pretty satisfied.

But what if you could see the 100 search queries behind those clicks and their individual performance? Suppose you saw that 88 of those clicks, representing $55 of your spend, produced no sales at all, and in fact, included words that suggested those people were really non-qualified prospects. And just 2 specific search queries were alone responsible for 10 of the conversions but only $5 of your spend. At this level of detail, you see that you really have a small number of super-profitable queries (which should be converted to keywords) and a large number of queries which should be either turned into negatives or at least bid in a different way.

The truth is a version of this scenario is happening right now in every ad group you’re running – but you can only see it and take the right action by looking at the search queries and their performance. We included full search query reporting, and related features to take advantage of these queries, because we believe that using them is the most effective way to improve paid search results.

Q. Why do you hate Broad Match so much?

Craig: Broad Match is a tool that is frequently misused. Adding a keyword on Broad Match is a way of telling the search engine “here, you figure out how to spend my money.” That’s fine in certain narrow cases and usually for limited amounts of time, but beyond that it’s a recipe for waste and missed opportunities. We have a great Match Type Analysis Report in ClickEquations where you can see the % of your spend, revenue, conversions, etc. that are resulting from the keywords of each match type. As a simple rule of thumb, I’d say that no account should have more than 50% of their spend occurring on Broad Match keywords. Ideally, I’d say 30% is more of a goal.

The right way to use Broad Match, and all the match types for that matter, is to build what I call a ‘Match Type Keyword Trap’ which has been described in detail over a series of blog posts and a white paper at our site. Basically, over time search queries that produce good results should be promoted to Exact or Phrase Match, so that less and less of your sales result from Broad Match keywords. The Broad Match that ultimately remains should be either a way to catch new or strange queries you haven’t seen or considered before, or for those circumstances where a huge range of queries are infrequently used to express the same search intent. Using Match Types in this way is another example of working in ‘High Resolution’ and results in increased impression share, better Quality Scores, higher positions and lower CPCs – everything a paid search marketer could dream of.

Q. I’m supportive of your suggestion of an Include match type. Can you sum up your thoughts on this for the readers?

Craig: Right now Phrase Match is literal – the words in your keyword have to appear in the query in the exact order. There are many cases where users might search for a set of words in many different orders and with many different delimiters: ‘cheap hotel in Spain’, ‘hotel with view in Spain’, ‘travel package to Spain with hotel’, for example. It would be nice to buy the keyword ‘Spain hotel’ with an Include Match Type and have it cover any query that included all the specified words in any order regardless of any other words in the query. It would allow us to apply one Include Match where today you would need 10 or more Phrase Matches or worse yet just have to use Broad Match because there are so many combinations.

Q. What is the best way to organize search queries to gather insights for optimization?

Craig: Search queries are naturally organized into the ad groups and campaigns where they were matched, and they’re useful at each of those levels as well as on a per-keyword basis. By looking at the collective list of search queries in an ad group, you can judge if your text ads are appropriate, and see when you really have to split keywords into smaller ad groups in order to better tailor the text ads to the search queries.

Looking at queries for an entire campaign helps you decide if you can trust the summary reporting you’re getting at the campaign level. If a few queries are generating all the cost or revenue, then it becomes obvious that the averaged data isn’t really that informative. On a keyword basis, the queries obviously let you add more negatives or expand to new keywords or match types, but they can also help make better bid decisions by clarifying whether the keyword performance was skewed by a few poor or very good queries.

Q. What are some best practices to utilize once you’ve analyzed the query data?

Craig: My favorite new use is to look at the number unique queries are being matched to individual Broad Match keywords. I’ve been shocked to find single keywords getting matched to 500 different search queries or more! This is a huge opportunity to dive into the query list and add negatives and expand the keywords and match types to better take control of that targeting. We have a report in ClickEquations Analyst that sorts keywords by number of unique queries, and it’s a real eye opener. I used it just last week to first indentify this problem and then bulk import about 40 negatives (in this case it was a medical keyword that was attracting a huge list of specific medical conditions that the product couldn’t treat) and almost 80 new keywords and phrases.

Opinions expressed in the article are those of the guest author and not necessarily Search Engine Land.

Related Topics: Channel: Search Marketing | Search Marketing Toolbox

Sponsored


About The Author: has been a search marketer since 2003 with a focus on SEM technology. As a media technologist fluent in the use of leading industry systems, Josh stays abreast of cutting edge digital marketing and measurement tools to maximize the effect of digital media on business goals. He has a deep passion to monitor the constantly evolving intersection between marketing and technology. You can follow him on Twitter at @mediatechguy.

Connect with the author via: Email | Twitter | Google+ | LinkedIn



SearchCap:

Get all the top search stories emailed daily!  

Share

Other ways to share:

Read before commenting! We welcome constructive comments and allow any that meet our common sense criteria. This means being respectful and polite to others. It means providing helpful information that contributes to a story or discussion. It means leaving links only that substantially add further to a discussion. Comments using foul language, being disrespectful to others or otherwise violating what we believe are common sense standards of discussion will be deleted. Comments may also be removed if they are posted from anonymous accounts. You can read more about our comments policy here.
  • http://ppcisme.com Erez PPCisme

    Craig, I agree with you that broad match is the worse and there is nothing better for us than “included match” if it existed. But there are few things I would like your opinion on because I somehow disagree:
    1) You said that you want only 30% of your account budget to be spent on broad match type. Do you think this is reasonable? I mean, there are so many weird and unique queries that you couldn’t even imagine and I believe it’s more than 30%.
    2) I agree that you should add your best converting queries in all match types but this make sense only if large % of your sales are coming from specific keywords, but many times most of your sales will come from the long tails, and you just can’t have all of the long tails as keywords. Having too many keywords with low search volume is worse I think than having broad match keywords.

    Basically what I’m trying to say here that adding phrase and exact match types is recommended as long as there is enough clicks volume to justify it – that’s my opinion anyway.
    BTW, I’ve also encountered situations in which a certain search query triggered on of my broad match keywords although I have the exact query as an exact match type – this makes the whole match type “thing” a little more tricky, but that’s a different story.

  • http://www.clickequations.com Craig Danuloff

    Erez – Thanks for the questions. The 50% and 30% goals are rules-of-thumb and in this more than most areas the account specifics/situations really matter. The density of queries in your head will determine how effectively you can capture things in Exact or Phrase – remember it isn’t that having lots of Broad is inherently bad because of the match type, only in that it usually means we haven’t scrubbed our queries – adding negatives and promoting other keywords to more specific match types. If Broad is making money, Broad is fine!

    In terms of exact queries matching against the Broad versions of the keywords, yes sometimes those do slip through. Shouldn’t be more than a few % of the clicks on those keywords though – AdWords help is clear that more exact versions should match higher within your account.

  • Stupidscript

    Thank you for these insights. I definitely agree that an “Include” match type is desirable. If I remember correctly, “Include” is how “Broad” used to work … before the days of keyword “expansions”, which, if I remember correctly, used to be opt-in, but are now simply the way “Broad” is implemented.

    I am having an impossible time correlating user queries with the term/phrase that was matched. Where did you get your chart? I can’t find anything like that in the AdWords interface, and I’ve been looking really hard.

  • Stupidscript

    Ahhh … I see now that the chart with Google queries and the keywords that matched them are part of ClickEquations’ toolset. So that’s answered.

    Regarding the “include” match type … isn’t that how Google’s original Broad Match was set up? Prior to “extensions”? I remember being asked to approve inclusion in Yahoo’s expansion option (if we didn’t opt-in, then matches continued to be made based on the same idea as you have for Include … in the query in any order = match), but I do not recall being asked to approve the same change at Google, apparently around October 2006 when expansions went into effect.

 

Get Our News, Everywhere!

Daily Email:

Follow Search Engine Land on Twitter @sengineland Like Search Engine Land on Facebook Follow Search Engine Land on Google+ Get the Search Engine Land Feed Connect with Search Engine Land on LinkedIn Check out our Tumblr! See us on Pinterest

 
 

Click to watch SMX conference video

Join us at one of our SMX or MarTech events:

United States

Europe

Australia & China

Learn more about: SMX | MarTech


Free Daily Search News Recap!

SearchCap is a once-per-day newsletter update - sign up below and get the news delivered to you!

 


 

Search Engine Land Periodic Table of SEO Success Factors

Get Your Copy
Read The Full SEO Guide