Experiments with Exact Match Keywords in PPC

Last week our staff attended MozCon, virtually of course. There were a lot of good presentations and content. However, one small part of one of the sessions really got me thinking.

Francine Rodriguez of WordStream gave a very good presentation on automation in PPC. Her presentation, along with all the Moz video bundle, is worth the price of admission.

During part of her presentation, she covered how keyword match types have changed in the last 1-2 years. And her argument was that match types are essentially dead. Part of her research found that in one of her exact match keyword campaigns, 80% of the clicks and 41% of the spend turned up duplicates for exact match keywords. This meant that 20% of the clicks and 59% of the spend were on the targeted exact match…so what about the other variants? Her presentation talked about 15% of the 3.5 billion searches a day were for queries that Google has never seen before and how now it can be important to build in function words like “in” and “the”.

Her analysis of these exact match keywords turned my attention to the instances where I am using exact match keywords. With all of the changes in the keyword match types, I was very aware of the wide variety of actual queries that may trigger ads for keywords, EVEN if I am using phrase matching. I use phrase matching most often in my keyword match type selection. As a part of this approach, along with long-tail broad match and a fair amount of modified broad match types, I always diligently study the actual search query reports. These reports give me good insights on what the market is actually typing in and gives me ideas on keyword variations, ad copy, extension language and even landing page copy.

However, admittedly, I think I found a gap in my work. Exact match keywords. While I don’t use them as often as the other match types across the accounts we manage, I do use them frequently. I knew, based on the grumblings on Twitter and industry articles, that you can see some misguided actual search queries for exact match keywords. But, I guess I was too naive. I didn’t spend much time looking at search query reports for exact match because, well, I thought exact match meant exact match with some small variance for intent and spellings. So, post-conference, I dug into our accounts that use exact match. Wow, was I wrong:


The first example of exact match that I encountered was my exact match campaign with pool slides. Since the campaign launched, there were 4,280 clicks for that exact match variation. While I had 3,031 clicks for pool slide or pool slides, there were 30% of actual search queries were not for these two variations. An additional 1,249 clicks being for queries that were not the exact match variations I had set up to target. This is obviously quite a large number. Another keynote is that 30% of our click spend also went to other queries that were spin-offs of our exact match variations. Now, I did not see as many outlandish examples of poor queries triggering our ads. However, this did alter my strategy for variations and ad copy, and I did need to add some negatives from this research.

Laser Distance Meters

With this information, I went onto two other situations of branded exact match. One of the examples is an authorized dealer for a manufacturer of laser distance meters. The other is an informational campaign in regards to car insurance law. These are exact match brand terms. Evaluating this, I did see that in the reported queries for the laser distance meter did deliver what I was looking for. Since January 1st, this term received 417 clicks and 6,543 impressions out of 464 clicks and 6,712 impressions. For these percentages were much more aligned with what I want from exact match at a variant rate of 90% of clicks and 97% impression rate. Most of these variant queries that trigger ads were for spelling and translation queries.

PSA Campaign

Transitioning to the brand name for the informational PSA campaign, things do tilt to more of a 50/50 situation. Although our volume was much lower, I was seeing a 54% variant match rate with also a 47% match on impressions. Now, with this evaluation, the other term that was getting the other clicks and impressions were for those that were appending the domain extension (a .org) to the brand name. But, if that is a search scenario I was trying to avoid, then this would have been another good example of exact match variants not delivering the approach we wanted.

So the key takeaway here is to be aware and know that Google is becoming more and more loose on the interpretation of the intent of those queries and what queries will trigger ads, no matter what your match type is. While this minimizes the need for dumping tons of keywords into an ad group, you need to keep studying your actual search query reports. Additionally, both Francine and an additional presentation by Dr. Pete Myers at MozCon referenced staying ahead of the curve in terms of keyword orientation, intent and new queries. Keep using Google Trends and your search query reports to stay ahead on your PPC keyword selection strategy. Good luck!

Joseph Ford

Joe Ford is a Managing Partner at Netvantage SEO. In addition to overseeing day to day business operations of Netvantage, he directs paid search strategy and management. Ford is on the Marketing Committee for Impression 5 Science Museum, and the Executive Board of the Capital Area IT Council. He was previously a member of the Lansing Regional Chamber of Commerce Executive Board of Directors for 8 years. Additionally, Ford is an adjunct faculty member in the Eli Broad College of Business at Michigan State University.

Leave a Reply

Your email address will not be published. Required fields are marked *