Has Google’s Broad Match ‘Over-Expanded’ Itself?

by Tim Daly, Friday, April 20, 2007
MediaPost Publications

AS PAID SEARCH engine marketing matures, so do the strategies marketers employ to manage their campaigns. When Google introduced the expanded broad match a while back to the regular broad match that is so frequently used in campaigns, there was immediate interest, but there were also some issues lurking. As defined by the GoogleAdWordsHelpCenter, “With expanded matching, the Google AdWords system automatically runs your ads on highly relevant keywords, including synonyms, related phrases, and plurals, even if they aren’t in your keyword lists.” Now you will not have to build out extensive keyword lists and take into consideration the long tail. Sounds good, right? But we discovered a few interesting eye-openers when we investigated this further.

The expanded broad matching has been constantly changing over time, and recent expansion efforts by Google seem to have numerous matching inaccuracies. For instance, Google initially recognized the search terms cookware and bakeware as one and the same. The problem with this is that these are two completely different products. If consumers were to place their plastic cookware in the oven and use it to bake something, they would have quite a mess on their hands.

Curious about the effect that this natural extension of the broad match is having on advertisers, we conducted a test to find out. The foundations of the test: Change all keywords in a relatively small campaign to exact match and see what happens to performance data. After just one week, we evaluated impressions, clicks, click-through rates and cost-per-click — and what we found was astonishing. Impressions decreased by 24.38%, yet clicks only decreased by 15.68%, thus resulting in an 11.36% increase in the click-through rate. Since the click-through rate plays an integral role in the Quality Score algorithm, the average cost-per-click decreased 23.77%. Most interesting, the performance changes were instantaneous, suggesting that expanded broad match negatively impacted Quality Score and Ad Rank.

So, what does this all mean?

In summary, these results suggest that Google’s expanded broad match is potentially serving less than optimal ads, lowering click rates, increasing cost-per-clicks and negatively impacting your Quality Score. Because some marketers are blind to Google’s ad serving, bid changes are being made to the wrong “broad matched” keywords, causing changes to Google’s Ad Rank. As a result, the wrong ad gets served and advertisers may then unknowingly pay more than they should for the ad placement. In worst case scenarios, Google may be sending traffic to the wrong page on your Web site given the ad it chooses to service — which in turn could negatively impact your conversion performance.

To address this potential problem, advertisers should seriously consider focusing on building out the “long tail” of key words and limiting the use of broad matching, now more than ever. If a keyword or phrase is relevant to your offering, then you should be taking the time to include it in your PPC search campaign.

How can you tell if your campaign is affected?

There are most likely fewer than 500 keywords that drive 80% or more of your media spend, so isolate a few of those and change the match type to a phrase match. As long as the phrases are in a logical order, like buy red shoes as opposed to red shoes buy, you should be able to assess the impact. Look at all of the performance metrics through to your conversion numbers and return on advertising spend (ROAS) for even date ranges.