Go to Top

Impact Of AdWords Ads on Overall Traffic: Google Study

Google published a new AdWords study on their research blog titled: “Impact of Organic Ranking on Ad Click Incrementality.” The first thing to notice is that the official title of the study is a bit puzzling. AdWords advertisers typically are interested in knowing if they are paying for traffic that they could be getting for free.

While this study is clearly meant to adress this, it could have been titled: “Are you paying for traffic that you could be getting for free?” Instead, Google opted for something about “Ad Click Incrementality” which basically means “Are my ads sending me incremental visitors that organic traffic wouldn’t otherwise provide if my AdWords ads were shut off.”

After reading through the full study, I now appreciate the official title incrementally.

Search Engine Land gave the study a brief summary. The broad-stroke, non-scientific conclusion is that it’s probably worth it to advertise in paid search for these reasons:

1. For sites in the study, most ads showed for terms that had zero organic rankings. And any click that happens on an ad that doesn’t have an organic listing accompanying it, would be sending incremental traffic.

2. Even when there were organic listings which ranked well on a page, paid search ads contributed significant incremental visitors that the website wouldn’t have otherwise gotten.

Another one of the main points of the study is that “even when advertisers show up in the number one organic search result position, 50% of clicks they get on ads are not replaced by clicks on organic search results when the ads don’t appear.” A very interesting conclusion, since (presumably) a large share of number one organic rankings would be covered by advertisers bidding on their own trademark.

This is a glancing reference to another age-old question: “should I bid on my own trademark?” Going strictly by traffic, the Google study indicates “yes.”

From what I’ve seen, using Google Analytics and sales/lead data, none of these conclusions are surprising to me. They’re consistent with my experiences. Although I still come up short on the reasons why, when bidding on your own trademark, it brings in more traffic, even when you are ranked number one organically for trademark queries.

Using Google Analytics, I have measured trademark traffic both before and after showing ads on trademarks, and it does seem to be the case that ads increase overall trademark traffic.

One hypothesis I have to explain this is that people get used to seeing Google’s yellow box full of ads at the top of search results. Maybe someone who is browsing around begins focusing more on top ad spots and stops paying attention to organic listings, as if the yellow box was validation of relevancy for their query.

Another hypothesis I’ve had is that visitors are interested in seeing every listing for a site – maybe because they aren’t easily finding the information they are originally looking for from the first organic listing, they choose to click the ad. It could be variable across sites, but I doubt Google would help find a generalized answer to what’s going on here.

Of course, Google didn’t measure any trademark queries directly in their study, so it’s all still an assumption. And if you look at the full study, Google also didn’t measure directly the incremental clicks gained specifically on ads that rank for terms that have a number one organic listing.

Instead of direct measurement, they normalized data that includes incremental clicks from all ads, then took a limit to exclude incremental clicks for ads that appear on pages that don’t have any organic listings. From an analysis perspective, I find this confusing because it seems like it would put focus on only a portion of their set and also still include values for ads showing on pages with organic listings lower on the page.

The study also doesn’t give numbers for the mean percentage of clicks from ads that appear next to organic listings. Guessing these numbers from box plots, however, everything seems to add up.

Rather than go by any of these studies, it’s better to do your own tests. You can just try a change and compare data across time periods, but Google also outlines a methodology for creating randomized tests by varying geo-targets.

The idea here is to have a test group and a control group, split up by random geographies for making comparisons, and doing what’s called a “randomized, pretest-posttest, control-group design.”

Googling that type of experiment and looking through Google’s geo experiment methodology, you should see statistical methods for analyzing data from that kind of experiment.

But if terms like “linear model” and “analysis of covariance (ANCOVA)” make you uncomfortable, it’s a rather tough read. Analyzing difference in data without statistical models is possible, if you acquire a lot of data and can see large differences, but it’s a good idea to have the math on your side. After all, you might not know what is “a lot” and what is “large.”

facebooktwittergoogle_plusredditpinterestlinkedinmailby feather

About Steve Loszewski

Steve Loszewski leads the paid search team at Pure Visibility. He is individually qualified in AdWords, has the Google Analytics Individual Qualification, is an Oracle Database 10g Administrator Certified Associate, and is a Sun Certified Programmer for the Java Platform SE 6. Steve has been managing AdWords accounts since 2005 and also has experience in SEO. Most of his time is spent in the trenches, working with keywords, ads, bids, landing pages, placements, etc within the AdWords Interface. You can find him on Google+.

Leave a Reply

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

You may use these HTML tags and attributes: <a href="" title=""> <abbr title=""> <acronym title=""> <b> <blockquote cite=""> <cite> <code> <del datetime=""> <em> <i> <q cite=""> <strike> <strong>