Google Study: PPC Ads Do NOT Cannibalize Your Organic Traffic

Does Paid Search Really Cannibalize Your Organic Traffic?

Though I’ve never met him personally, I admit to being a big Hal Varian fan. For those who don’t recognize the name, Dr. Varian is the Chief Economist at Google and like me, one of the oldest guys in his company.

Over the past few years, he and his team of researchers have made my life simpler by providing pithy answers to some of search’s mythically difficult questions, like “How do search auctions work?” and “Does ad position effect conversion rates?

Last week, his team at Google released the results of their research that answers a question that paid search managers across the world get asked on a regular basis: “Why the [bleep] are we advertising on our own [bleeping] brand terms when we are ranked #1 for those [bleepety-bleep-bleeping] terms already?  [Bleep]!”

Though the Google research team posed the question in a slightly more scientifically-fashionable way for their research, they essentially set out to answer the question of whether or not paid search ads cannibalize traffic from corresponding organic listings for the same keywords.

Their findings, in true Varian-esque style, were simple, direct and memorable. They found that paid search ads give you a 89% incremental lift in site visitors – above and beyond traffic you would normally expect from your organic listings.

You can download the study at the Google Research Blog:  “Incremental Clicks Impact Of Search Advertising” by David X. Chan, Yuan Yuan, Jim Koehler and Deepak Kumar.

Is A Google Study That Proves Google Paid Search Works Valid?

You don’t have to be a committed cynic or skeptic to question the results of Google’s research on its own search properties. It’s only natural to raise the question of self-interest, but since Google has made no bones about the fact that it is their own research, they are being upfront and candid.

In their report, they provide pretty good detail on their methodology and their statistical methods though it is provided only in summary form.

But, in fact, the study seems to support the prevailing conventional wisdom in our industry and even some earlier studies on the impact of paid advertising on brand terms and natural traffic. It certainly supports Brad Geddes findings in his Search Engine Land column last week, “Should You Bid On A Keyword If You Rank Organically For That Term?

At most search conferences, and in columns, we often are reminded by experienced search experts such as  Sid ShahGeorge MichieMona Elesseily,and Josh Dreller, to name a few, of the importance of managing and bidding on brand terms even when you have good organic positions.

Even when the advice is more anecdotal than data-driven, such as keeping your competitors from dominating the paid ad spots, brand building and controlling messages, most of us buy into the idea that it is a good idea to bid on brand and other high ranking terms.

A related independent research study published in 2008, by a NYU undergrad business student, Priti Kumar also seems to support Google’s conclusions, too.

The study, Search Advertising in Electronic Markets: A Study on the Impact of Keyword Wordographics (PDF) was more generally about the effects of wordographics on search advertising campaign success but in the process Kumar’s research mostly disproved the idea that paid search ads cannibalized brand keyword traffic, at least in the case of the large national retailer involved in the study.

Google’s own study was rigorous. They extracted data from 446 campaigns running in Germany, France, Great Britain and the United States from October, 2010 through March, 2011. They examined campaigns that had been paused after running ads for some period of time and then, using some clever deductive modeling, estimated the incremental impact of paid search ads on total search volume.

If the study has any Achilles heel, it might have been the fact that over half of the campaigns modeled were US campaigns, and were heavily dependant on data in the holiday period. There is some hint that the research is ongoing, and if so, it will be good to see if the same 89% lift holds true across other parts of the year.

Trust, But Verify Using The Nuclear Option

If you are really serious about testing the impact of paid search on your own campaigns, the quickest and most effective test is simply to simply to turn off your brand campaigns for a short period and then evaluate the impact on your traffic and your conversions, which I call the “Nuclear Option.”

I refer to this as the Nuclear Option because it often chokes off so much traffic so quickly that you notice the impact on top-line revenues immediately. We don’t deploy it often because it can have such a detrimental impact on revenues. More often, just threatening to turn off the brand campaigns is enough to dissuade others in your organization who are whining about paying for clicks they believe they should be getting free. They usually back down before the test goes online.

Or, you can take a studied approach along the lines Brad Geddes outlined last week. Both Brad and Google provide formulas for evaluating the economics of your own incremental paid search ad cannibalization studies. The math involved in doing your own testing is straightforward and is actually superior to the Google study, because instead of looking purely at click volume, making it a better truer test of your particular campaign and market space.

Google stops short of recommending taking the campaigns offline, though, and suggests modeling instead of the Nuclear Option. Most of us don’t have in-house economists, however, but if Hal Varian wants to volunteer any of his scientists to to devise models for any of our accounts, we will welcome them with open arms.

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

Related Topics: Channel: SEM | Paid Search Column


About The Author: is President and founder of Find Me Faster a search engine marketing firm based in Nashua, NH. He is a member of SEMNE (Search Engine Marketing New England), and SEMPO, the Search Engine Marketing Professionals Organization as a member and contributing courseware developer for the SEMPO Institute. Matt writes occasionally on internet, search engines and technology topics for IMedia, The NH Business Review and other publications.

Connect with the author via: Email | Twitter | LinkedIn


Get all the top search stories emailed daily!  


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.
  • geoprofi

    Actually, I’ve just read the research paper and I do not think it supports the statements made in this article.

    The research shows that for most advertisers (89%) removing PPC SE traffic will result in loss of SE traffic that will not be compensated by their organic listings. However, as we very well know, only a few advertisers also occupy the top 3 position in the organic listings (which are comparable to PPC traffic). Hence, this study proves nothing, than any web marketer can’t tell you and prove you in a matter of minutes.

    If the research was made so that it included only advertisers who had top organic rankings, the results would have been very different, as I have seen more than once in my practice.

  • Wouter Blom

    owning a website with 50-70 000 visitors a month and a monthly budget of 3000 euro’s i also see no evidence that adwords in cannibalizing my organic volume

    We rank organicaly mostly between position 3 and 7 on 450+ keywords.

    We ran multiple tests because we could always spend the money missed during one day in a upcoming day. So we tested with daily intervals of adwords – no-adwords. and weeks of adwords and than back to no adwords.

    Currently not running adwords for reasons of profitability :-(

  • Aaron Bradley

    I agree with @geoprofi. I was unable to find in the study any indication of what organic rankings the advertisers how paused their campaigns possessed (no whether these were branded or non-branded keywords). If you’re not ranking organically for a given keyword *of course* your traffic is going to decline when you turn off a paid search campaign for that keyword. This “study,” indeed, proves nothing.

  • Matt Van Wagner

    Hi Geoprofi.

    Thank you commenting.

    Your suggestion to get research results specifically for the case when top ads and organic spots appear together – is excellent, especially since the Google study did mention that low incremental ad clicks (IAC) “…may occur when the paid and organic results are both similar and in close proximity to each other on the search results…”

    Priti Kumar’s earlier research suggested that cannibalization was not an issue when branded-terms appeared in high organic and paid ads spots, but it was limited in scope and not conclusive, so if Google were to include that type of research next time around, it would be really helpful to brand managers.

    On your first point, however, I believe you may have read too quickly because the 89% Google refers to is incremental additional clicks attributable to paid search, not to a majority of advertisers. No matter how well they rank currently, if advertisers can get incremental lift from paid search ads, and it is profitable, clearly it’s a win, so the study is useful, even if it may be perceived as self-serving.

  • Matt Van Wagner

    Hi Aaron – Thanks for chiming in, too.

    I agree that additional research along these lines would be great to have – maybe Google is already working on it…

  • Mary Kay Lofurno

    I believe the answer here is test, pure and simple.

    I have worked on sites where I have tested this and seen both outcomes, depending upon the product/industry/objective.

  • lemazel

    I would discount Google’s study immediately. I wouldn’t even consider giving it any credibility whatsoever. Do you trust a cigarette company’s research on the effects of smoking? The oil industry’s research on global warming? Halliburton’s research on the positive economic impact of war? Anybody can do a study and get any kind of results they’re looking for – and even make the study look reasonable and unbiased to a reader. The easiest way to do that is to carefully select your subjects. And it’s really hard to demonstrate that you weren’t biased in the selection process. Sounds like Google mainly looked at e-commerce websites in this instance. The fact that the subjects were mainly in the US is a red herring.

    It’s not to say that Google’s research necessarily gives an incorrect of misleading conclusion – just that it should be completely discounted. And frankly, the same goes to a certain paid search marketer who wrote the post. Making a commission on paid search ad spend are we?

    The big giveaway here is that you are making blanket statements about a reader’s business. You shouldn’t generalize: the type of business makes all the difference in the world. E-commerce is hugely different from, say, webmail. Even if you’re a zillion-time Amazon user, when you go to the site, chances are you’re there to buy something. If eBay hijacks your visit in a bit of competitor brand bidding, you might just buy the same thing there instead. Might even be cheaper. But if you need to login to your Hotmail account to check your email and Yahoo pops up on paid search, might you decide to login to your non-existent Yahoo account instead? Nope. Might you decide all of the sudden to create a Yahoo account? Nope. Might Microsoft spend a fortune buying clicks on it’s own brand that it would get from organic anyway? Yep. Does Hotmail sponsor its own brand? Nope. Does Amazon? You betcha.

    The most annoying thing of all is that while you admit that the only real way to test the impact is to… wait for it… *test it*. You then go on to suggest that nobody actually has the goujons to carry out the test – so devastating would the results be for the business. Most people just ‘model’ the impact. What the bleepety bleep?!?!

    Maybe I should just forget all of that tedious a/b testing and just model the results from now on. Nice.

  • Matt Van Wagner

    Agreed, Mary Kay. Thank you for commenting.

  • Matt Van Wagner

    Hi Lemazel,

    I appreciate strong opinions, thank you for voicing yours.

    The study was based on a question that itself a broad generalization. The answer, of course, will be a generalization. That shouldn’t be a surprise to anyone.

    If you want to really answer the question for your own company, you either have to build a model, or you have to set up a valid experiment and work with real-world data. Do you have an alternative suggestion for answering the question specifically or more generally?

    As I said at the outset of the post, what I like about the Google study is that, generally-speaking, it helps provide insight the relationship between paid search advertising and organic traffic. It is not the first study attempting to provide the insight. It won’t be the last.

    As to your cutting remark about making commission on ad spend, we don’t make commissions on paid search and I am not aware that anyone in this market does. Are you aware that search engines do not pay commissions? However, I don’t mean to disappoint you by not feeling appropriately stung by your sharp rebuke, so If it not too much trouble, can you please tell me what bleepety-bleep you are talking about and I promise I will make time to be hurt by the remark. Much obliged. Cheers.

  • George Michie

    Matt, we’re conducting a study on this as we speak. We suspect that the answer will vary widely depending on how much money the given brand advertises offline, how long they’ve been in business, the degree to which their name is unique (Zales vs and as others have mentioned, whether or not you rank #1 for your own brand.

    We’re also going to look closely at the extent to which pulling the ads increases affiliate traffic, as they too can cannibalize brand search.

    As a former retailer, it just doesn’t make sense to me that most brand advertising isn’t cannibalized in most circumstances. Different if your brand name is “online poker dot com” but for Lands End, I have a hard time believing they’ll lose significant traffic overall by pulling the ad.

    I think testing risks little, here, but we sometimes struggle to convince clients to try it.

    Non-brand, competitive search is a totally different ball game.

  • Matt Van Wagner

    Thanks, George. Good luck with your study. I hope you’ll be able to write up on your RKG blog or here at SEL.

    We have several brand clients in market segments where there is no clear dominant player, and/or whose names are bland and easily confused. We’ve done some purposeful testing, and in other cases expiring credit cards have caused unscheduled ‘tests.’ We have definitely observed the direct impact of brand ads being offline. I suppose famous brands like Lands End who have strong offline marketing efforts that pre-dated the web would be less impacted by a nuclear option test.

  • mvaneeghen

    I consider the following to be true: Campaign * Website * Product = Result. Therefore the results of such a study are true for the websites that have been included in the research.

    Moreover, the same study on the same websites would produce a different result a year later as the quality of their SEO and PPC strategy may vary, their website is different or their product has become less desirable.

    In short: This study is BS and Google is full of it.

  • George Michie

    I don’t think this study is worthless, nor do I think they cooked the results. They haven’t detailed answers to the critical questions. They have not differentiated between brand and non-brand search which is crucial, and they failed to thoroughly disclose results relating to the organic link position. To their credit, in section 3 they do acknowledge that: “A low value for IAC [Incremental traffic from ads] may occur when the paid ad and organic results are both similar and in close proximity to each other on the search engine result page.” They go on to explain that this happens when both are near the top of the page.

    Disentangling brand from non-brand in those critical cases would provide very valuable insights.

    One more point in Google’s defense. Their published research does not always support their short term financial interests. Varian’s paper showing the relative invariance of traffic value by position — which we and others have been saying all along — does not really help their cause. Cooking the books to show that the top position converts better than others would create much greater incentives to jockey for position at the top of the page. To their credit, they did not make that claim. They could have gone a step further (as we did) and pointed out that conversion rates at the top were actually slightly worse than at the bottom, but… :-)

  • Matt Van Wagner

    I concur completely George.

    I sure hope Google continues collecting and mining the data and publishing more ad studies like this – the insights are interesting directionally. Even though they are macro-level observations, and not intended to be correct for any one advertiser’s environment, they are enlightening.

  • Matt Van Wagner

    Thank you for your feedback, Wouter. Hope you can find a profitable sweet spot for your AdWords campaigns. Have you tried adCenter for Bing/Yahoo? You may not find as much traffic, but we’ve always found it to have lower CPA…

  • GuavaUK

    If Google released a -1 button I’d seriously be using it on this post.

    I have seen the data myself in numerous analytics accounts, and even from Googles own researchers quantifying both incremental traffic gains and cannibalisation effects from running PPC and SEO together.

  • Matt Van Wagner

    Thank you for your comment, GuavalUK.


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


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