Experiments With Google's Contextual Advertising

from the let's-see-how-this-works... dept

When Google first announced their contextual advertising program that would let content sites post text ads that (supposedly) matched the content of the page, I thought it was a great idea. If it worked, it was a really powerful solution. Text ads weren't intrusive or annoying like flashing, bleeping, popping ads that were gaining in popularity. More importantly, if the ads were really relevant, then they could actually (gasp!) be useful to surfers, instead of just annoying. Plus, with the simplicity of signing up for Google's adwords program, and their absolute dominance of the search market, they already had the basic critical mass to become a full-fledged ad network. However, I started to have doubts. First, there were a number of "ifs" in my initial assumptions. The biggest one concerning whether or not the ads would really be that relevant. Furthermore, I began to agree with others who pointed out different search modes: browsing and searching. When people are "searching" they're looking for something that will answer their questions - and thus, are willing to click through on ads that might be relevant. When it comes to browsing, people aren't looking to go elsewhere, but are content with the content they're seeing at the time. Thus, on content pages, people are generally browsing, and ads would have much lower click-throughs. The end result of these lower click-throughs could lead to annoyed advertisers, who start pulling their ads from content sites. At least that's the theory. A number of people argued with me about this both on the site and via email, and I finally decided that the easiest way to settle the argument is with some empirical evidence. From the advertiser side, I've been running experiments for a few months, and haven't been that impressed. The clickthrough rates from content sites are running between 10 and 20% the rate of clickthroughs on search sites. However, the question remained as to how well they worked for content publishers. So, we're now starting an experiment with Google ads on Techdirt. We've placed the ads out of the way (beneath all the content), to keep them from interfering with the reason you're (hopefully) at this site. However, we are interested in your feedback, both good and bad. Either feel free to respond to this post, or send us email at googleadfeedback@techdirt.com. My initial reaction is that Google seems to think this site is all about anti-spam software and/or ambulance chasing lawyers. I'm not sure if this says more about Google's algorithm or Techdirt.
Hide this

Thank you for reading this Techdirt post. With so many things competing for everyone’s attention these days, we really appreciate you giving us your time. We work hard every day to put quality content out there for our community.

Techdirt is one of the few remaining truly independent media outlets. We do not have a giant corporation behind us, and we rely heavily on our community to support us, in an age when advertisers are increasingly uninterested in sponsoring small, independent sites — especially a site like ours that is unwilling to pull punches in its reporting and analysis.

While other websites have resorted to paywalls, registration requirements, and increasingly annoying/intrusive advertising, we have always kept Techdirt open and available to anyone. But in order to continue doing so, we need your support. We offer a variety of ways for our readers to support us, from direct donations to special subscriptions and cool merchandise — and every little bit helps. Thank you.

–The Techdirt Team


Reader Comments

Subscribe: RSS

View by: Time | Thread


  1. identicon
    rafer, 2 Jul 2003 @ 1:55am

    Google ads

    Our experience at Wifinder correlates with yours, except that i look at it differently. Like any other site, we have an 80-20 rule (20% users = 80% of traffic). Our core group of users clicked through like maniacs at the beginning, showing just how well the ads were targeted. Then clickthrough dropped like a rock. We are a tremendously focused site; we only get educated Wi-Fi public access users. As a result, our Google ads hardly ever change, and our users quickly got too bored to pay further attention. Somehow Google needs to figure out how to rotate ads or let us do it.

    link to this | view in thread ]

  2. identicon
    Anonymous Coward, 2 Jul 2003 @ 8:48am

    Spam? & Adsense

    Hmm. Spam ads. Maybe it IS you...

    We're running Adsense on 2 of our publications and are very happy with it.

    You're missing one very important point - yes, maybe the click-thru's on the content sites are lower, BUT the real question is "what is the conversion rate?" Because as an advertiser, the ads don't cost you anything until the click-thru's come. Yes, it would suck if you were being charged by the impression, but you aren't.


    link to this | view in thread ]

  3. identicon
    whit, 2 Jul 2003 @ 3:30pm

    Re: Spam? & Adsense

    BUT the real question is "what is the conversion rate?"

    Just for the fun of it, I'll still maintain that the real question is how well google's ads do compared to other ads placed in the same location.

    If the experiment only includes google's ads it doesn't seem like much of an experiment -- you get clickthrough rate x, which we'll stipulate is lower than what you would get for the same ad when placed by targeting on google.com.

    What would be interesting to me would be setting it up your ad serving so that one of (at least) three types of content is randomly selected for display at each page load:

    1. Google's targeted ads.
    2. Non-targeted text ads.
    3. Links to non-advertising content.

    (You could, of course, add any number of other types of content.) If you log the ad type, time of display, and time of click (if any) each time something is displayed, then you've got an interesting comparison set up; not only can you compare clickthrough rates for the different content in a reasonably equitable way, you can also chart, slice, and dice the results in any number of interesting ways...

    That's a fair amount of work to undertake for a "hmmm, I wonder" sort of experiment, though, so it's probably not practical. :)

    link to this | view in thread ]

  4. identicon
    gluefreak, 2 Jul 2003 @ 8:11pm

    No Subject Given

    The ads I'm seeing on techdirt right this second are TOTALLY IRRELEVANT to everything I'm reading on techdirt -- at least so far as I can tell. Three of the four ads are about some Baycol lawsuit. The fourth is about BestBuy and MSN. Certainly these aren't going to get much clickthrough.
    Over at Google weblog there is a tool that lets you test which ads Google would serve to your site. When I try it on my various sites, the results are terrible... Very discouraging.
    BTW Google weblog URL is http://google.blogspace.com/

    link to this | view in thread ]

  5. icon
    Mike (profile), 2 Jul 2003 @ 11:12pm

    Re: No Subject Given

    Yup. So far, I'm not that impressed. I have no clue what the deal is with the Baycol lawsuit ads. My only thought is that these seem to have insanely high clickthrough prices, so Google puts them on some sites so that it appears to boost their earnings.

    However, the ads that show up on the individual posts appear to be somewhat better targeted, though they often have no sense of "positive/negative". In other words, I'll completley trash a concept, and Google will place an ad promoting that idea. For example, I posted something trashing Business 2.0 for their decision to shut out internet traffic, and the ads link to places to subscribe to Business 2.0.

    link to this | view in thread ]


Follow Techdirt
Essential Reading
Techdirt Deals
Report this ad  |  Hide Techdirt ads
Techdirt Insider Discord

The latest chatter on the Techdirt Insider Discord channel...

Loading...
Recent Stories

This site, like most other sites on the web, uses cookies. For more information, see our privacy policy. Got it
Close

Email This

This feature is only available to registered users. Register or sign in to use it.