My name is danah boyd and I'm a Principal Researcher at Microsoft Research and the founder/president of Data & Society. Buzzwords in my world include: privacy, context, youth culture, social media, big data. I use this blog to express random thoughts about whatever I'm thinking.

Relevant links:

Archive

Who clicks on ads? (Revisited with data)

Two months ago, I ruffled some feathers with a post called Who clicks on ads? And what might this mean? Lacking any good public research, I pointed to a blog post by an AOL Global Advertising Strategy guy talking about research they did on AOL ad clickers. The report was by no means generalizable to all ad clickers, but it made a significant point: ad clickers are not representative of the population at large. Still, there were folks that were annoyed that I wasn’t pointing to public data, especially when I continued on to make my own hypotheses about who these heavy clickers are.

This week, in a study called “Natural Born Clickers,” Starcom USA, Tacoda, and comScore found that “the 6% of the online population accounting for most of the click-throughs skews toward male Internet users ages 25 to 44 with household income under $40,000.” [see news brief; anyone have the full report?] “The study also found that their heavy clicking did not reflect high spending levels offline. They were also more likely to visit auction, gambling and career sites. The findings suggest that high click-through rates don’t necessarily boost branding campaigns.” In other words, “the click is dead.”

This study finds that the age and gender of heavy clickers differs from what the AOL report found. (This probably says more about AOL users than anything else.) Yet, their findings also support (but do not confirm) a portion of my initial hypothesis that heavy clickers are:

  • More representative of lower income households than the average user.
  • Less educated than the average user (or from less-educated environments in the case of minors).
  • More likely to live outside of the major metro regions.
  • More likely to be using SNSs to meet new people than the average user (who is more likely to be using SNSs to maintain connections).

Folks tend not to like to hear that heavy clickers skew towards lower income levels, but I still believe this to be true. (For the record, 2006 median U.S. household income was $48,201.) Also, I should note that the population who uses SNSs to meet new people most likely skew male and 25-44, although not exclusively.

Hitwise also came out with new data this week: Yahoo search draws a younger audience, but Google users are more likely to spend more online. What I find particularly intriguing about their report is this graph:

Now, I don’t know what all of these labels mean, but terms like “Affluent Suburbia” and “Upscale America” lead me to believe that the Hitwise bubbles are saying that people who spend lots of money offline are also the most likely to spend more than $500 offline.

Now, if you put these two reports together, you get a funny image of what’s going on. Wealthier users are more likely to spend money online, but they are less likely to click on ads. Poorer users are more likely to click on ads, but not likely to spend money online except in a few verticals. Wouldn’t this then mean that Google is more likely to get the eyeballs of those likely to spend money, but statistically less likely to make money off of their clicks? This would seem to conflict with the TechCrunch post that suggests that the Hitwise data proves why Yahoo is in deep doo-doo. Given that both Yahoo and Google search generate revenue through click-throughs and not impressions, wouldn’t these two reports conflict with TechCrunch’s assessment?

Print Friendly, PDF & Email

12 comments to Who clicks on ads? (Revisited with data)

  • Kat

    The bubbles’ labels are Mosaic demographic segments: http://www.appliedgeographic.com/mosaic.html

  • ts.info

    This study is being interpreted in nearly polar opposite views by different observers:

    Some view it as the end of the click, others view it as only that click *measurement* should be abandoned. The study’s authors, for instance, believe that the CPM model is more important than CPC, and click performance is not the correct measurement:

    �While the click can continue to be a relevant metric for direct response advertising campaigns, this study demonstrates that click performance is the wrong measure for the effectiveness of brand-building campaigns,� said Erin Hunter, executive vice president at comScore. �For many campaigns, the branding effect of the ads is what�s really important and generating clicks is more of an ancillary benefit. Ultimately, judging a campaign�s effectiveness by clicks can be detrimental because it overlooks the importance of branding while simultaneously drawing conclusions from a sub-set of people who may not be representative of the target audience.�

    Several other views of this study espouse the end of CPC and interruption advertising, with either CPA or “featuretisement” being the future:

    http://www.computerworld.com/action/article.do?command=viewArticleBasic&taxonomyName=internet_business&articleId=9062938&taxonomyId=71&intsrc=kc_feat featuretisement

    http://news.softpedia.com/news/Heavy-Clickers-an-Advertising-Nightmare-78627.shtml

    http://www.accuracast.com/search-daily-news/ppc-7471/bad-news-for-display-advertising/

    The official press release http://www.smvgroup.com/news_popup_flash.asp?pr=1643

  • Google uses both (or “either”) Cost-Per-Click (CPC) and Cost-Per-1000-Impressions (CPM) advertising for their Google Adsense program. CPC advertising is obviously more expensive for each unit, but Google is smart enough to realize that CPM in the aggregate can make a significant income (although doesn’t necessarily sell any actual products).

    So, I guess what is actually going on here is that Google continues to make money, despite any actual efficacy of their product for selling goods. The trick, and Google works hard to create this, is that Google has convinced everyone that their product works. This is why Google boxes up their search algorithm, keeps tight-lipped about revenue sharing for advertising, and is diligent in affecting a “clean” internet (see their recent hi-jacking of 404 pages, their identification of virus-laden pages, etc.). By “clean” I mean, naturally, good for selling advertising.

  • This is great information and actually confirms my beliefs about click through ads and rates. I just finished a new business plan and part of the model is ad based. I modeled click-thru ads as a very small percentage of total ad revenue for a variety of reasons but what you discuss here is one of them.

  • “pointing to public data”, even. 🙂

  • brad

    Who clicks!?! Who even *sees* the dang things?

    One day a year or two ago I noticed that I couldn’t recall seeing a single banner ad in – how long? Months? Years?

    It’s funny we talk about clicks versus impressions when it seems increasingly likely that the majority of people acclimate to banner ads and eventually stop seeing them at all. The brain has a marvellous capacity to filter out what it regards as noise. With 90x% of us regarding banner ads as noise, that’s a lot of invisible banner ads.

    I’ve felt quite isolated in this attitude, surrounded by marketers as I am, but it seems I’m not alone.
    http://www.useit.com/alertbox/banner-blindness.html
    (You can tell it’s science because they use eye-tracking and heat maps 😉

  • sarah bluehouse

    I wonder if there is a concordance to use of public/friends terminals and emotional availability of ad-blocking software.

    Also I know that I’m kind of proud of my trained ability to not see adverts.. unless they scream, pop, jump in size to cover the box i actually *wanted* to click. In which case, I hate them. alot.

    But I suspect much of affluent america feels the same way, because affluence comes with the “I am not a tool of the man, but I am the man, therefore, the content behind banner adverts is an utter waste of time.”

    but the click-throughs, may not be *actually* immediately worth actual sales, but govern taste, which defines buying purchases, which then leads the consumer back to the store, or to purchase a similar product elsewhere… kind of collective marketting amongst secret pals.

  • With click-through rate around 1%, I still believe the race is far from over.

    TechCrunch might implicitly assume that improving average spending per buy is hard (as in “brick & mortar”, as in “this has been tried for decades”); click-through is solvable by several expected evolutions:
    – technical improvement in targeting technology – Google is supposed to be better and improving faster;
    – opening niche markets.

    The more I see the methodologies from studies by private American boutiques, the more I miss good old planified, public statistic & government census: it might be grey, but as least you have a sharp picture of what happens.
    OK, that was not helpful. But when you grow up with .5% confidence intervals, you miss them.

    > people who spend lots of money *offline* are also the most likely to spend more than $500 *offline*.

    You are denouncing a tautology or there is a spelling mistake?

  • eas

    Keep in mind that anyone running a pay-per-click campaign has every opportunity to track how those click-throughs end up and then adjust the price per click they are willing to pay to something rational.

    This might impact people who are trying to do brand advertising online, but I’d imagine that anyone of any sophistication that uses pay-per-click advertising to drive measurable sales is already setting their price accordingly.

  • Kevin Cantu

    The closest match I can think of to CPC advertising in magazines is for phone-sex. So I’m not surprised that there’s plenty of consideration given to CPM.

    Not all of us really want to click the monkey to win something that will make our penises bigger and help us meet hot single girls next door…