A Bad Day For Spam

On Thursday Google  pulled the trigger on “content farms“, making a big change to their algorithm which they hope will result in less low quality sites appearing at the top of the results.  This has been a growing problem over the past couple of years, but in the last 6 months, the IPO of Demand Media and a lot of public criticism have resulted in the issue getting to the top of the agenda.

In their blog post, Google alluded to how big the content farm spam problem was – the change notably affects 11.8% of searches.  Although Google have never publicly stated how many searches they handle per day, estimates upward og 2 billion have been thrown around.  For convenience, let’s half that figure to 1 billion.

11.8% of 1 billion is 118,000,000 searches that were affected by content spam

Given that the number 1 result in Google for a query gets around a  30% CTR, that means that about 35,400,000 clicks on spam results.

Assuming that these sites are running fairly optimised AdSense campaigns, they should be getting a click through rate on their adverts of at least 5% – that means about 1.8 million clicks per day.  At about $0.10 per click, that means about $180,000 per day is being paid out to content farms – just over $65 million per year.

Google won’t publicly confirm the share of payments from AdSense that go to publishers, but even on a 70:30 split, Google’s change will have cost them somewhere in the region of $30 million.  Small beans to them, but a big statement to publishers about what is needed.

User Preferences…

Google recently launched a blocking extension for Chrome that allows people to flag websites that they don’t want to see again as they crawl.  While Google claim that the tool was not used in determining the identifiers for low quality websites that they were removing from their index:

we did compare the Blocklist data we gathered with the sites identified by our algorithm, and we were very pleased that the preferences our users expressed by using the extension are well represented.

The Fall Out

There was a predictable storm of attention about the change, and the initial impact in the US / com version of Google was that a lot of websites got nuked and lost most of their search traffic.  What’s not entirely clear yet is what the longer term impact will be – are the sites only going to be impacted in the search results, or is their ability to pass PageRank around networks going to be affected too.  If so, we might see a bigger change in the approach of SEO companies who currently make use of networks of interlinked content websites to create relevant back links to their clients’ websites.

5 replies on “A Bad Day For Spam”

  1. Actually James, Google do share split with publishers. Currently it’s reported as being 68% for the pubs, at least that’s what my backend tells me.(Not my arse)

    1. You’re completely right. Until last year, Google did not make the revenue share model public. In their Adsense Blog, last May, Google said that the revenue share was 68% for content publishers, and 51% for search.
      What’s quite interesting about the post I’ve linked to here is that they say that the rev share is the same for all content publishers. I was under the impression that larger publishers were able to negotiate special rates.

      1. Ah yes, and on that point who indeed knows…Publically perhaps, there has to be an appearance of unanimity with fair equitable distribution/rev share etc. That said, Google IS the Internet powerhouse with arguably the biggest display network of both publishers and interested advertisers. I don’t know of many publishers who can afford to ignore Google’s inventory, so maybe Google really are in a strong position to say take it or leave it. We’ll probably never know! 😀

  2. Actually, they didn’t say that it affected 11.8% of the searches performed, they said that it affected 11.8% of their queries. Depending on how you read that, a single query might be performed thousands of times a day. Also, I have seen estimates that Google gets 3 billion searches per day. Even considering that the assumption you made about all of the affected results had a content farm in the #1 spot (which I know from previous investigations to not be true, often they would show for 2-5) I still think that your final estimate of $180,000/day to content farms is probably very low.

Comments are closed.