Forbes reported a couple of weeks ago about the latest skirmish over numbers between Google’s Shuman Ghosemajumder and Tom Cutler of Click Forensics. Reporter Andy Greenberg went right to the heart of the discussion by asking Google to address the July 2007 report by Click Forensics.
Within online content networks, Click Forensics estimated that more than 25% of all clicks were fraudulent, up from about 22% in the previous quarter.
Instead of answering a direct question and helping find common ground, Shuman responds with characteristic misdirection. First he attacks the methodology of other third party auditors with a critique that doesn’t apply to Click Forensics, then uses another third party study to claim that Google is actually charging for fewer clicks than they should.
That’s just one particular set of numbers. The auditing firm, Fair Isaac, for example, estimated in May that on Google’s content network, 10 to 15% of clicks are fraudulent. On ads placed next to search results, they said that there was a negligible rate of click fraud, less than 1%. That implies an overall click-fraud rate of around five to 7%. The number of clicks that we proactively throw out is less than 10%. So then the question is really: How much are advertisers getting for free thanks to our detection methods?
Shuman is a politician or a magician and perhaps a little of both. Instead of watching his lips, think about what he actually said and take time to do some math. The information Schuman provided proves that 22-28% of the clicks on the content network are invalid clicks.
Schuman tells us that the clicks on Google.com are only 1% invalid. Accepting that number (which seems entirely reasonable) and his implicit assertion that this represents 50% of all PPC clicks, along with the estimate that Google doesn’t bill for 10% of all PPC clicks, we can conclude that the invalid click rate for the clicks that are not on Google.com is around 19%. (1% Invalid for 50% of all clicks with 10% invalid for all clicks means 50% of the clicks have a 19% invalid rate).
Take that 50% of PPC clicks that we can infer from Shuman are not taking place on Google.com and break it down a bit more, and you wind up with some of these clicks on the Google Search Network (of which AOL and MySpace are the largest members). Google doesn’t share any data about the percentages of clicks from these partners, but if we assume that 25% of the remaining 50% of PPC clicks – or 12.5% of the overall PPC clicks -come through the Google Search Network, we wind up with 37.5% of PPC clicks coming through Google Content Network. If we assume that invalid clicks from the Search Network is half as much of a problem as on the content network, we can say that:
- 100% of the PPC clicks have 10% invalid (Google Provided Data).
- 50% of the PPC clicks have 1% invalid (Google Search, Schuman’s Data) .
- 12.5% of the PPC clicks have 10% invalid (estimates for search partnership).
- 37.5% of the PPC clicks have 22% invalid (our estimate based content/adsense advertising).
Still using Google’s 10% of all clicks are invalid, let’s go a little further and suppose that the overall share of clicks from Google.com aren’t 50%. Let’s say they are 60%, which is what we have observed with our campaigns. (again, Google won’t share these figures at the moment)
- 60% of PPC clicks have 1% invalid (Google Search)
- 10% of PPC clicks have (search network) have 10% invalid
- 30% of PPC clicks have (adsense/content distribution) have 28% invalid clicks.
Clearly, this analysis is riddled with assumptions, but it is time for Google to stop responding to Click Forensics with intentionally confusing data. Come forward with a breakdown of invalid click by channel and let’s have an honest and open discussion about the remaining difference. Google might just find out that a little transparency goes a long way.
From SEO Consulting company Alchemist Media, Inc.
