Saturday, September 23, 2006

Is click fraud really getting worse?

Both Business Week and the NY Times had articles on click fraud this week. Both claimed that the problem was "growing", but neither presented any evidence that the problem is any larger (in percetage terms) than it was, say, six months ago. Sure, they have some fresh faces and newer quotes and numbers and some colorful anecdotes, but nothing that they presented suggests that the problem is any worse than before for the overall online advertising market.

The Times quotes a number of 14% for fraudulent clicks. BusinessWeek quotes a range of 10% to 15%. Well, my recollection is that this is the same range that we were hearing about six months ago. So, the overall impact is unchanged. The magnitude of the problem would not appear to be worse.

Now, whether the actual problem is worse, better, or unchanged, is not clear, but it is also not made clear by these "new" articles.

One interesting metric that neither article offered is what percentage of raw clicks are filtered by the ad processors such as Google and Yahoo.

To truly understand the market, we need the following metrics as well as ratios between them:

  1. Ad impressions displayed.
  2. Ad clicks.
  3. "Fraudulent" clicks filtered by the ad processor (Google, Yahoo, et all)
  4. Clicks received by the advertisor.
  5. "Fraudulent" clicks filtered by the advertiser.
  6. Cost of a click to the advertiser.
  7. Cost of a non-fraudulent click to the advertiser.
  8. Cost of the click to the ad processor and the net revenue from the click after publishers gets their cut.
  9. Clicks "converted" to active prospects by the advertiser.
  10. Cost of a converted prospect.
  11. Clicks converted to "sales" by the advertiser.
  12. Cost of a converted sale.
  13. Suspicious clicks estimated by the advertiser.
  14. Estimated cost of a suspicious click.

But, the articles did not enlighten us on most of these metrics.

Judging from at least one of the quotes, it seems as if some advertisers are judging almost any un-converted click as "fraudulent", at least if the conversion rate is below industry "norms" or even below the advertiser's own expectations.

Another theme that came through from at least one of the quotes was an unrealistic expectation that if you had a certain conversion rate from one source of clicks one month, that a lower conversion rate in a subsequent month or from another source was a per se indication that click fraud was occurring. Rather, this might also indicate that either the advertiser is naive or playing the reporter as being gullible. It may be that the advetising campaign has run its course and that the advertiser has neglected to upgrade its products, services, or offers, or that it doesn't fully understand the needs of some niches of its audience. It might also be true that the market for that advertiser's goods and services may be saturated or that some niches are saturated. To suggest that a lower conversion rate automatically implies fraud is quite a suspicious leap. I'm appalled that the reporters were not a little more suspicious of the advertisers' claims.

It may simply be that some advertisers feel that their online ads are being less effective than they would prefer. That may well be true, and may in fact be the source of the anxiety expressed in the articles, but it isn't an angle that the articles pursued in any depth.

After reading these articles, the thought that leaps to my mind is that I am experiencing journalism fraud. I have an expectation that the bulk of an article in a publication such as Business Week or the NY Times will have very high value and that alleged "news" is "new." I feel that only 15% of the text of these articles was new and noteworthy and substantive, and hence that 85% of the text of these articles was a fraud since it was promoted as being new and news and noteworthy and substantive, but was really old and mostly recycled "noise" or insubstantive.

Final note: Please do not click on any of the ads displayed on my blogs and web pages unless you really are interested in the advertiser's message or offers. To click when you're not seriously interested in at least seeing and hearing what the advertiser has to say would of course be click fraud.

-- Jack Krupansky


Post a Comment

Subscribe to Post Comments [Atom]

<< Home