Last week there was buzz about how paid search is slowing, this week Fathom Online, which maintains a monthly keyword pricing index, reports that in fact prices dropped for paid search terms, by an average of two percent. Release is here.
I don’t know how you all feel out there, but I sense a backlash of sorts building toward paid search. The story has been too rosy, for too long. Journalists and Wall St. analysts are starting to look hard for chinks in the armor.
One of the most interesting, from my point of view, is clickfraud. I’ve been talking at length to a fellow who is a significant advertiser on Google and other paid search networks, and he is literally exhausted from chasing down all the fraud that is plaguing his ad buys, and really angry with Google for not being responsive to his requests for relief. I’ll be writing up that story and posting on it shortly, but if you have any information or insights on clickfraud, send em my way, or post em here.
Update” Charlene Li calls Google a “one trick pony” in this Bweek article, and SES show had a session on clickfraud, covered in this MediaPost piece. Why does every piece covering clickfraud end with a vague handwaving by someone saying “Smart companies are working on this, don’t worry about it…”?
17 thoughts on “Other Shoe on Keyword Prices, Clickfraud”
Take a look at this:
dude, the x’mas shopping season is over.
everything has declined. and surely by
more than 2%.
Fathom’s index consists of the top 500 generic search terms. As I understand it, this index doesn’t pretend to be a statisical sample of the entire population of keywords.
Could it be that the long tail is at work here, too? And that volume is growing among keywords outside the top 500, which doesn’t get captured by the index?
John posted about Safa Rashtchy’s analysis a couple days ago. In Safa’s report, he says it’s impossible to assess the keyword market by looking at prices in an index without considering volume:
“The Search Measurement Trap. We note that it is not too difficult to be misled by search pricing if one only measures a sample of keywords, no matter how big the sample is. With 12-15 million keywords now in use and expanding, there is really no statistically relevant metric that can show the actual pricing. The best approach is to contact advertisers, who are telling us pricing remains very strong….The main factor driving search revenue growth is in fact not pricing but increased advertiser spending (and new advertisers) and increased search volume. Search volume is in fact the most important factor to watch and we see no slowdown in that.”
Click-fraud has historical roots in the music industry.
“This pay-for-performance gaming of the system has a counterpart in the music industry. In London, I heard cases where a record company would hire students to hit the record stores around town and buy up copies of an artist they would want to promote. The volume in sales would push that artist up on the charts and generate a hit. In the US this practice was applied to radio stations where a radio syndicate would be paid to play a hit song more regularly than other songs.”
If you are going to “comment” on the industry shouldn’t you understand it better?
The companies that should worry are not companies like Google, Yahoo and other PPC companies.The money being spent on advertising through their networks allows detailed tracking and ROI analysis that is unmatched in any other type of advertising medium.
The companies that should worry are TV properties that sell ad space no one in the world is even watching. Someone actually paid to advertise during John McEnroes talk show!
Also if you can recommend any blogs with content that is not just postilations and ineundos it would be much appreciated.
I do think we are standing near the top of the PPC mountain of inflated expectations and are overlooking the vast wasteland that is the trough of disillusionment.
In fact for the last 6 weeks, I’ve been fascinated with click fraud (though mostly “postilations and ineundos”) over at my blog: http://gotads.blogspot.com
See that URL for a link to a SEMPO released survey of its members with 5 questions on clickfraud, though that survey was done way back in the halcyon days of Dec. 2004…
Sure Google is working on it, but your rhetorical question about the vague “It’s OK” feeling is why they are going to be burned by it. Eric Schmidt and others are selling that tonic, and if it’s not true that Google can solve it – and fast, that’s when the Google “Come to Jesus” moment will hit (as John Heilmann foreshadowed)
I wrote about Google being a one trick pony last year (http://aqute.typepad.com/aquteresearch/2004/12/google_one_tric.html)
and many of the blog entries on (http://aqute.typepad.com/aquteresearch/google_/index.html) allude to this theme
An interesting article about the “Army of clickers” by the Times of India
I like the comments on the issue, particularly Ian’s connection to the pop music gaming.
While partial gaming and messing with costs will likely be unavoidable, I’d think that the most egregious problems could be avoided.
I don’t understand the technical limitations but I know from watching one of those crime shows on tv that the fbi can supena records on chat board visitors, and trace the visitors to actual ip adresses of where a page was accessed from.
I should think that the pays for clicks could have vast swaths of ip addresses made un payable.
For example, why couldn’t I only offer to pay for clicks on my lake tahoe ad that derived from ip addresses in the united states? Or only pay for one click a month from a given ip address or other such limitations?
Is that hard technology to roll out? I’d think that the basic tools were in place that could be coded toghether, but maybe I’m wrong?
Here is the possible replacement for paying Google for search keywords.
EnCirca.pro is offering generic keywords in the .pro domain. The search engines give extra weight in its rankings to domain names exactly matching the search keyword.
Until now these were available only to licensed lawyers and accountants. Now anyone can get them.
Google, Overture, FindWhat…all of them have the ability to block alot of the click fraud. They will, in order to keep the industry going, CONTROL IT. The companies in India and elsewhere will continue to be around because they are not the only ones making money on PPC fraud. I wonder what percentage of Google’s income can be attributed to these scams. When the price goes up and the ads get hit…
The people in India can click away all they want to. Unless they are using open proxies those clicks are being blocked by most, if not all, of the PPC providers. On top of that most open proxies get blocked too, if the engine has half a clue about what it’s doing with its click algo’s.
On the net, PPC banner ads were the first use of the pay per click model. This was a mainstay of the adult industry online for many years. As a matter of fact the same problems existed with click fraud then as they have now. PPC banner click fraud was actually the downfall of many adult companies online.
Fact is there is no great technology to stop click fraud right now. Don’t assume that just because a PPC player is larger that their technology is better. That’s just not true. Fact is the technology to spot and detect click fraud is pathetic across the board.
The problem is a hard one to solve. Can it be done? Yes, I think it can. Will it be? That all depends on whether or not the engine solving the problem can withstand the revenue hit from blocking all of those clicks. Google could lose as much as 30% of it’s revenue if they implement a working, automated, solution.
PPC advertisers need to force the PPC’s into full disclosue and compliance or else nothing will change. Unless, one of these PPC players wises up and makes them all look bad 😉
I’ve been using Adwords for about a year and I am beyond uncomfortable with their system. I tried Overture and felt that it was a huge waste of money. I track my lead sources very closely and after a sizeable investment (by small biz standards) I could not attribute as much as one prospect inquiry to Overture despite high click throughs in short amounts of time. I brought on AdWords and at first it had measurable, good results. . . and then the ROI began to drop – a lot. There is no reason for higher click throughs and substantially lower conversion pcts. except for click fraud. I’ve gone from a believer to a cynic on PPC.
I think there’s a straightforward solution to click fraud that falls somewhere between ridiculously simple and diabolically clever. Namely:
Offer a money-back-guarantee, no-questions-asked, on every click delivered. “Pay for only the clicks you like,” Google could say.
This is fleshed out in more detail in my blog entry: “Killing click-fraud (and the competition) with one stone” — http://gojomo.blogspot.com/2005/03/killing-click-fraud-and-competition.html
Paid search is still very much a developing market. The danger with click fraud, in my view, is not the actuality of some number of advertisers suffering but the potential PR-driven/perception driven loss of confidence (esp. among new, unsophisticated advertisers) in the medium in its adolescent stage.
Ultimately this may spawn new models that would yield greater value for advertisers (e.g., CPA or Cost Per Lead) but less revenue for engines — at least in the near term. http://blog.kelseygroup.com/kelsey/index.php?/weblog/more/will_click_fraud/
My interest in click fraud detection is manifested through my blog http://noclickfraud.blogspot.com. In it I have noted that Google is rather good at ignoring multiple clicks and not charging the advertiser and that the real challenge is to identify a single click as suspicious.