Interesting post on Traffick positing the theory that Google’s AdSense “faces extinction” unless Google does something about it. The author notes that AdSense doesn’t work so well for publishers with strong repeat audiences (I can attest to that), that click fraud is growing (I sure have proof of that with folks I’ve spoken to lately), and that new options are threatening AdSense’s base (like AdBrite and BlogAds).
net net, I don’t think we’ll see AdSense going anywhere, but I agree that upgrades are due, and I sense they are coming shortly. First up might be verticalization – so you can buy in large consumer verticals like autos, travel, etc. Second might be opeing up the network to let developers build mini-networks of endemically related sites. Now that’d be nice, eh?
UPDATE: Slashdot picks up this Newsweek article on click fraud, and my post above as well. Welcome, slashdotters, if you want to know more about Searchblog, head here.
11 thoughts on “Traffick: AdSense Teetering?”
Actually, the post is by my colleague Cory Kleinschmidt, but I agree with a lot of the points. Google has always been obsessive about the possibility that AdWords (and Google Search) would lose legitimacy — hence their care in designing the page layout, rules, etc.
But it has felt like a different standard of care was taken in building out AdSense so quickly. What Google is now doing is attempting to remedy this after the fact. The nice thing for them is that they can figure out some other place to earn revenue from within two years to replace that revenue, should they want to scale back their involvement in this type of advertising.
Ironically, I was browsing John’s blog today and sent him an Email about his copyright (which should be updated to 2005) … so here’s a second note that the Slashdot thundering herd is coming to his excellent commentary above …
I am one of the slashdot herd. But one of the smart ones. There’s not many of them. Especially if you factor in the slashdot editors. Anyway, enough about that.
The thing I don’t get is why the heck click fraud took so long to be abused (one) and noticed (two). I remember the days of AllAdvantage, how easy it was to setup a bot to click all over the screen and rotate through webpages whilst generating revenue from their banner ad app. They started cracking down, but that was years ago. This really isn’t too much different.
At the risk of being hoist by my own petard I suspect the more intelligent members of the /. community don’t both posting these days due to the poor S2NR.
Adsense click fraud has been around since Adsense came online and it was nearly a year ago that Count Scubula got busted by the feds for selling click-fraud software. It has just taken the mainstream meejay, and hence their fanboyz at /., a long time to catch up.
People are looking for all sorts of complicated ways to tie clicks into sales or solutions to detect fraudulent clicks. Some of this is sensible but because Adsense is an auction it self regulates a bit like the stock market. If advertisers are making less money with Adsense, due to click fraud or whatever, they will bid less for words so reducing their advertising spend. If the program eventually proves completely unviable there are still affliate programmes like Commission Junction that will finance small websites.
Evolution in action.
Eventually, I think Google will have to do something similar to what you’re offering on this site – go to time-based charges for ads instead of click-based ones. I won’t proceed to guess how it would be done, whether it would be based on a period of time or on how much time it had been available for display on various websites, but it will change to something else. Also I think your ad rate is too high, which is why nobody has bought your banner. $175 a week is a lot of money unless someone gets, I would guess, on the order of several hundred thousand click-thrus.
Those who click with no real interest will fade out with the momentary publicity. The next step is to find confirmation mechanisms that correct the extra (fraud or accident) clicks. The idea’s great.
But as you – Mr Battelle – are discovering (and showing at the same time), it’s relationships that tell over the long haul.
I’m spending >$100K/yr on Overture and Google and am seriously concerned about click fraud. Unfortunately, my alternatives are limited. I can see real site activity tied directly to my advertising budget. Many of the words I use are >$10 so it only takes a couple of bogus clicks to get expensive.
I’ve had trouble with a local competitor and found somebody clicking an ad every few minutes. Didn’t get charged but he didn’t know that.
I’ve contacted both Google and Overture asking for copies of the IP Addresses I’ve been charged for. Been refused on multiple occasions as this is “proprietary” information. Presumably I’ve got this “proprietary” information in my logs. The tough part is correlating clicks to charges. I know they won’t give this up as you would easily be able to see the fraudulent clicks and competitor clicks.
They’ve been very helpful offering to analyze my logs if I have specific questions. But finding those suspect clicks is the hard part. I’ve never gotten a refund when I’ve sent in logs for analysis.
The new problem is all the search sites that simply post Google and Overture ads. Many of these sites disguise the sponsored nature of their lists.
I can smell a class action suit coming. Google and Overture are aware of the problem and not providing users the tools to fix it. This is the same as a waiter simply telling you what dinner cost and that you should trust him.
Does anybody have any idea what the commission is from Google and Overture for these reposted ads?
BlogAds will never compete unless they decide to either (a) offer automated signup or (b) respond to their emails.
One of the problems is that Google has no way of tracking what is happening with a click after it has left their servers. Yes, they do have SiteStats ROI tracking tool, but I am afraid the rate of its adoption by advertisers isn’t all that great. As for trying to distinguish between the legitimate and fraudulent visits, the only possible way is to use a statistical scoring system, much the same way as spam filters work. Tooting my own horn here, but this is what the company I work for, Clicklab, is currently doing. We are applying several dozen tests to each visitor session to detect fraudulent signatures, and calculate the Click Inflation Index as the ratio of the sessions deemed fraudulent to total number of sessions. Feel free to download a white paper on click fraud from http://www.clicklab.com or email me at “de @@@ clicklab.com” with any questions.
Reading this post several months down the track, AdSense is certainly producing mixed results for publishers — I have seen a gradual decline in earnings as have others who post to Webmasterworld. As a PPC ad manager I no longer allow Adwords to appear on Content sites and Google has reduced the payout to publishers with its “smart” pricing, so I am part of that problem.
does anybody know of a good overview of the most prominent problems publishers and advertisers are having with Google Adsense? If so, please mail email@example.com
Thanx very much in advance