[Editor's note: We're bringing back blogger Kevin Laws for a second post, given that his one yesterday stirred the crowd. This time he writes about click-fraud.]
Recently, the Washington Post joined BusinessWeek and the New York Times in misreporting the problem of click-fraud. Each publication focuses on the hoards of clickers hired to perpetrate click-fraud and how it’s really damaging advertisers. Except it’s not really hurting advertisers.
BusinessWeek story even goes over the top, implying that Yahoo and Google are actually profiting from click-fraud, so have no incentive to stop it.
Click-fraud clearly exists. There really are people of evil intent hiring hoards of clickers to defraud Google and Yahoo’s ad networks. Google and Yahoo do take a cut for every click sent. However, all publications are misleading about the extent of click-fraud and about who it hurts.
BusinessWeek uses the example of Martin Fleischmann at MostChoice.com who spends $2M a year in cost-per-click (CPC) advertising and estimates he’s lost $100,000 in the last 3 1/2 years. Of course, BusinessWeek doesn’t do the math for you showing that’s just over 1% of his clicks, barely a rounding error.
This is not because click fraud is just 1% - it’s probably closer to 10% or so, from legitimate estimates I’ve seen. Martin isn’t losing 10% because he isn’t stupid. Using his MBA from finance-heavy Wharton, Martin is surely calculating the value each click brings him in terms of actual converted customers. He runs a mortgage site, and makes money only when he generates a legitimate lead for a mortgage company. If 100 clicks generate 1 lead worth $100, then each click is worth up to a dollar.
If click-fraud causes Martin to get 200 clicks, but still only one lead, then each click is only worth fifty cents. Martin adjusts his bid. He might lose a little money if he didn’t notice for a day or two (thus the $100K loss over many years). After all, what’s the difference between a “legitimate” prospect who doesn’t turn into a lead and somebody who never intended to in the first place?
After adjusting his payment, Martin still pays up to $100 and still gets $100 in revenues. Google and Yahoo still earn $100 in revenues for the clicks. So who is losing?
The real victims are the honest publishers. Every one of those honest clicks sent to Martin is now worth half as much because of the fraud. If you happen to carry Martin’s ad, you earned $1 for a click before, and now only get $0.50. Click-fraud doesn’t hurt advertisers - it hurts the people that carry Adsense.
Furthermore, the ad networks rely on publishers for the traffic they sell to people like Martin. If honest sites stop hosting Google Adsense, Google loses its network. Rather than profiting from it, Google and Yahoo are more motivated than anybody to solve this problem.
The ad networks are not innocent, however. Both Google and Yahoo are well aware of who really loses. They both have strong interests in not appearing to cheat honest publishers (and that would be a far larger class-action suit than the advertisers filed). It is far better to have a few disgruntled advertisers than to admit the real problem, so they are silent in the face of these reporting errors. Instead of correcting the logic, they are satisfied with merely pointing to their statistics on the smaller-than-perceived extent of click-fraud.
Click fraud is a problem, but it’s a problem for publishers, not advertisers, and Google and Yahoo are very motivated to solve it. Publishers are now the ones that should be evaluating alternatives to Adsense, not advertisers.
14 Comments
-
Wild Bill said:
Great article. Also checked out Vast.com, wow…thats exactly what I’ve been looking for for my business. Awesome site!
-
J said:
$100,000 is 5% of $2M, not 1%.
-
T said:
The article said he lost $100,000 over three and a half years, spending $2M a year. $100,000 is 1.4% of $7M.
-
Raja said:
There is a problem with the logic.
If the advertiser reduces his bid to $0.50 then he may not get many clicks and may not get the $100 customer. You are assuming that all the other advertisers will behave the same way and adjust their bids accordingly. This is not reality. Advertisers can’t make bid adjustments without considering competing bids. Click fraud certainly hurts the advertisers.
-
Peter Cranstone said:
Kevin,
What if the search engine had more information about Who was click on the link, What device they were using to click on the link and exactly “Where” they were when the clicked on the link. If it’s a real person, using a real device and the location changes each time could that help prevent click fraud?
Peter
-
Krish said:
This is convoluted logic. A little over the top. Click Fraud hurts Advertisers who end up paying more and that’s a fact - or am I missing something here which I doubt.
-
Kevin Laws said:
Addressing a few of the comments…Raja, if an advertiser drops their rates and drops out of top position, then whoever takes over would have the same click fraud problem, and would reduce their rates, too. That’s why there is some friction (after all, Martin did lose his estimated 1.4%).
Krish, it is the conventional wisdom that click fraud hurts advertisers, but it’s just not true. Even for “legitimate” clicks, your conversion rate might be 2% or so. Thus 98 of 100 people just leave. Any advertiser takes that into account in pricing the rate they’d pay. I fail to see why it’s so convoluted to point out that if that click through rate dropped to 1%, the advertiser would drop the rate they’d willing to pay. How do you think the rates get set, if it’s not based on actual conversion?
Peter, the search engines have plenty of information to reduce click-fraud. If you sign up for an account with Google or Yahoo, they have a “conversion tracking” feature which allows you to put a little pixel on your “Thank you for purchase” page. Ostensibly, this is so that Google and Yahoo can provide you with good data on who converted, which is useful. However, they use this data to track conversion rates by traffic source and knock out certain traffic sources. This is how they’ve managed to keep click-fraud under control at all (to the 10% or so level that most reputable estimates seem to come in at).
-
Raja said:
Kevin,
Your assumption that all advertisers drop the bid because of click fraud is not true in most cases. This may only happen if click fraud increases the cost of click to the point they are forced to change the bid. I suspect in most cases, click fraud increases the cost of click but still remains lower than revenue per click. In such cases, advertisers do not want to tinker with thier bids, but take click fraud as cost of doing business. This is true in our case, and I suspect this to be true in most cases. As you can see click fraud certianly hurts the advertisers the most.
-
alpha24seven said:
Not true. Not true at all. You are completely ignoring the “perception of quality issue”. Once advertisers start to believe that click-fraud is diminishing the quality of their campaigns in ANY way — they will place less value on the campaigns, which will at the end of the day place less value on the company providing the campaign infrastructure.
I’m building a thesis for a short regarding click fraud. Your position is completely incongruous with the advertisers I’ve spoken to.
-
Martin Fleischmann said:
Kevin, while a few points you make have a germ of truth, your article is mainly just wrong. You’re using made-up numbers and trying to say what is happening at MostChoice without basis. We lost more than $100K but I limited it to that, and our spend wasn’t $2M all 3 years, but beyond that you miss the real point of the problems inherent in click fraud — and what types of firms are really behind it.
BusinessWeek gave me the opportunity to publish my own take on the problem a few weeks ago, about 3 weeks after the original article came out, you must not have seen it. You and others would benefit from reading it and responding to it directly: http://www.businessweek.com/bwdaily/dnflash/content/oct2006/db20061016_120791.htm
-
Kevin Laws said:
Martin, I appreciate the first source information - your contribution is invaluable. I’d be happy to correct the numbers (which are not mine, but are the ones that appeared in the BusinessWeek article). If you send along the real numbers (even crunched, if you want to protect trade secrest) and I’ll include them in an addendum.
The hypothetical, I hope, is clearly presented that way to help explain the point - I’m not implying those numbers or rates are what you actually pay.
I have just read the opinion piece you published in BusinessWeek. You are correct, I had not read that before writing this. I was pleased to see that you did recognize this exact phenomena (that it hurts the advertisers if click fraud is evenly distributed).
You may be surprised to learn that I agree with your other points as well. To the extent that click-fraud is unevenly distributed and sporadic, it would hurt some advertisers more than others.
So perhaps I should be more specific - my background being economics, I’m tending to look at advertisers *in aggregate*. It’s like saying that trade, in aggregate, increases incomes - that’s true, but there are certainly individuals hurt and others helped by trade.
To the extent that click-fraud is an unpleasant, random event, however, it has the same effect I’ve described. If it drives you away from Google instead of leading you to reduce your rate, it has the same net effect on publishers - it reduces the rates they make and reduces the take for Google.
So while I appreciate your points, it doesn’t appear to change the conclusion - advertisers can limit the fraud through careful management, the rates do go down (whether through advertisers leaving or reducing bids), so it does affect publishes, and Yahoo and Google do have every incentive to fix the problem.
Neither the BusinessWeek article nor the New York Times article addressed any of those points - though clearly you corrected some of that in your own response.
-
Martin Fleischmann said:
Two new things: first, looking at your site I see why you are defending search affiliates. Certainly honest publishers of all sizes (Yahoo included) make less than they should because of click fraud and that’s part of my article (I told Yahoo that years ago), but it’s still the advertisers who are most hurt.
Second, like you I guess I should also promote our site (www.mostchoice.com) with a link to it so people can see that we’re not a “mortgage site”, another example of shoddy research on your part. You shouldn’t go asserting things or other people’s thoughts based on guesswork.
-
Martin Fleischmann said:
OK, just saw your response which I hadn’t before sending my last one (hadn’t refreshed). I’m glad you agree with my points, I tried to be reasonable and incisive on it, while BW over-dramatized a bit and wasn’t clear about diff. between Yahoo & Google and content & search networks.
But my central “economics” point disagrees with your bottom line — while bids do drop to some extent, the uncertainty that spikes of bad clicks create is inefficient and unfair for everyone in the system. At the very least it costs advertisers time and attention they don’t want to spend to monitor it and fight for refunds. You’re not in favor of letting the thieves siphon off money from the real sites as am I, but advertisers pay for this thievery, Google and Yahoo made money from it when it was a “dirty little secret” and most advertisers didn’t know better (many still don’t, big pubs are just now getting it).
Now it’s started to hurt Yahoo (and damage Google’s reputation) and the problem has become too big to ignore so they both take it more and more seriously. But their incentives were clearly to cover and ignore before since they made a % of the incremental fake traffic. I’ve had lots of emails from advertisers thanking me for sharing our experiences and wondering how they can recover some of what they now clearly see (as they finally look at their logs) was taken.
You’re right that publishers should be pissed by some lost revenue, but not that it doesn’t hurt more the paying customers who were robbed more directly. No need to worry about which group is bigger, we need to focus on how to stop the thieves by changing the system incentives first and foremost.
-
An Advertiser and Publisher said:
If you were an advertiser, you would have a totally different perspective.
I placed 1 ad and racked up a huge bill on AdWords with few conversions. I’m sure it wasn’t just my clever wording that got people to click.
As a publisher I care that Google and Yahoo can match ads well and that they have inventory. That’s an even bigger issue to me, than click fraud is.
As a publisher I don’t care one bit if there is click fraud. That’s all good for me.
