Click Forensics, a company that certainly benefits from press about high click fraud, has come out with another scary statement: Click Fraud accounts for more than 28% of clicks on content networks, which I assume means AdSense and similar types of syndicated networks. The overall rate of fraud is more than 16%, the company claims. Seeking Alpha covers it here.
The thing is, we’ve heard this before, and before that, and probably before that, and the response from Yahoo and Google is always the same: Click Forensics has got it all wrong. We catch nearly all fraud before anyone has to pay for it. All of this is overblown and misunderstood.
So why does Click Forensics keep at it? Who’s right here?
PPC Programs certainly are prone to “click hassles”. While there may be true “fraud” there are also the unproductive clicks from salespeople seeking leads, job seekers, and curious competition seeking info (not necessarily commiting fraud). The actual percentages are impossible to quote in blanket form. Certainly it is higher for certain industries. It would be great to know more about the methodology used to interpret/identify click fraud versus simple irrelevant clicks.
In managing client PPC Programs we see far less questionable ratios of clicks-to-contacts for long tail searches. One might assume that this is due to “click hassles” being more common with easy to find, top level, general terms.
Additionally, we see an increase in effectiveness when Ads are not run on “content networks”. I suspect that one would find a higher percentage of unproductive clicks in these areas.
My humble opinion is that it would be better, especially from a ROI perspective, to utilize SEO for visibility on top level terms, use PPC to reach niche searches that are highly qualified and less visible to the unwanted click monsters, and avoid content network ad distribution.
Does click fraud exist? I certainly think so. However, these reports would have better value if they discussed how to minimize it, control it, and get the optimum results from your PPC program.
Back when we were dealing with a variety of clients advertising online, from Nextel to Disney, Outpost to P&G, the thing we would always focus on with them was the top two desired outcome from visitors. Be it sales, software downloads, lead generation, questionnaire fill-outs, newsletter signups. As long as we could keep our CPA (cost per action) down to at or below the goal, we didn’t worry too drastically about “click fraud”. When your measuring results on the other side, it’s clear when things are not converting, and doesn’t really matter whether its click fraud or not.
N: Mark Grimes
P: 503-502-0185
W: http://www.ned.com/
C: – a better world, every day
A: 4888 NW Bethany Blvd, K5-222 Portland, OR 97229
2007 Ned Accomplishments & Milestones
=====================================
78,850 youtube neddotcom channel video views
59 microloans: http://www.kiva.org/lender/neddotcom
64 better world videos: http://www.youtube.com/neddotcom
345 members: http://www.ned.com/
2008 Ned Goals & Deliverables
=============================
1,000,000 web site page views monthly
500,000 youtube neddotcom channel video views
1,000 registered ned.com members
500 Better World Island monthly project revenue
150 microloans
25 good quality inbound web site links
10 ongoing monthly ned.com sponsors
7 new videos at neddotcom YouTube channel
6 projects anywhere in the world
3 microfinance workshop sessions in Portland
2 events where five or more Ned members meet FTF
2 thriving real world locations
2 or more totally unexpected spectacular things
boss1: “I want…same revenue this year.. …. recession …. .so… , tell me your thoughts..”
employee1: “ …. this is x%, that is y% …”
boss1: ‘mmmm…no … ‘
employee2:’…it was %….it would % ….may this one…..’
boss2: ‘hmmm looks good.. … but…you know’
employee3: ‘….microhoo …is …’
boss2: ‘yeah that’s …%…’
employee4: ‘….mobile … %….may …raise…%’
boss3: ‘this looks better, …..not enough……’
employee5: ‘…click fraud … ‘
boss123: ‘… great…keep things up….to % …’
employee1: ‘but ..we ….lose…….’
bose3: ‘…advertisers choice…’
bose2: ‘…. one after one…’
employee1: “ahh”
bose1: ‘..okey gentle…..I… ..calling 100 spartons, we s……prepare legal …..’
From our experience, the bad apples in the click fraud debate very clearly come from the “content” or contextual side of the search engines.
But the real problem is the lack of transparency from the engines. Advertising with Yahoo Search is a particularly vexing problem. A huge chunk of our traffic from the Yahoo “search” network comes from parked domain traffic that Yahoo in our estimation fraudulently classifies as search. This parked domain traffic rarely, if ever, converts, and is one of the big reasons that the overall ROI from our Yahoo campaign(s) sucks so badly.
So if the engines and particularly Yahoo can’t give us transparency, maybe then can first get honest about what constitutes true search traffic on their network.
They also missed the important caveat: Content networks aren’t the primary source of clicks. Saying 28% of content network clicks are fraudulent is meaningless. What % of ALL PPC clicks are fraudulent?
This debate reminds me of the Brian McNamee (Click Forensics) vs. Roger Clemens (Google-Yahoo) saga. He-said, he-said.
However, in this case, it’s hard to believe that GOOG & YHOO would be lying with all they have at stake (market caps in the tens of billions) compared to a little startup in Austin that has everything to gain but not much to lose comparatively speaking.
Having lived in Austin and tracked Click Forensics the past year or so, if CF is right, it’s hard to believe that GOOG & YHOO would have rested on their laurels over the past year 1) by not driving the click fraud % way down with all the engineering resources they have and/or 2) refund advertisers who have been victimized by click fraudsters they’re detecting, which I understand happens to a degree (arguably it’s not fraud if the advertisers are given proactive credits/refunds).
If McNamee-Clemens warranted a Mitchell investigation and Congressional hearing, I would definitely think CF vs. GOOG/YHOO would with all that’s at stake here.
At present things are not transparent enough. Search is currently saved by its easy to demonstrate KPI’s.
Once these KPI’s drop, which they certainly will… as search becomes more and more expensive. Then once this happens we will begin to see marketing teams having to justify the reduced ROI on thier campaigns to mgmt. Mgmt will question marketing performance. Alternatives spending plans will be sought.
Also, when times become tighter for GOOG and the rest it will be interesting to see what happens to CF. It would be very easy and tempting for them to relax their (hidden) rules a little. By doing this though they would only accelerate the issue for marketers.
We are beginning to see this with the more mature ad serving network. DoubleClick et-al are gradually becoming more and more transparent. This is being driven by the advertiser who is demanding more bang per buck in response to search.
They also missed the important caveat: Content networks aren’t the primary source of clicks. Saying 28% of content network clicks are fraudulent is meaningless. What % of ALL PPC clicks are fraudulent?
Whatever the number is, it is the search engines dirty little secret. It’s so sad because it’s cost users Billions. Why users don’t revolt is beyond me. Maybe, it’s because of the lack of good options.
You must remember this: A click is just a click!
😀 nmw
The click fraud thing has been around for ages. I am only surprised that they haven’t done anything about it up to now. I mean it is no good just releasing statistics to show that click fraud is being committed. They must also suggest how they are going to deal with it or else this will remain just another of many surveys.
I don’t think those are unreasonable numbers, especially depending on your industry, the networks, match types, and phrases used. Loads of small businesses do their own PPC and know nothing about how to do it properly. The default settings leave you wide open to non-productive and fraudulent PPC. People bid for top position on single words, broad match, globally, and with content matching turned on.
Even if they figure it out, how much time and effort are they going to spend to get it fixed? Get the charges refunded? Most advertisers simply aren’t big enough to get reps and all the helpful extras that might mitigate some of the damage.
Yahoo is far worse than Google on this, though it’s improved after Panama went into effect. Before, we were getting a few thousand a month refunded on just one client whose campaigns were set up properly. We wouldn’t have caught any of it if we hadn’t been using 3rd party site analytics software and had tracking codes on all of our campaigns. It took a lot of work to get Yahoo to agree that people in Vietnam weren’t looking on ski boat sites to find a personal injury lawyer in the Midwest. Gotta love those high quality “partner sites”… Content networks are the worst. No transparency, no control.
We were in PPC at the very beginning when Google was just a search engine and Yahoo was the only PPC. We have quit on all PPC because of fraud, you tell me: 50 clicks in one day from various IPs registered in New Haven, Connecticut at $18.00 per click for a medical software program of interest to only a few doctors.
Guess what is in New Haven? Guess whose vacations I was paying for – Google’s response, a $3.00 “fraudulent click” refund.
They do have everything to lose if this becomes known – and everything to gain by keeping it secret – remember when Microsoft claimed no viruses were affecting their OS systems. A click fraud program is a lot easier to make than a virus to attack the OS.
At one time PPC was an affordable option particularly for new websites. Now it is getting diffcult for new websites to rank high in the search engines naturally, and money spent on PPC goes down the drain on fraudulent clicks. This scenario can not improve, as dishonest people will try to beat the system, even if the big search engines come up with a strategy to combat the fraudulent click business.
I’m sure that Yahoo, Google, and others have “technology” that can “filter” clicks based on some criteria they feel indicates fraud. That’s not the entire solution, however. They do not have the ability to detect (without a doubt) that a particular click is fraudulent, such that an advertiser would never be billed for that click again, regardless of where or how it originates. Likewise, they do not have a technology that enables them to identify all of the (human) sources of fraudulent clicks, regardless of whether they are the actual perpetrators of fraud (ie. human clickers), or individuals who write software that generates fraudulent clicks.
As the amount of clicks rise that the engines and ad networks filter, their costs increase and their profit decreases. The engines not only do not make money off the fraud – they lose money, because not only are they unable to charge the advertiser for the filtered clicks, they incur a cost for performing the filtration.
However, I’ve been saying this for years now, and there still seems to be those who just can’t or won’t believe it’s true. Nothing I can do about it now. I’m out of the search industry. I’m in an industry where the people take fraud seriously, and discuss it openly (without giving away secrets). They frankly discuss the existence of fraud, the vulnerabilities that are afoot, and strategies to deal with the problems. The industry isn’t without problems – note the recent routing announcements gone astray that caused people all over the world to lose access to YouTube, but you will find the professionals in the industry publicly engaged with their customers and the public at large in identifying the causes of such problems and proposing solutions. (They even have remarks about click fraud, geotargeting, and other search-related issues as well!) You ought to go over to the NANOG mailing list and read what they have to say. It would be nice if there could be open, honest, and frank discussion about click fraud as there is on NANOG about errant route announcements, and other problems, but that isn’t happening.
I had a rant about this…
http://aboutresultsmarketing.com/marketing_blog/2007/09/30/beware-biased-companies-research-supports-productpart-1-click-fraud/
I think Click Fraud is a little overstated, and that ClickFraud Network is there to push their agency platform, and be “the mediator” in Click Fraud, thus forcing Google and Yahoo to make them consultants.
Don’t get me wrong, I think that Google and Yahoo should allow a third party into their networks to monitor clickfraud. Why exactly should I trust them to refund my money???
But bottom line, with Clickfraud’s shady “wolf crying” method, I don’t want them to be doing it… There are better companies with better technologies.
In addition, don’t you think that web marketing has “less fraud” and is “more accurate” than TV and Radio advertising which has only Arbitron and Nielsen “algorithmic evaluation” ratings.
Anyone who can’t make money off of PPC is because they have a competitor that is making more money per sale. (Or has brand strategist blunderheads just throwing money at PPC.)
I’ve been in both places before, so don’t feel to alone if PPC just doesn’t work for you 🙂 Just please, don’t kvetch that the problem is click fraud.
I agree, it’s another factor that hurts your bottom line, and yes, partially due to advertising dollars going down the drain- because of fraudulent PPC’s. We all want to maximize our ROI, and ultimately the problem can be resolved by establishing control over your PPC’s, decreasing fraudulent activities, and finally increasing your ad campaign results. I recommend looking at http://www.ClickFacts.com for more content and practical solutions to this growing problem in digital advertising.
Perhaps ClickForensics keeps at it because there is something to their claims. Apparently Yahoo has been convinced.
They are certanly wrong… it’s not only click fraud, but also cost per action fraud (which can be much more expensive to advertisers). Google and Yahoo don’t have full control, on the CPA side, google can’t even check that the “visitor” is using the same IP to convert several times a day.