Still crap.

On the upside, they raised the domain exclusion list option to 500.

On the downside…. you’ll need it.

Tonight I gave a talk to the Cambridge SEO Meetup group about 10 ways advertisers can reduce their exposure to click fraud. If you would like to download the presentation, you can do so here.

I just finished my presentation regarding click fraud and click quality to the Click Quality Council (try saying that 3 times fast.)

I was pretty nervous, not because I don’t feel confident in my knowledge about the subject matter, but because I know how fired up I can get about this stuff. So, to the audience who have checked out this site as a result, welcome. I encourage any and all questions about click fraud.

As Tom Cuthbert mentioned, my presentation was based on a blog post I wrote a few weeks ago.

You can find it here… 10 Ways to Reduce Your Exposure to Click Fraud

I’ve spent a LOT of time studying click fraud. You can find other posts here detailing some of my battles. In that time, I’ve caught dozens of parties engaging in it and studied them to figure out how they’re pulling it off so that I can fight back against it. I’ve created detailed reports and submitted them to the PPC businesses (Yahoo, Google, MSN, etc…) pointing out issues and have had a lot of success in getting credits for the bad traffic to my clients.

Unfortunately even though I may have been able to get credits for my clients, those credits usually end up just being eaten up by the same culprits. Also, the PPC businesses have become much more stingy when it comes to these giving these credits. If you can get any credits at all these days, it’s usually a tiny fraction of what you believe you are owed.

When people come to me now asking for help with their click fraud problems, I tell them not to expect to get any recovery from the PPC businesses and that they need to get in front of the problem and be much more vigilant in their click fraud prevention efforts. You can’t expect the PPC businesses to do it for you.

Here are 10 methods that any advertiser can use to reduce their exposure to click fraud:

1. Know your traffic sources – Most PPC systems have their primary display point (Yahoo.com, Google.com, etc…) but also have additional distribution partners for their ads. The majority of bad traffic and clicks come from the downstream distribution partners. Find a stats program that gives you full referral information. By seeing the full referral, I can see the downstream partners and measure the quality of the traffic I’m receiving from them. Make sure you have your tracking strings tagged with the correct PPC network identifier so you can know which distribution network the offending site is in.

2. Use the domain/channel exclusion options – Once you have a decent sample size of traffic from various referrers, use the domain blocking feature to prevent yourself from showing up on those sites. In this instance, I’m referring to Yahoo and Google in this instance as they are the only two at this point to allow this sort of blocking. Yahoo currently has a 250 domain limit, but you can push them to expand that number for you. Google allows you to opt out of entire channels like the Parked Domain channel.

3. Use whois to block domain groups – One thing that PPC businesses DON’T look at when they’re identifying click fraud is the registrant of the domain. Usually, if somebody is running a decent click fraud racket, they’re spreading the bogus clicks over a number of domains in order to fly under the radar. If they’ve got an actual site, can use the Reverse IP function at Domaintools.com to identify other domains they may own. Don’t be surprised if you see those in your logs, as well. Another option is to buy a Registrant Search report to identify domains they own and add all domains owned by that person/business to your block list for future preventative measures.

4. Use conversion tracking – A lot of catching bad traffic ties into basic web marketing. If you’re using PPC to sell something, you should have conversion tracking in place to know which keywords are producing your sales. If you have conversion tracking in place, you can see which sites are driving your sales. Google or Yahoo gives you a 6% conversion rate yet Site-A-Bingo-Winner.info has delivered hundreds of clicks and no sales? Block it.

5. Turn off or lower content bids – Yahoo, Google and MSN all have a lower tier Content Match distribution group. For Yahoo it’s Yahoo Publisher Network, for Google it’s AdSense. Those are those skyscraper text ads you see all over. They don’t usually perform very well. You have the option to opt out of them completely or edit your bids for those channels.

6. Don’t bid on high cost terms – According to a recent Click Forensics report, the overall industry average click fraud rate was 16.3% for Q1, 2008. It would be a mistake to take this number and assume it is universal for every term. The fact is that the more expensive the click, the higher the rate of click fraud.

7. Don’t bid for #1 – In line with #6, it’s assumed that the number 1 bidder for a term will be bidding the most and that click will be the most expensive. If you move down to the number 2 or 3 position, you may find you get less questionable clicks, not to mention less general curiosity traffic.

8. Restrict your geographic targeting – It’s not a mystery, but a lot of bad traffic comes from outside North America. This may sound prejudiced, but it has proven itself out time and time again. Traffic from Russia, Turkey, India has been problematic for a long time. New campaigns now (at least I know this to be the case in North America) start with North America as the default. There is problematic traffic in North America as well, but this may be one tip at reducing if you have your floodgates wide open.

9. Daypart your bidding – Humans tend to do most of their searching during the day. Bots (specifically the less sophisticated ones) will run 24 hours a day so the fraud rate increases at night. Turn your traffic off at night to reduce these opportunities.

10. Don’t use 2nd tier PPC systems – It’s tempting for new advertisers to see CPC’s on FoofyDoofy.com PPC system costing one tenth of what it would cost for the same term on Google. There’s a reason for that. Most of those systems deliver extremely low quality traffic. If you really want to test it out, throw $50 at it and track it very, very closely to see if it performs for you.

Conclusion:

Click fraud and garbage traffic exists. It’s not going to ever go away. Trusting the PPC networks to look out for your best interest will just leave you frustrated and with lighter pockets. There’s no one way to cut it out entirely but if you follow the steps above, you can SIGNIFICANTLY REDUCE your exposure to click fraud and low quality traffic. This is not a one time fix. Fighting click fraud requires constant vigilance and adjusting your strategies appropriately.

Boy, did Yahoo really screw the pooch this time!

Microsoft got tired of Yahoo’s game. Now Google might now want to do the ad distribution deal because they don’t feel like dealing with the regulatory commission hassle. And even if they did, what leverage does Yahoo hold anymore? NONE. I hope Yahoo enjoyed being the popular girl everyone was trying to get to go to the big dance. Now she gets to sit home on prom night eating her Haagen Dasz all alone. Well, that’s not entirely true. She’ll still be getting lots of attention… by way of class action lawsuit from investors for this total debacle.

So what does Microsoft do now? They were trying to jump start into being real competition for Google. Their MSN er… I mean Live.com really hasn’t done it. How do they go and get in front of massive amounts of eyeballs? Some are suggesting that AOL might be the next stop.

What about domain portfolios? Say… Marchex? The stock is weak because the public doesn’t understand them. Not to mention the parting gift you could make on Yahoo’s bottom line by taking away one of their bigger distribution partners.

As my wife can attest to, when something angers me I have a really hard time letting go. Some might call it perseverance, I think she just calls it annoying.

So, following up from last time when Yahoo said they found no problem with the traffic… They kindly suggested using their groundbreaking domain exclusion tool and change my tactic settings.

So, based on what they’ve told me (and what I’ve known for awhile about their detection systems), here’s the way to commit click fraud on Yahoo*…

1. Register a bunch of domains and park them.
2. Get a fresh proxy list, have a friends and family network, employ a PTR system, set up an arbitrage system on a second or third tier PPC system,
find open WiFi networks, employ a botnet, build a browser helper object or any other number of ways to generate clicks from multiple IP addresses.
3. Distribute small number of clicks on each domain but never have multiple clicks from the same IP address.
4. Stamp the CTR by flooding the domains with impressions that don’t click.
5. Profit.

Sure, eventually enough advertisers may ad the domain to their block list to make it ineffective, but for $7 I get to start ALL OVER AGAIN.

All they’re doing is looking at the lowest hanging fruit. The click fraud guys are MILES ahead! The pattern is right in front of Yahoo’s eyes if they even bother to check the registration of the domains. The domains I’ve added to my exclusion list are based on domain names residing on the same servers as what I’ve seen to be problematic domains. Of course, they can just go register a few more domains and start again. Before you know it, I’ve hit my 250 max.

And I’m one of the GOOD GUYS! I’m sure the click frauders have many more ways than what I’ve detailed above.

It’s 2008. Try to keep up.

*Obviously, I DO NOT recommend or advocate doing any of the steps above. If you do, you’ll be stealing and that’s illegal. So don’t do it. But if you’re an advertiser, you need to know just how easy it is to defeat their ‘sophisticated’ systems.

After three weeks and me calling and saying “Hey, remember me?”, Yahoo finally came back with the results of the traffic review. It was the standard boilerplate response of finding no problems with the traffic. Despite providing IP addresses, geo identification, terms, referrers, domain ownership evidence, timestamps and pageviews, Yahoo couldn’t find any issues. Really. I feel like I’m taking crazy pills!

So anyway… $100 down the drain. But not really, as I made sure to also point this info out to a few respected journalists and executives at click fraud detection businesses.

I’ve been a click fraud hound for many years now.  People that know me closely will tell you that I tend to obsess over this stuff.  Yahoo has been a thorn in my side since I started because their traffic quality plummeted after they lost MSN and started filling in with no-name search partners and parked sites.  Click fraud is actually what first led me to my interest in domains and learning about that whole world.  I learned that not every domain parker is involved in click fraud, but they’re sure getting a black eye from it by not speaking out more against it.  It surprises me that they don’t do more to protect their golden goose.  But I digress…

Last night I set up a new campaign in Yahoo.  First time in a long time because of my bad experiences with them.  But, I figured it would be as bad now that I’m bidding on MUCH less expensive terms.  WRONG-O!  The following is what I learned from a 50 click test.

I loaded thousands of geo-specific terms.  Los Angeles (keyword), Montana (keyword), Tuscaloosa (keyword).  You get the point.  I turned it on around 10:45 PM.  Using Yahoo’s close-to-real-time reporting I watched the clicks adding up.  I was seeing CTR over 75%.  When I got close to 50 clicks, I paused the campaign and went in to my logs to check the clicks.  I took the IP addresses and ran them against a geo-location database.  I wish I could say I was shocked by the results.

Now, what someone in West Virginia is doing looking for a Chicago (keyword) is beyond me.  Sure, it could happen.  But not 99% of the time like occurs above.  This is one of those things where taken at face value maybe wouldn’t look like bogus traffic.  Yahoo certainly didn’t seem to have a problem with it as they charged me for all of them.

Yahoo has finally taken a BIG step towards solving their traffic quality problems.

For years now, their have been complaints from far and wide about their distribution network (read: parked domains).  And for years now they’ve done nothing about it, leaving it up to the advertisers to solve the problem by adjusting their bids.

Now they’ve enabled a feature allowing you to load in a list of domains on which you don’t wish to appear.

It’s a major step, and it will be good for those clean domainers out there because advertisers may begin to feel more comfortable with the traffic and start up the bid races again, resulting in greater income.

However, just sitting here thinking about it, I can foresee some problems.

1.  As an advertiser, you can only block the domain once you’ve been hit enough to determine it’s bad.
2.  Once a domain has been blocked by enough advertisers, the fraudster will just go buy a new ‘clean’ domain.  Rinse and repeat.
3.  Some decent person may come along and buy a domain and find out too late that it’s already been poisoned and added to many advertisers’ blacklist.
4.  Once a domain is on the advertiser’s blacklist, it may never be reviewed again for quality despite new ownership.

Some people will get hurt by this, I suppose.  It will be mostly those advertisers that don’t utilize the feature or have adequate tracking and by domain acquirers that don’t do their due diligence prior to picking up a new domain.

*Caveat* This new feature just started.  So far, I think it’s only available to platinum level advertisers.  And whether it actually works or not remains to be seen.

Apparently TrafficClub (Moniker’s parking service) is delaying payment because an upstream partner thinks they smell click fraud.

I’ll tell ya, I would hate to be a parking company these days. Their margins are getting seriously squeezed. Yahoo and Google are taking bigger cuts(why? because they CAN) and there are so many parking services now that domainers will hop to a competitor for a higher rev share.

On top of that, they’re getting slammed for allowing click fraud. I’ve been deep into tracking click fraud for a couple of years because I’ve been an advertiser for one of the most expensive PPC areas. Obviously, the amount of attempted click fraud grows exponentially with the price of the click.

Click fraud drives me insane because it’s such a preventable problem yet Yahoo and Google do so little to really prevent it. Well, I think Google actually does a decent job.

What I find curious is that G or Y may do these investigation and they may find some evidence of bogus traffic. They then withhold payment. But credits back to the advertiser are rare. So, who keeps that money?