Search Engine Watch
SEO News

Go Back   Search Engine Watch Forums > General Search Issues > Search Industry Growth & Trends
FAQ Members List Calendar Forum Search Today's Posts Mark Forums Read

Reply
 
Thread Tools
Old 11-07-2006   #1
Discovery
Jerry Nordstrom
 
Discovery's Avatar
 
Join Date: Sep 2004
Location: San Diego
Posts: 1,029
Discovery has much to be proud ofDiscovery has much to be proud ofDiscovery has much to be proud ofDiscovery has much to be proud ofDiscovery has much to be proud ofDiscovery has much to be proud ofDiscovery has much to be proud ofDiscovery has much to be proud ofDiscovery has much to be proud ofDiscovery has much to be proud of
Has PPC Bot the Farm?

I recently read a detailed article in wired about the latest in the Bot Wars. I certainly understood that bots were an issue, but I didnt fully understand the magnitude of the issue until I read this article.

Junk Clicks, and Blue collar click fraud are one thing, but bots that can methodically take down DNS servers, crushing e-companies at will, and no doubtely deploy complex click schemes to fool the likes of Google and YSM content programs could spell disaster for PPC as we know it today.

I have to say, with the increasing amount of click fraud I have dealt with over the past 4 years and the lack of response, even the straight out denial by some SE's has me wondering about the long term growth of PPC.

In my view if the SE's continue to ignore and deny the magnitude of the issue in order to gain short term, but large profits, we may lose PPC as a viable advertising medium altogether. If they admit it and address it we may have a chance to create a system to defeat it.

Have your concerns with bots grown over this past year?

Discovery
Discovery is offline   Reply With Quote
Old 11-07-2006   #2
jbgilbert
Old Timer
 
Join Date: Jun 2004
Location: Texas
Posts: 100
jbgilbert has a spectacular aura aboutjbgilbert has a spectacular aura about
Has my concern grown?

Yes and No...

For certain topics, I would consider click fraud to be a real issue.

For other topics click fraud is not a problem at all.

However, it probably needs to be dealt with better than it is being dealt with today -- that's an assumption since I don't know exactly what the SEs are doing about it today.

Will PPC be doomed because of click fraud? Probably not, but many companies in the "wrong" topical area may get driven from PPC because a combination of click fraud and escalating click prices will make it next to impossible to produce an acceptable ROI on a PPC campaign.
jbgilbert is offline   Reply With Quote
Old 11-08-2006   #3
Astir
Member
 
Join Date: Aug 2004
Posts: 15
Astir has disabled reputation
Blue collar click fraud? Is that when real people actually click on the ads as opposed to an automated system?

I don't things have changed qualitatively in the past year. I no longer advertise on the second-tier PPC services (too much fraud in the past), so I don't know if things have changed there, but I doubt they have.

Google AdWords doesn't have nearly as much fraud (that I can detect) as they did two years ago. There is still some, but overall they have done a decent job.

MSN has never had a fraud problem as far as I can tell.

Yahoo is getting worse, as their people have become increasingly desparate for new revenue and willing to take on low quality affiliates.
But it's the same type of fraud as far as I can see.
Astir is offline   Reply With Quote
Old 11-08-2006   #4
cogbox
Member
 
Join Date: Nov 2006
Posts: 8
cogbox is on a distinguished road
It's how they respond to fraud that matters.

I think the biggest issue is the service position that the site takes when fraud is discovered. Here's an example of the bad:

I recently created a new campaign on Ask.com. I've used them off and on over the years, but generally had a better experience back when they had the Branded Response/Featured Listings program. Haven't liked the re-branded Looksmart interface.

So, the new campaign went live and we burned through $321.71 in 24 hours--fine if it is good quality traffic. Problem was that of the 382 clicks, our tracking only picked up 244 visits, and of those only 3 visited more than a single page (normal Ask.com search visitors view multiple pages 40%+ of the time). 95% of the traffic came from two affiliates - Findology.com and ABCsearch on just three phrases (not the most common phrases normally).

I sent this information to Ask.com, and they responded with boilerplate about what defines a "valid" click for them. Among their rules defining a bad click:

There are multiple clicks on the same listing for the same visitor, more than once within a 2 second time span. A visitor that clicks more than once within 2 seconds could be someone trying to inappropriately inflate clicks to that website.

So, the same visitor clicking on a listing every 2.01 seconds is considered valid!

Their email asked for more information about the bad visits, which I promptly sent--and they promptly ignored. Just as they've ignored the four email I've sent since then on the topic.

Needless to say, that campaign is dead to me. As is Ask.com.
cogbox is offline   Reply With Quote
Old 11-09-2006   #5
sherwood
Member
 
Join Date: Jul 2006
Location: Cambridge, MA
Posts: 36
sherwood is on a distinguished road
cogbox, you're not the only one to have problems with Ask:
http://forums.searchenginewatch.com/...2&page=1&pp=20

Of course, now that they supply Lycos a few people will be tempted to try them - hopefully they do their research first.

To fight click fraud and its devaluing effect on PPC, the SE's need to realize that it's:

- 25% technology issue
- 25% customer service issue
- 50% convincing their CFO that they MUST sacrifice click fraud revenue to maintain trust with advertisers, even if that sacrifice is substantial. (I think this is the biggest reason why fraud seems
sherwood is offline   Reply With Quote
Old 11-14-2006   #6
Discovery
Jerry Nordstrom
 
Discovery's Avatar
 
Join Date: Sep 2004
Location: San Diego
Posts: 1,029
Discovery has much to be proud ofDiscovery has much to be proud ofDiscovery has much to be proud ofDiscovery has much to be proud ofDiscovery has much to be proud ofDiscovery has much to be proud ofDiscovery has much to be proud ofDiscovery has much to be proud ofDiscovery has much to be proud ofDiscovery has much to be proud of
Astir - Yes the phrase "blue collar click fraud" was the phrase that came to mind to distinguish "human - mom and pop variety" click fraud from BotFraud.

2nd tier engines have traditionally been huge JunkClick factories and many still are... especially the likes of Findology and ABCsearch. We have found a few that are changing their ways and we are getting good results. Perhaps, they are realizing that some are getting fed up with the bigs and that putting on a white hat may be profitable.

JB and Cogbox, my biggest concern is that the SE"s are not addressing the issue. It is hard to say if this is an intentional stance on the issue. They say they are doing things behind the scenes; however any solution that truly attacks click fraud and junk clicks must involve the advertisers in some way. So how can they be attacking this issue and never working with us to do so?

Do SE's have any incentive to stop click fraud when they can potentially derive huge profits from it?

What exactly does each SE's advertiser network look like? Could they not report to us which "partners" would make up 80% of the traffic on each SE? What if they truly broke this network down and showed us their names, which are tool bar vendors, which are search partners and which ones are REALLY content partners?

What if they passed through all click data, instead of cloaking to make it difficult for advertisers to determine who is referring the traffic?

And what if (and you knew this was coming) they allowed us to block the sites/domains/IP's/ToolBars that were producing traffic that did not convert; from WITHIN our interface - NOT by a call to CS and an Inspector Clouseau style investigation.

Bottom line; without transparency, clear plans or features for advertisers to self police all click fraud/junk click activity there is a lack of trust between advertisers and the SE's. Some Bots may be created or have been created to exploit this weakness in our relationship in order to drive a click fraud wedge into the heart PPC.


Discovery
Discovery is offline   Reply With Quote
Old 11-14-2006   #7
Mel66
Oversees: Google AdWords
 
Mel66's Avatar
 
Join Date: Jul 2004
Location: Michigan
Posts: 775
Mel66 has much to be proud ofMel66 has much to be proud ofMel66 has much to be proud ofMel66 has much to be proud ofMel66 has much to be proud ofMel66 has much to be proud ofMel66 has much to be proud ofMel66 has much to be proud ofMel66 has much to be proud of
Quote:
Originally Posted by Discovery
What if they passed through all click data, instead of cloaking to make it difficult for advertisers to determine who is referring the traffic?
On that topic, has anyone noticed their percentage of "unknown/unidentified" referrers increasing lately? In every pocket of my PPC campaigns where performance has declined sharply in a short period of time, I inevitably find that the top referrer for said campaign is "unknown." It's worst in Yahoo, but even in Google I'm seeing this. "Unknown" is in the top 3 referrers for many of my PPC campaigns.

Are these all bots, or what? It's becoming a big area of concern for us, and I fear it'll get worse as the holidays draw closer. And it smells fishy to me when it's largely non-converting traffic.

Melissa
Mel66 is offline   Reply With Quote
Old 11-15-2006   #8
cogbox
Member
 
Join Date: Nov 2006
Posts: 8
cogbox is on a distinguished road
Unknown Referrers

Quote:
Originally Posted by Mel66
On that topic, has anyone noticed their percentage of "unknown/unidentified" referrers increasing lately?
Actually, I have noticed an increase in the number of unidentified referrers lately. For example, here are the percentage with unidentified referrers for one client's adwords campaign:

june: 4.7%
july: 6.2%
aug: 5.8%
sep: 8.2%
nov: 11.0% (so far)
cogbox is offline   Reply With Quote
Old 11-16-2006   #9
Mel66
Oversees: Google AdWords
 
Mel66's Avatar
 
Join Date: Jul 2004
Location: Michigan
Posts: 775
Mel66 has much to be proud ofMel66 has much to be proud ofMel66 has much to be proud ofMel66 has much to be proud ofMel66 has much to be proud ofMel66 has much to be proud ofMel66 has much to be proud ofMel66 has much to be proud ofMel66 has much to be proud of
Wow, our percentages are 2-3 times higher. And increasing. Interesting.

Melissa
Mel66 is offline   Reply With Quote
Old 11-16-2006   #10
Jeff Martin
 
Jeff Martin's Avatar
 
Join Date: Jun 2004
Location: Dallas, Texas
Posts: 364
Jeff Martin is just really niceJeff Martin is just really niceJeff Martin is just really niceJeff Martin is just really nice
Having run a free click fraud service for awhile, helping advertisers out when I can and having been given the opportunity to work the Lane's Collectibles case I believe that the PPC networks don't turn a blind eye to click fraud. They do have measures in place to help filter this out and it will continue to be an ongoing process just like the fight against spam and the fight against viruses goes. The main problem is communication.

Boiler plate email responses serve to anger advertisers.
Lack of transparency serves to place mistrust in the PPC network/advertiser relationship.
Ignoring or refusing independent 3rd party supportive auditing data frustrates advertisers.
Not communicating the steps or methodologies you have in place to protect advertisers creates doubt in advertisers.
Only providing summary activity information instead of detailed auditing only gives advertisers a piece of the picture.

And the humdinger for a lot of us:
Having a conflict of interest where the PPC network makes the sole decision in deciding whether certain activity was fraudulent or not and then keeping the money is just not acceptable.

As always I encourage all PPC advertisers to build their own PPC auditing tool, buy one from the number of services out there, or use VeriClix. Make sure you are auditing - consider it your insurance policy.
__________________
Jeff Martin - SEW Moderator
Vericlix
Jeff Martin is offline   Reply With Quote
Old 11-18-2006   #11
Discovery
Jerry Nordstrom
 
Discovery's Avatar
 
Join Date: Sep 2004
Location: San Diego
Posts: 1,029
Discovery has much to be proud ofDiscovery has much to be proud ofDiscovery has much to be proud ofDiscovery has much to be proud ofDiscovery has much to be proud ofDiscovery has much to be proud ofDiscovery has much to be proud ofDiscovery has much to be proud ofDiscovery has much to be proud ofDiscovery has much to be proud of
Hi Jeff,

I agree with your points. I just wonder why the SEs don’t simply allow every advertiser to exclude any domain/ip/tool bar from search or content. This would result in a peer reviewed network. Making this solution even more attractive is:

It allows advertisers and SE's to work together, this alone builds trust between us not skepticism.
A lot of actionable data could come from such a peer review. For example: What one may think is a click fraud site could simply be a poor performing site for their product or service. By cross sampling data across industries summary data could be provided to advertisers about the best performing sites for their industry. Additionally, if virtually all advertisers point to xyz website as a low conversion site the SE would have an easy time identifying who should be ousted from their network... and we could hold them to it.

Sure SE's need to have an aggressive protection plan to stop a ton of click fraud from hitting their system up-front. But they then need to supplement that by creating a system like a peer reviewed network to better understand the integrity of what does get through. Junk, Fraud, poor performing, excellent or other.

By implying that everything that goes past their defense is a legit click and not providing tools for advertisers to identify the source and opt out of certain publishers from running their ads, to me is where the blind eye comes in. (Google does have domain blocking, but only for the content network) The current "prove it was click fraud and we might refund you" haggle we must go through with a CSR seems to be intentionally difficult, time consuming and a deterrent for most to even bother. Especially when we have little data to prove our point, outside of our conversion ratio just sheet the bed over a 4 hour period and 1000 clicks.

Discovery
Discovery is offline   Reply With Quote
Old 11-18-2006   #12
andrewgoodman
 
andrewgoodman's Avatar
 
Join Date: Jun 2004
Location: Toronto
Posts: 637
andrewgoodman is a name known to allandrewgoodman is a name known to allandrewgoodman is a name known to allandrewgoodman is a name known to allandrewgoodman is a name known to allandrewgoodman is a name known to all
Quote:
Originally Posted by Discovery
Hi Jeff,

I agree with your points. I just wonder why the SEs don’t simply allow every advertiser to exclude any domain/ip/tool bar from search or content. This would result in a peer reviewed network.
I totally agree with the peer reviewed network concept. If designed properly it places the correct type of pressure on publishers, and allows advertisers to "shop smart".

That would be a fairly radically new concept, though. Did any of the ad networks or media planning firms in the banner world ever offer something like this? Usually it would be "try a buy, watch it fail," and by the time you'd been through $100,000, the publisher or the network could then adjust to kick out or restrict the participation of nonperforming Site X. That type of looseness doesn't fly in the new regime of online ad platforms like AdWords, where advertisers have been trained that they need to burn less money before making needed adjustments. But you can see why the networks are resistant to change. They are still better in some ways than their old competitors, the ad networks.

Probably the reason we are not there yet is that neither Google nor Yahoo yet have the right ratio of good to bad publishers, and are using only band-aid measures to choke off the bad ones.

Under site targeting, Google allows URL exclude, so that's a start. But sharing the "who's excluded" data in some way would be great for advertisers. Even by hiding it, though, they can put it to some good use by weeding out bad publishers.


Another problem is that with loose participation policies in the publisher networks, it still puts onus on advertisers to be constantly checking their stats for new, bad publishers. And wondering which are truly bad and which are just nonperforming for them. It almost feels like AdSense was initially flawed in its policies, so that advertisers face an ongoing struggle with "they can build 'em as fast as we can exclude 'em." That has been improving, but largely through algorithmic, band-aid measures such as smart pricing and more vigilance with discarding invalid clicks.

A more fundamental policy shift seems to be required, but at least at Yahoo that seems unlikely to happen. Yahoo is going to be targeting very small publishers for YPN and looking at the whole area only in the most optimistic of lights, rather than tempering that optimism with suspicion which seems warranted in light of rampant click fraud. This may leave advertisers with only one option: to opt out of content entirely. Let's hope not. Let's hope advertisers are given more sophisticated levers to run custom campaigns, including some of the content they want to enable.
andrewgoodman is offline   Reply With Quote
Old 11-20-2006   #13
Mel66
Oversees: Google AdWords
 
Mel66's Avatar
 
Join Date: Jul 2004
Location: Michigan
Posts: 775
Mel66 has much to be proud ofMel66 has much to be proud ofMel66 has much to be proud ofMel66 has much to be proud ofMel66 has much to be proud ofMel66 has much to be proud ofMel66 has much to be proud ofMel66 has much to be proud ofMel66 has much to be proud of
Very good points Andrew. I would like to add a couple comments.

First off, Yahoo needs to set up their content network so advertisers can use different tracking URLs in content than they do in search. Currently that is not possible - ads must be *exactly* the same in search and content, right down to the destination URL. That, among other reasons, is why we don't use Yahoo content - no way to track conversions separately (unless you use Yahoo's tool, which we are not interested in). There are some ads I'd like to run in content b/c I think they did well there, but it's not worth the risk if I can't track results. So Yahoo is losing money on this. I don't know if this is different in Panama or not - we are still on the old DTC system.

Secondly, there is the issue of low-quality partners in search networks. Google has this problem too, but not nearly to the extent that Yahoo (and Ask, for that matter) does. From where I sit, it looks like all you need to do to be a Yahoo search partner is put up an MFA site with a search box at the top and a few "related links," and voila! you're a search partner. Check out "toseeka" if you want an example. Nothing but Yahoo search ads on every darn page - no original content whatsoever. Sites like this are very low quality, yet they somehow qualify as search partners and therefore as a Yahoo search advertiser, you're stuck with them and have to factor their useless traffic into your CPCs. Site exclusion in the content network won't stop traffic from these types of sites. The engines (all of them, Google included) need to offer site exclusion across their entire network, not just in content. This would create a true peer review system like Discovery is talking about, and it would be very clear very quickly to the engines who their poor performers are.

Also as Discovery said, some of the partner sites do well in certain verticals but not in others. We actually get decent conversions from many domain parking sites. Others do not. However, there are some sites that seem to be universally low quality and those are the ones that we should be allowed to exclude across the board.

Melissa
Mel66 is offline   Reply With Quote
Old 11-20-2006   #14
briggidere
Member
 
Join Date: Oct 2006
Location: LINCS, UK
Posts: 22
briggidere is on a distinguished road
i think one of the important things is the fact that they are working on things to detect and prevent click fraud, but how much information can you provide people with to let them know what you are doing.

If we all knew the prevention techniques, then we could get around them. They need to keep these things to themselves otherwise it will compromise their situation.
briggidere is offline   Reply With Quote
Old 11-20-2006   #15
Discovery
Jerry Nordstrom
 
Discovery's Avatar
 
Join Date: Sep 2004
Location: San Diego
Posts: 1,029
Discovery has much to be proud ofDiscovery has much to be proud ofDiscovery has much to be proud ofDiscovery has much to be proud ofDiscovery has much to be proud ofDiscovery has much to be proud ofDiscovery has much to be proud ofDiscovery has much to be proud ofDiscovery has much to be proud ofDiscovery has much to be proud of
Hi briggidere,

Good to see a fairly new member from across the pond posting.

In my view I the SE's don’t have to tell us anything about the Herculean battles they fight against click fraud everyday. Like M.I.B we probably don't want to know the truth... Does YSM have a little strobe wand? I can't remember.. Anyhow, I believe that a peer reviewed network would go a long way in eliminating the poor performing publishers regardless if they are fraudulent, have irrelevant content or simply have bad visitors. The bottom line is that the publisher network will be largely managed by the best performance indicator possible... the goal conversions set by advertisers.

The simplicity is beautiful.

SE's need not disclose their secret investigative ways and can hand off some of the work load to advertisers.

Advertisers would determine according to their advertising goals which is a good partner for them.

Publishers with good content/traffic/goal conversions will be rewarded with more advertisers competing for space at higher prices.

There are drawbacks, if we had total transparency into the network partners and we could exclude any domain/ip/toolbar/whatever, then what level of "click fraud junk clicks" must we as advertisers accept? All? After all the onus of policing our personal network would now be pretty squarely on us.

Discovery
Discovery is offline   Reply With Quote
Old 11-20-2006   #16
projectphp
What The World, Needs Now, Is Love, Sweet Love
 
Join Date: Jun 2004
Location: Sydney, Australia
Posts: 449
projectphp is a splendid one to beholdprojectphp is a splendid one to beholdprojectphp is a splendid one to beholdprojectphp is a splendid one to beholdprojectphp is a splendid one to beholdprojectphp is a splendid one to beholdprojectphp is a splendid one to behold
I think we need to seperate the various options.

There is search split into (at least) three:
1. The top tier - AdWords on Google, MSN ads on MSN, Yahoo on Yahoo.
2. The second tier - Ads on good third party search networks. Google on Ask and AOL.
3. Search everywhere else - AdWords on Joe-Blow-Search.com

And then there is the content network, with probably a lot more layers between top and bottom, but similarly aligned.

The click fraud each element sees is vastly different, and needs different tools and solutions.

It adds a lot of layers to the process, but being able to restrict ads based upon multiple criteria would be tremexcellent. Similarly, splitting up each group would be very helpful.

For AdSense, IMHO there should be multiple cats of sites. Lets use three as an example:
1. Sites that everyone knows, say NY Times.
2. Sites deemed to be pretty trustworthy, say sites that have performed well in the program for a period of time.
3. Everything else.

Being able to target specific levels would reduce the need to find bad sites, would pay a premium to sites that are run well, and would be opt in, advertisers know the risk rather than opt out, advertisers have to hope and track down the bad ones themselves.

Ditto search, beacuse at the moment, two options on Google (everywhere or just Google) doesn't really cut it.

The USP of online marketing is its ability to be highly measured. The more that can be done to segregate the data and measure separately, instead of grouping multiple similar but different streams together, the better for all concerned.
projectphp is offline   Reply With Quote
Old 11-26-2006   #17
Torq
Member
 
Join Date: Nov 2006
Posts: 5
Torq is on a distinguished road
PPC frustration

I work for a large ad network (www.adonnetwork.com). I have used every combination of our sources only to be discouraged by discrepencies in numbers. I have finally discovered a combination that works with Miva. Has anyone else experienced good or bad ppc campaign results using Miva? They seem to be very effective with the campaigns I am running.
Torq is offline   Reply With Quote
Reply


Currently Active Users Viewing This Thread: 1 (0 members and 1 guests)
 
Thread Tools

Posting Rules
You may not post new threads
You may not post replies
You may not post attachments
You may not edit your posts

vB code is On
Smilies are On
[IMG] code is On
HTML code is Off