Click fraud is one thing. Impressions fraud anyone?

Ever heard of impressions fraud

[charliemunger]

I have noticed a *massive* spike on my biggest keyword on a few days in the last few weeks, a spike unique to that keyword, and to googles search alone (I am talking a six hour 2000% increase in impressions, with no difference whatsoever in clicks, and no change on content)

Is it possible an angry competitor is pausing his campaigns, spamming the impressions (hitting refresh 3000 times lol), and then bringing them back when everyone elses CTR and quality score has just been beaten down.

Sounds a little odd, I know, but how else you explain a 2000% increase in 2 x 6 hour periods on one keyword with no change in CTR is beyond me

Have never heard of impression fraud before, but there you go!

[gregbo]

Click fraud is getting press now. IMO impression fraud is also a concern. As a general rule, I think any method of charging or assessing prices based on raw traffic is a bad business model, at this stage of Internet protocol development and architecture.

[DamonHD]

Hi,

All this fuss about PPC fraud: it's not good but it's not limited to PPC and it's certainly NOT new.

Famous quote: "We know that half of the money that we spend on advertising is wasted: we just don't know which half."

In my area there are two free local papers. Most weeks only one turns up, often for months at a time, usually, so far as I can tell, because the oiks paid to deliver it have taken the money and dumped the papers over a hedge. (I pity the owner of the hedge too. B^>)

That is "impression" fraud, though the ad network (ie the paper) does not necessarily know about it.

There is wastage/fraud/snafus in ****ALL**** advertising, and at least in PPC advertising you have a chance of working out in almost real time what is working and what is not so that you can take your $$$ elsewhere.

Rgds

Damon

[CrimsonGirl]

^That's all true, but one big selling proposition of web advertising is its measurability. Google, Overture, and those of us in the Internet marketing business have been saying for years: with print ads you waste half your money, but you don't know which half; with web ads, you can tell what works and what doesn't. Any type of fraud undermines the credibility of this statement and leads to less advertising.

Impression fraud is worth fighting if we have any tools to do so.

[DamonHD]

Hi CrimsonGirl,

Absolutely agree that we should minimise impression fraud.

All I'm saying is that treating it like a new plague is overblown, IMHO.

Rgds

Damon

[gregbo]

There is wastage/fraud/snafus in ****ALL**** advertising, and at least in PPC advertising you have a chance of working out in almost real time what is working and what is not so that you can take your $$$ elsewhere.

Hmmm ... I'm not so sure I agree. If this were true, a SE or ad network could just use this technology to automatically weed out the fraudulent clicks.

[venrooy]

It's impossible to combat impression fraud, because it happens on googles site and not our own.

We should have access to every ip that sees our ad, or be able to put a tracking script on every page that our ad shows up on.

Right now impression fraud is making google money, and there's absolutely no way of proving it actually exists without their help.

[jezz037uk]

Has anybody had experience of using one of the many clickfraud prevention services available by subscription? Like many, I suspect that a percentage of clicks are not necessarily fraud, but at least "click abuse" and would welcome a way of monitoring / limiting these. Looking at the available services is confusing as they each tell a great story. Insider knowledge / experiences very welcome!

(Experience is something you get *just* after you need it!)

[deep_alley]

Had heard someone else complaining of impression fraud long ago. It is pretty rare but I dont think we can just forget about it because clicking the refresh button a 1000+ times is too ridiculous a thought. On google a competitor doing this makes a lot of sense because google gives so much weightage to CTR and other 'relevancy factors'.
There are reasons your impressions can go up (change in position, fewer competitors, more content ads etc) but I cannot imagine these reasons contributing to such a substantail increase as you have mentioned charliemunger.

[Melech]

I've seen some very generic keywords in my list (that show only in position 100 or higher) spike on a few days. Normally they get one impression a day, if that, but for a few days they get 10 or 15 impressions, which is odd given that you would have to go to the tenth or so page of search results to get them to show.

[mike_ppc]

Not necessarily. If you set your preferences to see 100 results and then search your KW, you will get an impression if you are in the first 100 results, even if your ad is down on the page.

[Melech]

Good point, and so I suppose it could be someone researching the field and grabbing 100-result pages of the common keywords. But I meant that the average (default) user would need to chase a keyword result pretty far to show my ad, and thus it's abnormal behaviour. But you're right, it may not be malicious.