Hacker News Reader

Comment by birken | Hacker News Reader

You think Google can't detect this? Enough advertisers have conversion tracking that Google can detect if a person is clicking a zillion ads and never actually converting, in addition to the other multitude of ways they detect fraud. I'm sure in the battle between fake traffic producers and Google there are probably tricky ways you can fool Google, but a person clicking every ad is not one of them.

Also... this doesn't actually hurt Google. If the percentage of fake traffic goes up, the cost per click will just go down over the long term. And what, advertisers are going to lose faith? Give me a break. It doesn't matter if the traffic is 0% fake or 90% fake, all that matters is if the traffic is making you a profit or not.

View on HN

> It doesn't matter if the traffic is 0% fake or 90% fake, all that matters is if the traffic is making you a profit or not.

Only under the assumption that no competing, superior technologies exists. The worse the efficiency of the ad-model, the more attractive the alternatives will be.

Criticizing Microsoft over delivering an inferior user experience twenty years ago is in the same category as the critique against Google: as long as they have complete control over the market, it doesn’t matter at all. But when competing, superior technologies start appearing, it becomes their downfall. And the worse the user (advertisers, in case of Google) experience, the greater the incentive for competing technologies to appear.

What alternatives? The key product here is Google search which is what attracts the consumers. The advertisers are dealing with Adwords not because Adwords is the greatest platform in history, but because Adwords is the platform where the consumers are. Ad fraud, no matter how massive, has no effect on the consumer side of Google search.

Google has a strong incentive to fix this because any money going to scammers is theoretically coming out of their own pockets as advertisers lower their CPCs to adjust for lower conversion rates, but it isn't an existential threat to the marketplace. Cut out the scammers and the CPCs adjust back upwards as conversion rates increase.

And again, Google has the tools to find the scammers. It's a game of whack-a-mole for sure, but at the end of the day the higher the level of fraud on the marketplace the easier it is to catch, so the scammers are somewhat limited in how much damage they do.

> It's a game of whack-a-mole for sure, but at the end of the day the higher the level of fraud on the marketplace the easier it is to catch, so the scammers are somewhat limited in how much damage they do.

This is a perfect summary of what is going on.

In this case I trust Google because their incentives are aligned with advertisers', and they certainly have both data (from Google Analytics and other sites) and means to distinguish between a human and a bot.

> In this case I trust Google because their incentives are aligned with advertisers

Is it?

The better the overall search algorithm, the less the need to hack around it by paying for placement..

Some level of 'bad results' helps ensure that ads are needed in the first place, and junk sites provide a ready source of bad results..

Now, when junk sites cause you to lose revenue, the equation gets unbalanced, but that is not to say having the variable equate to 0 is good either, because the revenue would lessen..

Obviously it can't simply allow the fraud because that would make google fraudulent itself, but at the same time they cant just outright ban crappy sites, since there are so many that are only marginally crappy but otherwise legitimate, and could only be unwilling participants in someone else's dishonest scheme..

I would argue that google is attempting to maximize a set of equations here, and since at some point there is a grey area between 'crappy scammish sites' and 'full on fraud', google has some leeway to what extent it polices the grey area, and, further, to some extent can do this at advertisers expense so long as 'the house wins' to borrow gamboling terminology..

> The better the overall search algorithm, the less the need to hack around it by paying for placement..

All the things you need are not necessarily known to you in advance. It’s entirely possible that someone has created a new product that no one knows about, and allowing these companies to pay for visibility allows you to discover these ideas that you might end up finding useful. The idea is that when advertisers pay for placement, they better have a useful product (that can recoup the cost), or else they’ll be throwing money out the window.

true, but which likely-related things are shown is also a decision for the algorithm to make, and one which can be skewed in a variety of positive or negative ways

> What alternatives?

I’m not claiming any alternative exists right now. I’m saying that the less efficient Google is at converting dollars paid by advertisers into visitors, the greater the pressure from future (unknown) technologies that eradicate the inherent inefficiency in Google’s model.

If I knew exactly how to capture Google’s ad revenue, I would be starting a company rather than writing comments on HN.

Actually Google do not have a complete market control. I managed a 100M+ ecommerce startup and our largest traffic source is Facebook and Instagram.

> You think Google can't detect this?

If they randomize it, no.

looks like they just need to start clicking a lower random percentage.

Reply on news.ycombinator.com