Hacker News Reader

Comment by aresant | Hacker News Reader

Editorializing:

Paid clicks on Google properties +43% - "yaaay we're driving more paid search!"

Cost-per-click on Google properties -16% - "ouch our advertisers are seeing less value on these add'l clicks!"

To me the second part of this is going to be most interesting to watch - if the clicks they are onboarding are lower quality, this is going to be a net negative.

View on HN

It doesn't work like that. Lower CPC does not mean lower value. It just means Google pushing into new eco-systems where CPCs are different from desktop search (e.g. mobile, YT, new publishers/partners) and it may take some time to reach equilibrium among the various actors depending on the ROI of the medium for the respective actors.

Precisely that. Change in Google's CPCs is a great example of Simpson's Paradox. CPCs on every property separately are growing, but aggregate is declining due to much faster growth in areas where CPCs started low

Also new growth will come from developing countries which will have lower CPC than the US.

If the half of the world without an Internet connection gets online and starts clicking on ads, their clicks are probably going to be less valuable than the existing rich world clicks. I don't see how that is a net negative for anyone, though.

Right, that would bring the average down but the sum up, which seems to be happening?

Well the use of ads is negative for everyone.

Google is selling ad inventory where it didn't previously. Google Shopping is slowly rolling out across the world, initially in new markets there is a lot less competition than established ones, so the clicks can be had a lot cheaper.

For example we are probably paying half for the same quality Google Shopping clicks in New Zealand than we are in Australia or the US. This won't last.

Your premise is wrong.

Watching an ad on Youtube is considered a click, but advertisers pay less than search ads.

Cost per click just reflects a change in where clicks are from.

How can it be a net negative? Net change in value = 1.43 * 0.84 = +20%

I think his point was that if CPC decreases monotonically forever, then eventually revenue will have to suffer regardless of volume growth. As others have mentioned, this doesn't take into account the mix of new ecosystems, though it's noteworthy that Google doesn't break down CPC by segment so we could see trends for desktop/mobile/YouTube/etc/

But if you look at trend on each individual property separately, CPCs are not declining anywhere. It's just effect of Simpson's Paradox on total number

I tend to agree with this analysis. Not only is the cost per click going down, but the amount of money they are paying out to get traffic is now at a staggering $7B a year run rate. Paying for more traffic and getting less in revenue per click shows up as a higher 'cost per revenue $' value which has gone up from 38% to 41%.

I have felt like the easy pickings have been behind them for a while but given enough money you can buy a lot of traffic. Not great traffic mind you but a lot of it.

i believe its that the supply of sold online ad clicks has grown significantly (fb) and therefore marketers have more bidders resulting in lower costs (not just for google but the whole market)

The problem is that institutional investors seem to really care about CPC. Even when RPM is up due to greater CTR, CPC drops have allegedly affected their stock price.

I can understand the usefulness of search ads and ads on maps, but if I see an ad on youtube it is unlikely I am going to go buy a product.

..and yet, TV advertising is still a huge business

YouTube can target ads much better than TV, and with TrueView ads the advertiser only pays if the viewer doesn't skip, so more of the revenue is going to be coming from ads that mutually add value (at least that's the theory)

I don't get TV ads either. What is the effectiveness of it?

Coke tastes better if you know it's a Coke. That's what a top-notch advertising department does for you. Well, that and early memories of drinking Coca-Cola on a warm summer night. It's the real thing, baby.

Put crudely, name and brand recognition. People view familiar things more favorably than unfamiliar things. If you've heard of it, you are more likely to buy it. If it has been on TV a lot, you've heard of it -- even if you didn't like the ad.

There's a correlation between TV advertising and increased sales over the long term.

http://adage.com/article/media/study-affirms-long-term-sales...

(The favorability thing has been lab-tested by showing people subliminal adverts, among other things.)

Building common knowledge of what choosing particular brands means. If you know everyone has seen those ads with some folks relaxing on a beach with a Corona, then you know that people will think you are laid-back when you bring Corona to a party.

This is why super bowl ads are more expensive per viewer, because the audience can expect their social group to have also seen the ads.

costs could have been driven down by competition not so much by lack of demand or falling quality.

Reply on news.ycombinator.com