There is a legitimate reason to oppose header bidding: it results in higher latency than using traditional ad services. Ads served at lower latency tend to get more clicks, and so they have more value. Unfortunately, that means that header bidding results in less money for the website.

...or so Google says, paternalistically. Yet for some reason this technique is used by about two thirds of publishers. I guess they're just throwing their money away? Somehow they fail to realize that Google knows best?

There are good reasons to not use Google ads if you are a news site: you are competing with Google for being a news provider. Even if you would make more without header bidding, you might not want to use solutions like Google ads since that means supporting someone who is competing with you. I personally think header bidding is the next good evolution of the ad markets (and the ad market seems to agree), but there is a cost to it.

How does Google compete with news providers?

How is a service presenting links to a publication competing with that publication?

Because that completely replaces the front page of the news site, a page people normally visit just to find out if there are any articles worth reading. I can see how that leads to less ad impressions on the news site itself if people only visit it when they found an interesting article (and just that article) on google news.

To be fair most newspaper websites form pages I have seen have been eyebleedingly bad; stuffed to the gills with anti-patterns, obnoxious amounts of ads and some of the worst user interface designs I have seen on the internet. is it any wonder people use google instead of newspapers own sites?

For a lot of people, the headlines are all they really want to read plus maybe a snippet of a sentence or two. News outlets that expect to sell those eyeballs lose out and consider this unfair competition from Google.

Google, for its part, has been perfectly happy to remove publications from Google News at their request. Or leave entire countries.

A competitor you can shut down at any time by asking isn't much of a competitor.

I agree with you that it's an exceptionally self-serving way of approaching the subject.

Client side header bidding requires an HTTP request / response for each bidder, followed by an HTTP request to the primary ad server (very often, Google Ad Manager), that usually runs its own ad selection process (against lime items from direct sales, and programmatic demand sources such as Google Ad Exchange). Server side header bidding offsets the call to each bidder to the server side and it somewhat improves latency, but it still takes some time to complete the auction, and there's still need for an extra HTTP request / response to the header bidding server.

> Unfortunately, that means that header bidding results in less money for the website.

Can they make up the difference in not getting ripped off by Google?

Yes, this is why many larger media sites opt to use header bidding. The decrease in CTR (click through rate) is offset by increased demand for each ad unit, driving up CPM (effectively the price of the ad).

The downside of this for websites is more middlemen extracting away value - but it’s usually still a net gain, otherwise publishers just wouldn’t use header bidding.

(I’ve worked in several large media companies and have implemented header bidding solutions for them in the past)

Probably not, since I doubt the ad network sends that much more of the margin it makes without Google back to the website.

You'd be surprised. Every time I've observed a switch to header bidding, revenue goes up like 50% or so. And I've heard the same from many other people. I don't think I've ever heard someone say it didn't improve revenue. And yes, that increase is after the new middleman takes his typical 20% cut!

I feel like a late loading ad that pops under peoples mouse gets more clicks than any other ad does. So are we sure about your statement that low latency leads to more clicks?

I'd be surprised if they don't have async mechanisms.

It's async, but since it's selecting the ad it still has to run before ad can be shown.

Could the latency problem be solved?

It's pretty fundamental: header bidding involves an extra hop through the client. It's encouraging that other commenters are saying that they are adopting header bidding despite this.

> Google's dominance of the ad market

It isn't just dominance. I certainly don't love advertisers, but this is illegal monopolistic price setting. This should be ample evidence for the DOJ to sternly penalize (or break up) Google and Facebook.

Google controls too much surface area on the web and mobile, and they're using it in harmful ways. They've been caught colluding, price setting, and wage fixing on multiple occasions. They're using their disproportionate power to alter the entire market to fit their needs.

How is this any different from Ma Bell? And given the pervasiveness of technology now versus back then, how is it not worse?

When we helped brief on this investigation, we postulated that header bidding might serve as a heuristic for the economic damage caused by the Google adtech monopoly. In time, mostly through private lawsuits, I expect some version of this to bear out.

Who is 'we' in this case?

Header bidding also means that websites blast your page visits to every provider that might want to bid on it, instead of "only" to the ad exchange and the winner. Just a massive surveillance spout to anyone at all.

Just because Google is against something doesn't mean that it is good for us as consumers.

This has more details about Google’s move to unified auctions and first bid pricing: https://digiday.com/media/what-is-a-unified-ad-auction/