The Strategic Implications Of Header Bidding

aripaparo-updated2The Sell Sider” is a column written for the sell side of the digital media community.

Today’s column is written by Ari Paparo, CEO at Beeswax.

This is the Year of Header Bidding.

OK, maybe it’s actually the Year of Ad Blocking, but header bidding is the second most talked-about topic in advertising, especially among ad tech insiders.

I’ve probably spoken to 20 publishers and ad tech vendors about the emergence of header bidding, and the common thread is the urgency and certainty around deployment. This isn’t just a fad, it’s a full-blown movement.

So let’s step back and speculate on the implications of widespread header bidding on the ecosystem’s various companies and entities. It may seem like a highly technical and obscure development, but it could have huge ripple effects on the entire digital advertising landscape.

Sell-Side Consolidation

Header bidding empowers sell-side players to get access to inventory at the same basis as juggernaut Google. Ironically, widespread usage of this technique will cause sharp consolidation in the landscape of supply-side platforms (SSPs) and exchanges. Let me explain.

While predictions of consolidation in ad tech have grown in volume, the industry lacks a forcing function to drive smaller players out of business. In manufacturing industries, for example, economies of scale can force smaller players out as their production costs become uncompetitive. In software businesses, there are often network effects that drive monopolies or duopolies.

But in ad tech, access to demand and supply largely remains unconstrained, allowing ad networks and related businesses to survive and thrive with little scale. All you need to build an ad network is a phone and a tag.

Header bidding has the potential to create a large barrier to entry for the sell side of the business, thereby driving consolidation. Talking to publishers that have or are considering implementing header bidding, the vital question they ask is which vendors to choose for the header. This decision, and the implied limit on the number of participants, is driven by the realization that page performance, or latency, is directly affected by the number and quality of header tags. This limit is a real, hard limit, and contrasts quite clearly with the very loose restrictions on ad tag deployment when implemented in a waterfall.

Suppose, then, that the average publisher is only willing to put four tags in its header, which seems to be the rough consensus among my contacts. The first two slots are going to Criteo and Amazon A9, both of which have unique, high-priced demand. If you’re a sell-side business, I bet you’re getting a bit nervous about securing your position in one of the other two slots.

Expand this logic to the entirety of the programmatic world, and there should be a forcing function – a scarcity – that makes larger and more effective sell-side companies increasingly able to provide yield to publishers, and correspondingly, makes smaller and less effective companies less able to survive.

Buy-Side Impact

The impact on the buy side is a bit harder to predict. In theory, a demand-side platform (DSP) or other buy-side platform should be agnostic to the implementation of header bidding. Without header bidding, a given bid would have a larger-than-average chance of winning on Google’s AdX. With header bidding, that same bid could be won on a different exchange. In either case the market dynamics and pricing should be similar.

This assumption breaks down in a couple of scenarios.

First, imagine that any single buy-side business, such as Criteo, has the scale to demand its spot in the header, while smaller buy-side businesses cannot aggregate enough demand to be worth a publisher’s attention. This puts the bids from a non-header source at a systematic disadvantage for the same inventory since a DSP bid going through a third-party SSP will need to pay the 10% to 25% transaction costs before competing for the impression. In reality, this is currently the situation for retargeting campaigns competing with Criteo.

Second, imagine that the sell-side players are competing for position in the header (see above). One way to get that position is to assure that more demand comes through your pipes at higher prices than the competition.

This could and should drive more companies to acquire or build buy-side businesses, which is what we saw with Rubicon’s purchase of Chango, AppNexus’ general buy and sell positioning and Twitter’s acquisition of TellApart. The implication for standalone buy-side businesses could be negative over time as access to quality inventory is redirected to in-house demand over exchange-traded demand.


The question everyone wants to know: What will Google do?

Make no mistake, header bidding is entirely an attack on Google’s dynamic allocation, the feature that gives it an unfair (or fair, depending on how you look at it) advantage when selling inventory from DoubleClick for Publishers (DFP). And don’t think for a minute that Google is happy about the barbarians at the gate. No. No. No.

There are three approaches Google can take: fight, innovate or change the conversation.

Google won't try to fight header bidding directly. That would not be in the interest of its publisher clients and could cause anti-competitive concerns. Turning on a policy in DFP that doesn’t allow headers would play right into the competition’s hands.

Innovating is always Google’s preferred path. Google’s continued leadership in programmatic direct deals is one way to make the yield question less important since quality supply will skip the RTB channels. Continuing to give advantages to its DoubleClick Bid Manager platform when buying AdX inventory is another tactic. Heck, it could even build its own header implementation and “optimize” the results to the benefit of publishers, all in the name of reducing latency.

Another glimpse of the Google solution can be seen in its recent announcements around Accelerated Mobile Pages. This technology is built to allow mobile web content to load much faster, with one … cough … unforeseen … cough … side effect … cough … being that you can’t run arbitrary JavaScript on the page. This kills header bidding dead, and, according to Google, is the future.

Publishers, Agencies And Everyone Else

As with most things in ad tech, the actual principals to the transaction are the least involved in these politics and mechanizations. For publishers, header bidding means a bit more yield and a bit less dependence on Google, both of which are good things.

For advertisers and agencies, there should be little to no effect. They should still buy media based on results, and choose vendors based on transparency and capabilities.

The big winners here will be consultants and industry gadabouts like myself, who will take advantage of this golden opportunity to add another obscure phrase to our slides and reinforce our self-created images as the only ones who understand this confusing marketplace. Same as usual.

Follow Ari Paparo (@aripap), Beeswax (@BeeswaxIO) and AdExchanger (@adexchanger) on Twitter.


  1. Dina Srinivasan

    The proliferation of ad tech hasn't solved Publishers' inherent predicament of life after print advertising. The ad tech industry has won this battle while showering them with complexity. A complexity which with all its moving parts (and gray behavior) has hampered the consumer experience with incredibly frustrating load times. Thus the rise of the number one topic you mention - 'ad blocking' - consumers saying no to the whole thing altogether. Will header bidding change this macro market trend in any meaningful way? I predict not. In our Silicon Valley driven economy, the user comes first; and markets that forget this will pay a steep price when technology catches up and consumers can force the pendulum back into their court (ad blocking). Smart publishers - like the Times - have foresight that life is about subscription dollars + meaningful ad experiences (and I'm not talking about paltry trends in Native). But yes, everyone in ad tech, will be able to ride out the ignorance and lag in market response in the meantime. And, may in fact be the long term winners in a shrunken ad market that is simply not worth the time of Publishers.

  2. Steve White

    Since most publishers continue to use DFP when entering the world of header bidding, using cumbersome array of string-matched line items, Google keeps control over the game. They can pick whatever is best for Google and force the market to bend its way. Publishers should strive to implement header bidding using more up-to-date techniques (e.g. dedicated header-bidding setup outside of DFP) to keep control in their hands.

  3. Alex Bozowski

    The "hard limit" on the number of header-bidding calls can be alleviated by server-to-server integrations, which are now offered by many bidders.
    This would reduce the pressure described in the article and again allow small player to keep in the game.

  4. "Hard limit" can be also removed on client side. For example, in Criteo implementation their script loads asynchronously on first page load, determines if Criteo is going to buy an impression and writes flags to cookie. Then on the second page load this variables are passed from cokie to ad server and then script again loads asynchronously to refresh cookie.
    This mechanism is not 100% optimal (creates 1 page load gap between bid and imperssion), but removes page load latency.


Add a comment

XHTML: You can use these tags: <a href="" title=""> <abbr title=""> <acronym title=""> <b> <blockquote cite=""> <cite> <code> <del datetime=""> <em> <i> <q cite=""> <s> <strike> <strong>