ExchangeWire recently attended PubMatic's Publisher Academy on the topic of 'Is Header Bidding Dead?', where Kirk McDonald, president, PubMatic, argued not, but that it is one stage in an evolution of ad tech and that the important goal is control over ad decisioning.

When you think about header bidding, it is important to ask yourself what the context is behind what you do and why you do it. We seem to always forget to talk about the why, with businesses in our industry existing ultimately because of the people, because of the consumer.

What's at stake?

Consumer attention and trust. We are in constant competition for consumer attention, now more than ever before and this is becoming more rapid as consumers move to new platforms; which is a challenge for both the publisher and the advertiser.

Defining a publisher as anyone who is a creator of content or services that are intended for a consumer and reliant on ads bought or sold to make the business model work, McDonald said that they need to work to find the smartest way possible to be intimately connected to their consumers.

With his years of knowledge and experience, McDonald referred to the simpler times of advertising; not in an attempt at romanticism, but instead, to show that not much has changed: "What we do today isn't all that new, it's just being done with new technologies."

Referring to the linear, sequential means of advertising in days gone by, which were based on a stimulus response, McDonald argued that digital has changed the flow of information: "Acceleration allows us to do things differently and, most importantly, we can personalise our messaging." This is an important development, as it allows us to gain and maintain consumer attention and trust; and not taking these two elements seriously means that the context is lost, regardless of the monetisation tactics you employ.

Is header bidding as strategic as we imagine?

No. According to McDonald, header bidding is a tactic, not a strategy; and the more important discussion around it is the strategic control of ad decisioning – are publishers really in control of ad decisioning and does header bidding help them to regain that power?

From creative management influenced by an editorial team, to business monetisation decided in a boardroom discussion, to ad serving and digital content delivery through to dynamic ad insertion – technology moved faster than our thinking, and there was no ability to keep up with real-time ad delivery. There has been an ongoing battle over ad decisioning control, and it has become strategic, no longer sitting within ad ops.

"We have to shift our thinking to the longer term", qualified McDonald, "and ask who is in control of ad decisioning, balancing the user experience against our need for monetisation, and whether we are best positioned, through technology and partnerships, to achieve that".

Why is the waterfall flawed?

McDonald stated that the waterfall has been created using static rules for a dynamic environment, driving potential lost revenue at each stage. For publishers, the decisions on which demand source can secure an ad impression is done through the creation of priorities – impression delivery is the same. The legacy, static waterfall approaches each impression as equal; it doesn't treat the impression's delivery differently, the result of which is that it creates rules based on where the impression delivers – not what the impression is. Therefore, you make the rule based on pages, and not necessarily any characteristics that define one impression as unique from another. Thus, you have a potential loss in either efficiency or efficacy of the ad delivered, resulting in lost revenue, as the impression is not monetised for its highest value, nor is it bought for its highest ROI. As McDonald said, buyers have found ways around this; which, for the most part, involves buying a secondary asset of publisher inventory, where said inventory isn't made fully available to the buyer.

This exists because sequential execution doesn't deal with the dynamic information that enters our current systems. The regulation of decisioning is one of most important aspects of what makes the flaw in the waterfall. The rules don't change through the process; and header bidding introduces a new dynamic, as it forces the ad server to move away from the outdated, inefficient, rules-based system and moves the ad decisioning algorithm away from the ad server, allowing it to be hosted on the publisher page.

What it means to retain ad-decisioning control

"In the waterfall system", explained McDonald, "there is a winner and there are losers – and the winner is the ad server". McDonald related that the ad server knows the publisher's lost revenue, the publisher opportunity, and also what the advertiser has lost in terms of inventory.

In comes header bidding, and there is so much promise and excitement around it, away from the legacy waterfall process: it increases competition at impression level, reduces passback cost, improves yield, lowers dependency on AdX (which, according to McDonald, publishers realise has to be part of their long-term strategy) and allows stronger brand control than in older waterfall set ups.

It's an opportunity to take back the control of ad decisioning, as it restores control, democratises the request at the top of the page and creates competition among all demand sources.

However, amongst the industry joy, McDonald was keen to point out something important; that header bidding is like the cassette tape: "There was a moment in time that cassette tapes made people more portable; the Sony Walkman was revolutionary. We may have thought that was it, but it wasn't, because after that we had the CD, then the MP3 and now music streaming. Who would have imagined this?"

McDonald likened the moment of celebrating header bidding to the wrapper tag, saying we should all enjoy this moment, but not get caught up in it, as it's still fundamentally about ad-decisioning. "Header bidding is not a complete solution", argued McDonald. "The solution lies in us understanding strategic implications." He asked the audience to consider that, whether buying or selling, 85% of incremental digital ad spend goes intentionally to Google or Facebook, who both have their own ad servers. Google launching EBDA (Enhanced Bidding into Dynamic Allocation) means they are 'supporting you' in your control of ad decisioning; but, if that's the case, are you in control?

What's the future of ad decisioning?

According to McDonald, if the header bidder is where we think we've perfected the wheel, we're a long way away. The future of ad decisioning is a holistic monetisation platform, which would allow guaranteed sales to work alongside indirect buys; auction-based buys to work alongside all IO-based buys; and to be able to optimise across multiple environments, as mobile, display and video are all merging, thanks to the consumer. This needs to exist alongside holistic inventory, with comprehensive reporting and analytics, as well as flexible integration partners. No small task, then.