Exploring ad tech and header bidding

Tun Khine
5 min readOct 10, 2019

--

Earlier this year an internet advertising report from IAB and PwC indicated that digital ad spending in the United States surpassed $100 billion. With an industry this large I wanted to learn and provide an overview of the tech powering the advertising industry and how ad tech dollars are used to grab your attention.

What is Ad tech and header bidding?

The birth of Ad tech takes us back to 1994, where AT&T ran the first true banner ad on Wired Magazine, which was then HotWired.com. Ad tech is the umbrella term for the software and tools that help agencies and brands target, deliver, and analyze their digital advertising efforts. This software and tools help advertisers make better use of their budgets to deliver the right content at the right time to the right consumers.

In the past, if advertisers wanted to show their ads on a particular publisher’s site, a tedious process of back and forth emails, calls, and negotiations, would need to take place. Typically purchases used a cost-per-thousand impressions (CPM) model, meaning they are bought in bulk and so the same ads were being shown to everyone regardless of their demographic and behavior. As the number of advertisers and publishers grew, so did the problem of communication and tracking these back and forth negotiations.

Enter Programmatic ad tech

With Programmatic ad tech tools, traditional human negotiations were replaced with machine learning and AI-optimisation. The goal being to increase efficiency and transparency to both the advertiser and the publisher. Often times this is done through real-time auctions where ads are bought at the same as a visitor loads a website.

The first instances of these auctions were known as real-time bidding with the negotiations and transactions happening around 100 milliseconds as a webpage loads for you!

Let’s say you visit a website that sells Star Wars memorabilia, but you don’t make a purchase. Later you visit your favorite news site, and suddenly there’s ads about Star Wars memorabilia everywhere! That’s thanks to RTB equipped sites that take visitor data and information on the website, then triggers a request sent to an ad-exchange. This information is then matched against available advertisers and a real-time auction takes place between those advertisers that match certain criteria. In essence the company selling memorabilia has simply stated “I want to show my ads on these websites — but only to visitors who previously visited my site and didn’t make a purchase.”

What is Header Bidding?

Before header bidding was introduced, ad space was auctioned off and delivered only once the ad placements began to load on a web page. In other words, once a web page’s ad placements began to load, the publisher’s ad server was called up and the so-called waterfall of ad serving would begin. In the waterfall approach, each exchange partner has a corresponding line item in a publisher’s ad server, and each sits at a different priority. If the first exchange partner in the sequence doesn’t make a purchase, it cascades down to the next exchange and to the next, until someone buys it, or until it ultimately gets passed to Google AdSense.

AdProfs

Header bidding is an additional auction that takes place outside of the ad server, in the header of a web page, loading before anything else on the page. The header typically contains metadata about the page and calls scripts used for formatting the style of the page, tracking, and so on. Because of this, it’s an ideal area to conduct a new auction. With header bidding, before the publisher’s ad server is even called, every impression is auctioned off to all demand partners simultaneously. Advertisers not only get to look at every single impression, which wasn’t possible before, but also get the “first look” at every impression, basically getting the pick of the litter.

AdProfs

Header Bidding Code

The header bidding code (which naturally lives in the page header) executes, and the visitor’s browser calls all the demand partners (AppNexus, Index, OpenX, Amazon, Criteo, etc..) simultaneously and says, “Hey, lookie here! How much will you bid for me?” Within that few hundred milliseconds each makes a bid. The highest bids from each partner are then passed from a visitor’s browser to the publisher’s ad server. The bid value is then matched with a corresponding line item inside the publisher’s ad server. Since the header bid load first and is prioritized above direct orders, it guarantees delivery. If the bid from the header is chosen to serve, the winning advertiser’s ad is shown.

The beginnings of header bidding code from Prebid.js

If you wanted to see how this all actually works, Chrome offers an extension called Headerbid Expert.

The future of header bidding:

With the growth of U.S. advertisers spending digital ad dollars via programmatic channels, more web publishers are adopting header bidding to impact their bottom line from programmatic advertising. In fact, according to ServerBid’s assessment of the top 1,000 websites that offer programmatic advertising, more than half used header bidding as of May 2018.

Companies like Pubwise are transforming the industry and shifting to create a “header-as-a-service” model, notably being the first of its kind. Pubwise provides their publishing customers flexible tools that manage the header and programmatic advertising ecosystem, advanced demand and content analytics, their network of top tier demand relationships, a machine learning tool called Smart Path Optimization Technology (SPOT) ensuring optimal configurations are presented to each site visitor, real human support to optimize these tools, and finally those advanced enterprise controls that puts their publishing customers in the drivers seat if they so choose.

The future seems bright when you combine header bidding with machine learning to create more intelligent and targeted ads. Once this adoption begins to happen industry wide, publishers are going to be able to track performance across multiple exchanges, while spotting bottlenecks and eliminating inefficiencies in real-time.

--

--

Tun Khine

Software Engineer, former International Civil Servant, Dad, and new ATLien