<?xml version="1.0" encoding="UTF-8"?><rss xmlns:dc="http://purl.org/dc/elements/1.1/" xmlns:content="http://purl.org/rss/1.0/modules/content/" xmlns:atom="http://www.w3.org/2005/Atom" version="2.0" xmlns:cc="http://cyber.law.harvard.edu/rss/creativeCommonsRssModule.html">
    <channel>
        <title><![CDATA[Stories by Colin Behr on Medium]]></title>
        <description><![CDATA[Stories by Colin Behr on Medium]]></description>
        <link>https://medium.com/@colinbehr?source=rss-74e63589271b------2</link>
        <image>
            <url>https://cdn-images-1.medium.com/fit/c/150/150/1*EKsJtCFyfgRaiqs5iMz3yg.png</url>
            <title>Stories by Colin Behr on Medium</title>
            <link>https://medium.com/@colinbehr?source=rss-74e63589271b------2</link>
        </image>
        <generator>Medium</generator>
        <lastBuildDate>Thu, 19 Nov 2026 14:21:31 GMT</lastBuildDate>
        <atom:link href="https://medium.com/@colinbehr/feed" rel="self" type="application/rss+xml"/>
        <webMaster><![CDATA[yourfriends@medium.com]]></webMaster>
        <atom:link href="http://medium.superfeedr.com/" rel="hub"/>
        <item>
            <title><![CDATA[Three Ways Header Bidding Will Gain Adoption in a Mobile SDK World]]></title>
            <link>https://medium.com/mmt-wtf/three-ways-header-bidding-will-gain-adoption-in-a-mobile-sdk-world-8573d7740db3?source=rss-74e63589271b------2</link>
            <guid isPermaLink="false">https://medium.com/p/8573d7740db3</guid>
            <category><![CDATA[header-bidding]]></category>
            <category><![CDATA[adtech]]></category>
            <category><![CDATA[programmatic]]></category>
            <category><![CDATA[mediation]]></category>
            <category><![CDATA[advertising]]></category>
            <dc:creator><![CDATA[Colin Behr]]></dc:creator>
            <pubDate>Thu, 06 Jul 2017 04:28:48 GMT</pubDate>
            <atom:updated>2021-04-19T19:25:45.250Z</atom:updated>
            <content:encoded><![CDATA[<figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*sQjiDSCepifdlNRx2CdT8Q.png" /></figure><p>It has long been pointed out that mobile apps do not have headers. Despite this, industry experts, publishers and other ad tech players have been predicting the rise of header bidding for mobile apps.</p><blockquote>Given the obvious limitations of the term ‘header bidding’ in the context of SDKs, “<strong>parallel bidding” or “parallel auctions”</strong> are more appropriate terms.</blockquote><p>There are a few key reasons why parallel bidding has become hugely topical:</p><ol><li>Header bidding in the web has seen tremendous adoption. <a href="http://www.businessinsider.com/header-bidding-gains-momentum-drives-up-publisher-ad-revenue-2016-5">According to some sources, publishers might typically expect to see 20–50% increases in revenue.</a> As a result, up to 70% of top publishers in the US have adopted header bidding according to Business Insider.</li><li>Mobile app activity and revenue has continued to grow at a tremendous pace, with mobile ad revenue now closing the gap against time spent according to <a href="http://www.kpcb.com/internet-trends">Mary Meeker’s latest internet trends report</a>.</li><li>Mobile SDKs are a black box, which makes it impossible to know the true payout for an impression won by a network. Even the main SDK based exchanges do not pass a bid price, which necessarily restricts publishers to more basic forms of waterfall mediation. This has been the status quo since the birth of the App Store.</li></ol><p>The impact is clear. A lack of transparency means that publishers make less money and advertisers often receive less visibility. User experience can also suffer, when publishers are unable to effectively blacklist specific advertisers or ad content.</p><p>There is good news — this world is changing quickly, with three key adoption methods for parallel bidding emerging in the industry:</p><h3>Pre-bidding</h3><blockquote><em>Definition: the use of multiple ad calls for each ad provider in a waterfall.</em></blockquote><p><strong>What is it?<br></strong>A limited, although substantial group of publishers have been using this method over the past 2+ years in order to maximize revenue from their waterfalls.</p><p>In a typical SDK based mediation waterfall, each partner is ranked based on historical CPMs. The look-back window may vary depending upon the feature set of the network partners and the mediation layer. The most common standard for ranking is based on country &amp; OS level of the user. This provides enough data to be meaningful (in most cases), while breaking down two of the largest differences.</p><p>Pre-bidding works in much the same way, the key difference being the existence of multiple ad calls.</p><p><strong>Why is this so different to a standard waterfall? </strong><br>In particular it allows different networks to be called multiple times, further separating out performance beyond the OS &amp; country level. Networks typically look at far more variables than a publisher to determine the value of each impression, giving them the chance to fill at different floor prices.</p><p><strong>How is this supported today?<br></strong>1. Multiple ad calls are supported by most SDK ad networks. They are usually referred to as ‘placements’ or ‘ad zones’. Both are confusing terms, since they can be used in order to create multiple different calls, either to show ads in different <em>placements, </em>or to call an ad network more than once in a given placement.</p><p>2. Floor prices can be a trickier ask depending on the network partner and business terms. Some network partners allow a true floor price, paying the desired CPM only as a minimum and going higher based on revenue shares. Other networks may support other features or workarounds, such as target price (not a minimum, purely the target for a given placement) or frequency capping at a placement level in order to boost CPMs at different levels.</p><p><strong>What is the impact of this in the market?<br></strong>The impact has certainly already been substantial for those publishers who support it. However, the impact has been limited by the lack of support for multiple calls within an individual placement by mediation layers. Current support for this feature is limited to just two main partners in the eco-system — Mopub and Appodeal. For other mediation layers, it’s a question of understanding which approach they think needs to be supported the most.</p><h3>Wrapper bidding</h3><blockquote>Definition: the passing of a predicted impression price at the SDK/client level.</blockquote><p><strong>What is it?</strong><br>SDK mediation waterfalls are often a very static affair. Once the app has initialized the mediation SDK, usually all other mediation SDKs are arranged in a pre-set order. This is largely a foregone conclusion, given that the mediation SDK will simply call for the ranking based on historical average performance.</p><p>However, most ad networks have a predicted value for the ad that they want to show. As their SDK gets called on app startup to check for an ad, the decisioning of the network will assign value to each potential ad, which determines the end ad shown to the user. Given that this price has already been calculated, it stands to reason that exposing it in a simple manner will allow for an increased level of accuracy and transparency.</p><p><strong>Why is this so different to a standard waterfall?<br></strong>Simply put, the passing of a price (real or predicted) is a significant change to aggregating the value of all impressions. Just as you wouldn’t expect a real estate agent to value all houses based purely on the county that they are in, you want to see a value for each individual impression.</p><p><strong>How is this supported today?<br></strong>Some networks have already begun to expose predicted value via their SDK. This has also begun to be a frequently explored topic within the industry. However, this is not yet the norm, and it is unclear whether mediation layers will support it. The increased transparency is a great thing for publishers, but the potential difference between the real price and the predicted price could be a problem.</p><p><strong>What is the impact of this in the market?<br></strong>This is still unclear given the lack of evidence on the market. We can assume that any level of granularity is better than nothing. At the same time, the potential operational overhead of dealing with predicted vs. real prices in an auction is also difficult to quantify. If publishers and mediation layers can figure out the right way to use the data, this might well provide the quickest path to a half-way point in the realm of real time bidding.</p><h3>Parallel bidding</h3><blockquote>Definition: true real time bids passed in parallel server to server, while ads are delivered to and displayed through their respective SDKs.</blockquote><p>Parallel bidding is currently pre-adoption. While there is plenty of discussion and talk, the standards for dealing with this in an SDK environment have not yet been established. All the same, I have little doubt that this will happen.</p><p>Critically, parallel bidding can only be a success if the mediator is required to transparently participate in the same parallel auction.</p><p><strong>What is it?<br></strong>If wrapper based bidding involves the passing of a price from one SDK to another, think of parallel bidding as a much more traditional approach to holding an auction. There is one big exception.</p><p>In a purely server to server (OpenRTB) auction, the buyer is sent a request, who then responds with a price and a creative. This transaction is real time, and binding in terms of the price. The advantage is clear, the transaction occurs immediately, without the need to account for any potential differences. It also occurs <em>before</em><strong><em> </em></strong>the creative is sent to the client side SDK to be displayed, which yields many additional advantages that won’t be tackled here.</p><p>In a parallel auction, each SDK is initialized as per standard protocol. At the same time, the mediator will send a request to each mediated network with a standard OpenRTB request, which can then be matched based on IDFA/AID and a time stamp. This is possible since each SDK should receive the same request at the same time. Each network then responds to the OpenRTB request, allowing the mediator to hold a real time auction and choose which SDK will show the ad based on the winning bid.</p><p><strong>Why is this so different to a standard waterfall?<br></strong>Besides the existence of a real price, this brings us back to a much more standard auction process. This would mark a significant departure from an approach where each SDK is entirely a black box, with real prices being passed on the server side. The act of actually bidding, is entirely different from simply fetching historical CPMs to determine an auction, ensuring a much fairer auction for all.</p><p>The most important difference of all — the presence of a competitive, real time, second price auction will force a higher average price from the mediator’s ad exchange. It will no longer be possible to unfairly outbid a mediation waterfall by arbitraging the difference between the real value of the impression and the average.</p><p><strong>How is this supported today?<br></strong>Simply put, it is not. OpenRTB spec does not account for the added complexity of bidding through a server side auction while delivering the ad to your own SDK. Until it does, expect this to be a point of ongoing negotiation between different players in the eco-system.</p><p><strong>What is the impact of this in the market?<br></strong>To refer back to the beginning of the article, real time bids that push up exchange payouts have had a tremendous impact on the market so far, but restricted to the web. We can expect publishers to see double-digit percentage growth in revenue, while networks will have a fairer chance to compete.</p><h3>Final Word</h3><p>The details and approaches of parallel bidding still need to be fully defined by the industry. The specifics of wrapper and true parallel bidding are far from outlined, but progress is still being made. As a publisher, with careful planning you can still take advantage of improvements around multiple ad calls, which will increase your revenue. As an advertiser, you will probably need to wait until full transparency through bidding becomes the norm in order to truly benefit.</p><p>If you or any friend happen to have additional thoughts on the topic, please do not hesitate to get in contact.</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=8573d7740db3" width="1" height="1" alt=""><hr><p><a href="https://medium.com/mmt-wtf/three-ways-header-bidding-will-gain-adoption-in-a-mobile-sdk-world-8573d7740db3">Three Ways Header Bidding Will Gain Adoption in a Mobile SDK World</a> was originally published in <a href="https://medium.com/mmt-wtf">mmt.wtf</a> on Medium, where people are continuing the conversation by highlighting and responding to this story.</p>]]></content:encoded>
        </item>
    </channel>
</rss>