Advertising seeks to convince an individual or a group of individuals to take a certain course of  action. The goal of advertising is persuasion, and for as long as humans have sought to influence  each other, advertising has been around in one form or another. Present day advertising utilizes  a variety of media and cutting edge technology to communicate targeted messages to audiences  across the globe. From the early days of the printing press, which enabled ads to circulate widely  in newspapers, to the invention of the television and Internet, advertising has evolved alongside  technology, refining the ways in which advertisers can target audiences.  

Advertising can be traced back to the Ancient world, where remains of ads have been found in  Ancient Egypt, China, Greece, Rome and Arabia. Sales messages written in papyrus and rock wall paintings were the earliest forms of out of home (OOH) advertising, like billboards today. Ads from the ancient world, and for thousands of years thereafter, would remain highly localized.  

For example, Figure 1 reflects a bronze plate ad for the Liu Family Needle Shop circa 990. Business was centered around local communities and commerce was restricted due to restrictions and difficulties with transportation. 

Figure 1

With the advent of deep water navigation in the 15^th Century,  the most powerful empires of Europe began globalizing the  world via major trade routes, which opened new markets. This, coupled at the same time with the  invention of the printing press, allowed sellers to mass produce ads to be placed in newspapers that could be circulated far and wide. As the world  entered the industrial revolution in the 18^th Century, the  capacity to produce goods and the  ability to widely distribute them  continued to increase dramatically.  

Figure 2

In response, sellers began to “brand” their goods, or turn everyday products into household  names. Early examples of this include James Lock & Co., a UK based men’s hat fitter established in  1676 and Hudson’s Bay Company, a Canadian retailer founded in 1670, see 1675 Tobacco ad in  Figure 2.  

A brand is defined as a product that is marketed in a uniform fashion, which ultimately establishes its own personality and image. Ranchers began “branding” their cattle two hundred years ago by  burning hot iron images onto their cattle to keep track and identify ownership.  With the diversity of media space available came the importance of creating a differentiated and  memorable brand, from here, a new business evolved: the advertising agency. These Madison Avenue organizations would assist advertisers in creating cutting edge messaging. In today’s model, agencies focused on messaging are called creative agencies. “Creative” is the industry term for the actual content of the ad, whether that be the picture for a billboard, the 15-second audio clip for radio, the 30-second visual ad for television, or the 3”x 2” message tailored for a newspaper. These early agencies would also assist in placing the messaging in the best  publications or locations. 

Figure 3

Advertisers within each agency would approach newspaper publishers directly to buy space on  the page to promote their brands. The benefit to both parties was clear: the business would gain  exposure and the publications would generate revenue. As America entered the 20^th century, a  variety of new media emerged to deliver ads to consumers, including radio and TV, which were  present in almost every home around the country. Radio advertising began to blossom in the  1920’s followed by television advertising in the 1950’s (see Figure 3).

The advertising industry was growing, and with it the responsibility of placing creative in well targeted publications. The job spun off into a separate business model, creating need for a new  type of company, “Media Agencies.” Once the assets or creative were produced, the Media Agency determined which medium and which publisher they should pay to show the advertisements. Media Agencies focused on aggregating advertisers in order to create leverage with pricing or buying power. By amassing a portfolio of large (high spending) advertisers, these agencies could force publishers (the biggest being the TV networks: ABC, CBS & NBC and big publishers like Hearst and Time) to lower rates across the board to win the business. While a strong portfolio dominated the early efforts by these Media Agencies, they also built strong attribution practices to determine the effectiveness of the ads placed across the landscape. 

Figure 4

The next medium for advertising was the personal computer in  

the late 1970’s and ultimately the Internet in the mid 1990’s. The ubiquity of the Internet  revolutionized the ways people would come to consume content and purchase goods on a massive  scale. The opportunities for advertising were (and remain) virtually unlimited, and the  advertisers and media agencies would have to begin to adapt to this revolutionary, fast- growing technology. Smaller agencies focused on the new Internet technology began popping up with  the aim of bridging the gap between web publishers and advertisers. These digital media agencies became indispensable, assisting advertisers in reaching this new flourishing new online audience. Advertisers began spending money on digital channels, seen in Figure 5. By the year  2000 the online industry reached $2 billion in overall media spend. Noting the chart to side overall  spend will reach nearly $100 Billion by 2019. 

Figure 5

These digital agencies focused their attention on the different formats available for advertising on  the Internet. Google.com launched in 1996 (see Figure 6). Today they are the dominating force in Search Engine Marketing. Google allows advertisers to customize  text ads, which would serve against individual Internet searches. 

Figure 6

Email marketing, the next incarnation of direct mail or catalog marketing, would target individual’s email addresses with offers catered to subscriber lists. Finally, the development of the banner advertisements and visual billboards online became known as “display advertising.” The first banner advertisement, or display ad appeared on hotwired.com,  the Internet publication of Wired Magazine in 1994. The ad in Figure 7, facilitated by Modem Media, was sold to  AT&T.  

Figure 7

With the expansion of the World Wide Web a new market emerged. Websites, much like newspapers in the early  days of advertising, began to focus on the number of  people who would be exposed to the ad. With billions of  web pages visited every day, the potential audience is  huge, allowing advertisers to reach larger audiences than  ever before.  Along with unprecedented reach, display ads allow advertisers the unique ability to measure and  target their audiences. As shown in Figure 8, “cookies” allow a technology provider to stamp a  user’s browser with an ID number.

Figure 8

In this way, a consumer can be tracked around the Internet,  much in the same way that animals can be stamped and traced. Internet technology allows direct  1:1 insight into the effectiveness of an advertising campaign. This is revolutionary compared to television, print, and OOH, which must be determined using in-store sales, loyalty cards, or unique  1-800 numbers to measure the effectiveness of an ad campaign.  

With the addition of cookies, the concept of “target advertising” became a reality. It quickly  became possible to serve ads on specific websites or subsets of a specific publisher. Using cookies,  companies began providing profiles of how users browsed the web independent of the specific  website in which the ad would appear. These capabilities, which are still evolving, provide much  more control for both advertisers and publishers in how ads are placed compared to all other  forms of traditional advertising.  With the clear advantages of display advertising, two types of organizations emerged as suppliers  of display inventory: publishers and networks. Major publishers continued to focus on the same  sales strategies as they had through other channels such as a magazine. They placed ads on the  top or the side of the webpage. Publishers would approach agencies to discuss specific sections  (news, finance, sports, entertainment, etc.), the amount of visitation, and the demographic  breakout of the audiences who regularly view these pages (as defined by third party audience  verification tools). These publishers would offer this space on a CPM (cost per mille, Latin word  meaning thousand) and compared the price per ad exposure or impression to the other channels,  like the circulation statistics used for pricing a magazine ad.

“Ad networks” were born in the late 1990’s, first as firms whom would represent publishers and  act as middlemen selling their inventory to agencies. One of the earliest networks was TeknoSurf  founded in 1998, which went on to become Advertising.com and was  eventually swallowed by AOL. Usually these were smaller publishers who were unable to fund their own sales teams. Boutique magazines, inspired bloggers, small newspapers, public record  aggregators, and others would begin driving traffic to their websites. Networks specialized in “traffic monetization,” handling the  operations behind selling and managing the banner inventory that would  show on the website.  

Figure 9 

Ad networks came to define any organization that would amass a group of publishers and present  their inventory in bulk to buyers. They could establish the same metrics as the larger publishers  and sell this inventory on a CPM basis. Because their clients were usually publishers focused on  producing content and not monetization, the networks could sell the inventory at a low CPM while  providing the publishers with a new, and at times substantial, revenue stream. 

To entice advertisers, these ad networks began building in-house ad delivery technology or ad  servers. This became an arms race, where the networks could customize the technology to offer  new products. Founded in 1996, Double Click was one of the first third party ad servers (see Figure 10). Now it is the largest and owned by Google. They used this technology to allow advertisers  to measure clicks and “conversions.” Conversions are, for  example, actions taken by a user who is exposed to an ad for Nike then later buys the sneakers from Nike.com. They created frequency capping, meaning a user would only see the same ad a certain number of times per day  (or per hour, month, etc.). They would sequence ads, meaning advertisers could create stories  played out over several pages. These advancements in technology continued to improve  drastically as computer space (servers) became cheaper and cheaper. 

Figure 10

Along with the explosion of search advertising, display advertising grew year over year from 2002  through 2008. This constant growth benefited both the major publishers and the networks.  According to www.adnetworkdirectory.com, there are currently 449 Networks – most of which  began during the upward swing of online advertising. 

The network model continues to grow today. Ad networks now come in many different forms and  vary from organizations that represent very prestigious publishers on a transparent and fair basis  to technology experts who use jargon to win agency budgets.  

With so many different types of content providers and middle men, the world of display  advertising remains very difficult to navigate. There are thousands of ad tech companies approaching agencies. Ultimately, ad serving has become a centerpiece for the entire display  ecosystem. Computers became extremely cheap and extremely fast at the end of the 00s, paving  the way for new developments in the Ad Tech world. In January 2007, the patent was filed for the  Online Ad Exchange, which marks the official introduction of Real-Time Bidding (RTB), the next technological step in the evolution of advertising. 

Download this White Paper:

This is the basic description of all players involved in the programmatic exchange ecosystem.  Display, Native, Video and Mobile In-App ads all use the OPEN RTB exchanges to help monetize inventory. 

WE MADE UP THE COMPANIES BELOW. FOR COMMON SSPs and DSPs, Google it.

This is how we see the ecosystem.

The publishers are the parties with content, money and resources.

The DSP & SSP, or the “Exchange” is full of venture capital money and great technology minds,  utilizing high speed processers and Hadoop capabilities across the web to create a fair and open  auction environment. 

Download this White Paper here:

Ever wonder how items you shop for online seem to follow you long after you’ve left the page?  It’s crucial to understand, as a consumer shopping online, or to as an advertiser looking to  increase ROI.

Real time bidding technology allows ad space on a web site to be purchased in the time period  between when the user calls for the page and when the page renders, as shown below in Figure  1. This transaction is accomplished through an instantaneous auction triggered by an SSP, or  Supply Side Platform, deconstructing the ad space characteristics into a string of text. The SSP is  also often referred to as an “Ad Exchange.” This string of text is an HTTP call, and is conveyed  very quickly across web servers (the Internet). The HTTP call which defines the space is sent to a  list of bidders. These bidders, or Demand Side Platforms, recognize the ad call and reply with  the amount they would be willing to pay to advertise in that space.¹ The SSPs then determine  the winning bidder and award the highest bidding DSP ad space at a price approximately  $0.00001 over the bid of the second highest bidding DSP.

Figure 1: Real Time Bidding in Action ²

• URL or Domain – typically the website (e.g. www.ebay.com), but can also be ‘masked’ by only providing the name of the network or ‘seller’
• Operating System – the OS of the computer loading the ad, be it Windows, Mac, etc…
• Browser – the browser of the computer loading the ad such as Chrome or Safari
• IP Address – The internet address of the computer loading the ad. IP addresses can be  converted into: 
  • Geo – the geographical location of the computer: Country, State, DMA, Zip 
  • Internet Speed – the speed of user’s connection (high speed, dial up, etc)
  • Business – 20% of the time this can be mapped to the business registered using  the IP Address 
  • Time – This is not actually passed in the Ad Call, but can be inferred as the  impressions are being sold in real time 
• Cookie ID – the identifying number according to the SSP of the user. This is a cookie, or text  file which is attached to the user’s browser. Usually this is a number at least 16 digits long.  At times a combination of numbers and letters. This allows the bidders to categorize users  on their end. With this cookie the buyers can target using the following: 
  • Frequency – How many times the bidder had seen the user. And how many  times the user has seen certain ads 
  • Data Segments – A host of companies offer categorization of users using cookie  IDs including: gender, age, income level, interest, browsing behavior, shopping  behavior, etc.  
  • Sequencing – Serving ads in sequence to a user 

Programmatic Media has a variety of different meanings. At its core the term ‘programmatic’  includes anything that is determined using machine learning. Since the early days of ad servers,  machine learning has been employed to select the most valuable ad to serve on the publisher  side. This technology, similar to how Google decides which search ads to show on Google.com,  has been around before the widespread adoption of the term programmatic. The term has  been a buzz term for the past few years as the buyers have gained control due to the RTB  revolution.

Part II. Types of Digital Campaigns

Programmatic Media includes all campaigns that employ technology and machine learning to  make advertising decisions on the buyer’s behalf. Each advertising campaign has the same basic parameters. Using a set of creative assets, the  marketer is going to run a certain amount of budget over a period of time with the intention of  driving a measurable result. Many different types of publishers and networks approach  marketers. Almost all networks use the RTB ecosystem in some fashion, but only four  organizations can claim to be ‘programmatic solutions.’

Most networks still use RTB solutions but do not inform the client of their techniques. By doing  this, networks are able to take higher margins than their typical business where they determine  a revenue share with their publishers. The four organizations inside the green circle depicted in  Figure 2 are those that are using the RTB technology. 

• Programmatic Direct: Programmatic direct includes campaigns run on specific publishers,  similar to an upfront insertion order (IO) deal, but is executed entirely through the DSP.
• Performance Network: The evolution of the Ad Network ultimately led to RTB specialists  who then started organizations that buy exclusively on the exchange and tout technology.  The Agency Trading Desks also fall under this category. These are groups who buy nearly the  entire inventory using RTB and sell the technology and ‘best in breed’ direct response  results. These organizations are not transparent in the costs of media, and therefore their  incentives are not aligned with the buyers.  
• Private Exchange: This is an RTB deal that allows an advertiser to control the targeting. The  publisher and the advertiser agree to floor pricing and the URL, but the advertiser can use  all other metrics in the Ad Call, through their DSP, to control the targeting. 
• Open Exchange: This is purely the advertiser buying inventory made available through the  exchange using a DSP. This is the most efficient way to purchase digital media, but the  individual controlling the DSP becomes very important.  

Download this White Paper:

FOOTNOTES 

¹ Actually, the DSP has an internal auction of its advertisers to determine the highest bid internally. This is the first of the instantaneous auction.  The DSP leads with the bid for the advertiser in their platform who bid the highest based on the Ad Call 

² Image sourced from livelyimpact.com

Real-time bidding (RTB) or what has become known as programmatic technology allows instantaneous decision making with regard to which ad will show on a website. Instead of deciding in advance which assets the ad server will display, now the decision can be made as the user  loads the website page. The ad server runs an auction to sell the ad space in real time, forming an online ad exchange driven by a Sell Side Platform (SSP). Not surprisingly, this technology has seen strong adoption and continues to thrive today. 

Figure 1: US RTB Digital Display Ad Spend

While the technology was developed in early 2007 it wasn’t until later that the market began  adopting at scale. The technology is now used  by much of the market and the amount of money spent through RTB continues to increase each  yea, as seen in Figure 1. ¹ 

RTB technology provides solutions to numerous problems that plague publishers and advertisers  trying to buy and sell digital ad space. On the publisher’s side, RTB technology solves issues of  forecasting and unsold inventory. On the advertiser’s side, RTB manages issues of distrust and  control.  

Forecasting Issues 

As display advertising became a ‘must buy’ for digital agencies, all major advertisers began  purchasing inventory. The most prestigious brands started cutting up-front deals with their  favorite publishers. All of the leading online publishers established ‘yield management’ teams  who were tasked with ensuring the rates created enough demand to fill the supply. These teams  would use data from the ad server, as well as other third-party applications, to determine how  many impressions were available in each section. They would then price out the inventory (available impressions) in the interest of maximizing revenue. Some publishers pulled talent from  the airline industry, as they were the most advanced in supply/demand analysis and inventory  management.

Figure 2: Actual Delivery vs Forecasted Availability Chart 1

The yield management team would use historical data to determine how many visitors on each  page. Using spreadsheets, advanced database techniques, or third-party software, these teams  would estimate how many ad impressions would be available for sale in the future. Sales teams  would, in turn, approach the market and sell the inventory, in pieces, to any advertiser willing to 

pay the premium prices. To avoid constantly providing ‘make-goods’, or future inventory due to  under delivery, the yield management team would forecast 5%-10% less than the exact amount  they anticipated. The inventory would vary by day and the forecasted inventory vs. sold inventory  would look like the chart in Figure 2.

Figure 3: Actual Delivery vs Forecasted Availability Chart

Unpredictability, an inherent characteristic of the Internet, creates a far larger loss of revenue  than shown above. The Internet provides the most up to the minute information possible, and  when major events happen websites can see 10x the traffic they had forecasted. The supply  spike can occur on a weather website as a northeaster approaches, a news website during a  natural disaster, a gossip website after the sudden death of a major celebrity. The advertisers do  not want ‘uneven delivery’, meaning the publisher cannot deliver all of the impressions for a 12- 

week campaign in one single day. This means that when a spike of 10x occurs, even if the yield  team anticipates this based on the morning numbers, they cannot deliver more than 2x or 3x the  amount they predicted. The monetization of that day would look like Figure 3.  

Depending on circumstance, this could equate to lost revenues anywhere from tens of thousands  of dollars to several million, all due to the inability to monetize inventory in one single day. 

Unsold Forecasted Inventory 

Premium publishers who can sell the ability to reach a specific niche audience are often able to  set their prices correctly as to sell through 100% of their inventory. Other publishers have the  opposite problem. The largest internet publishers, or portals to the internet, amassed such  enormous traffic they cannot sell all of their ad space at effective rates. The largest portals saw 

three billion visits a day at times. Three ads on a page per day equates to 9,000,000,000 available  each day. If organizations with too much inventory could determine a method to sell the long tail or ‘remnant’ inventory at even a $0.01 CPM they would make millions of dollars a month in extra revenue. While remnant solutions were available, they still could not handle the majority  of this unsold inventory. 

Distrust 

When Ad Networks started gaining popularity circa 2000, they all represented unique publishers  and could show differentiation with their ‘site lists’ or domains in which ads would serve. The  biggest publishers began running their ads across several different networks. This style of yield  management would become the core of the Internet publishing business. By rotating the ads, the  networks began having an issue of differentiation, as the same site would be available across a  variety of networks. Additionally, many of these ad networks were restricted to selling blindly or  undisclosed in order not to undermine the publishers’ relationships with marketers. 

To achieve the scale necessary, they began to avoid disclosing to buyers the delivery reports by  domain. They would insist they could not share how the delivery varied across the websites on  the ‘site list’. This would allow them to ‘dump’ a lot of inventory on a low-quality website.  Facebook apps such as Farmville, websites such as myclassmates.com and dictionary.com are  classic locations where you will see network ads in volumes, which are not conveyed to the  agency or more importantly the advertiser. These networks could also adjust their margin 

methodology and charge advertiser’s margins as high as 70%-80%. 

Some networks began using other forms of technology to show differentiation. Networks claimed  uniqueness by “focusing on a special audience” or “providing scale” or “using contextual search  technology,” but until the advancement of the programmatic RTB exchange there was basically  no difference between these organizations as they all worked with the same publishers. 

As these problems leaked out of the back rooms of these networks and into the ears of media  buyers, there was a clear need for greater transparency. By handling the auction for the  advertising space on an impression-by-impression basis, the RTB exchange offered a  technological solution.  

Control 

The networks were the first to implement and develop ad serving technology and continued to develop in the interest of their own bottom lines. The best advances in targeting were made  available to the publishers. Publishers could use the up to date ad servers to control what type  of ads would show on a certain section of a website, in a specific geographical location, or at a  particular time. While advertisers could request these targeting parameters, there was no  control on their side of how the ads were served. Someone on the publisher or network side  needed to handle the trafficking.

As the lack of control became a growing issue on the advertiser side, the RTB technology became  available. Programmatic pushed the ‘control’ into the hands of the “buy-side” or what is often  referred to Demand-side. This allowed the advertiser to manage how they want to target their  ads across publishers (in bulk, becoming ‘exchanges’). This was a clear benefit for all involved, as  the newfound confidence led to larger budgets. 

The RTB exchange provided solutions to these four issues, but the agencies we’re set in their ways and comfortable with the accustomed methods. Changing the way an entire industry  behaves takes time and takes ‘buy-in’ from big players early. How could the networks and  publishers ensure the advertisers would test this technology that offered ‘control’ and  ‘transparency’ into pricing and targeting? They created an auction environment promising the  buyers they would pay only a ‘penny’ more than the next highest bidder for each impression.  This Second Price Auction method created a ‘perfect’ market allowing the next generation of  media to emerge: Programmatic Media. 

Download this White Paper:

References 

¹ http://www.iab.net/about_the_iab/recent_press_releases/press_release_archive/1996_pr_archive 

² (or nodes, or boxes, or servers – all terms which refer to a computer, usually without a monitor and in a building with no windows  somewhere relatively unpopulated) 

³ Pemberton, Steve. http://homepages.cwi.nl/~steven/Talks/2011/05-07-steven-visualisation/ , CWI and W3C, Amsterdam, © 2011 ⁴ http://www.google.com/patents/US20070192356.pdf 

⁵ http://adage.com/article/digital/real-time-bidding-account-25-display-ad-spending-2015-emarketer/238300/

When online advertising hit its stride in the 2000’s, DoubleClick “ad serving” service became the  premier technology used by publishers to configure their website’s ad space. With  overwhelming market share, this technology continued to grow and balloon into a massive  organization. In 2007 it was purchased by Google. Doubleclick was making $300 million in  revenue at that point, primarily from the publisher technology. Doubleclick for Publishers, or  DFP, is a program used by major publishers across the world to set up the targeting schemes  advertisers demand from their publisher partners (see Figure 1). 

With the advancements of real-time bidding (see ‘RTB & Programmatic Media’ White Paper),  DFP would allow the publisher to sell the ad on the ‘Ad Exchange’ via Doubleclick AdExchange  (AdX). This is the ad exchange product produced by Google for publishers to use to sell their ads  in real time with very strong results. Although AdX is not the only exchange in the market.  There are several other competitors to the Doubleclick exchange which publishers can use. 

AdX integrates seamlessly with the DFP. This is not surprising, because they are both produced  by the same company and can build the technology into the same program. But, AdX is not the  only ad exchange on the market. Many other technology players have built exchanges, such as  AppNexus, Index Exchange, Rubicon, Pubmatic, Mopub, Smaato, OpenX. These systems do NOT integrate as seamlessly into DFP. In the below chart they are the ‘other networks’.

Because of Header Bidding, these other exchanges now seamlessly integrate with AdX, allowing the publisher to give these other exchanges equal opportunity to participate in auction for the ad space.

How does this impact Demand Side Platforms? 

At the end of the day, the ones most effected by Header Bidding implementation are the  Demand Side Platforms (see ‘How the Ad Exchange Works’ White Paper). The instantaneous  auction the Demand Side Platform (DSP) responds to happens when the exchange makes the ad  available while the user is loading the webpage. This is the definition of Real-Time bidding and  the technology driving Programmatic advertising. Before header bidding, the SSP would only  auction the ad when their slot was called in the publisher waterfall. Now, each time the page is  loaded, the SSP makes the ad available for auction, even if it rarely is responsible for selling the  ad unit. DSPs must ‘listen’ or ingest the ad calls from billions of web pages a day. With the  advent of header bidding, now DSPs see 3x – 5x the amount of traffic they saw previously. This  is an increased load on the need for servers to respond to all these available ad units. Many of  the smaller DSPs are having a difficult time keeping up and, in the end, they are not able to see  as many available ad spaces as the used to see.  

For prospecting campaigns, or video campaigns, this is not a big deal. DSPs can still pick off the  ads they want to target and typically get performance similar to what they saw before header  bidding. But for remarketing campaigns, which are the money-makers in this business, these  smaller DSPs are having a hard time ‘listening’ to enough inventory to compete with the larger scaled DSPs on remarketing campaigns. 

Another, buyer-be-ware change with the increased use of header bidding is the introduction of  ‘First-Priced Auctions’. Typically, an SSP would see all the DSP bids, award the ad space to the  highest bidder, but charge them the PRICE OF THE SECOND HIGHEST BID. This was done to  increase competition and buyer’s interest in this technology as it first appeared. As a buyer this  statement made this landscape very intriguing (although, still extremely complicated): “You will  NEVER pay more than the second highest bidder is willing to pay for each individual ad space”.  It was in the interest of developing a perfect market. Well, things don’t always work out like  that. 

Now, with the need for the exchanges to push the highest price to DFP through the header  auction, they are now passing the price they receive directly from the DSP to the publisher. This  is a ‘first priced auction’. Or, if as the exchange would justify “we are taking the highest bid we  saw to increase the publishers CPM”. This is not a bad thing. It does question why we have the  SSP layer at all. But, ignoring that statement, this is ultimately going to drive better pricing for  the publisher who must exist for us to have the ad tech ecosystem at all!

But, with the increase in first price auctions, DSPs must now build into their technology the  ability to identify first-priced auctions versus second price auctions and enable the buyer to  correct their bid strategy to avoid getting hurt by this change in auction mechanics. 

What does this mean for a Trading Desks? 

Well, to be honest, not much. If anything, it is a positive change in the ecosystem. Most trading  desk practices license technology from the largest DSPs in the world who have no problem  seeing all available bids, even with the increase in header-bidding auctions. Any reputable  trading desk, would already have implemented advanced bidding algorithms to drive the  strongest cost per acquisition. Whether or not they pass those savings onto their client is a  whole other story, hence the shift for a more transparent partner like ProgMechs. In the end,  not much changes with this introduction except that publishers are able to make more money  through programmatic channels. With that, more publishers make more inventory available,  allowing Trading Desks to see more inventory, increase performance, and continue to offer  stronger, more ‘high-impact’ style ad units.

Download this White Paper:

In the digital advertising industry, the term “Trader” refers to an individual who manages ad  campaigns through a Direct Service Platform (DSP) in a real-time situation. The term is used to  describe an industry professional who buys and sells commodities, currency or financial  instruments. As the ‘Wall Street Trader’ puts it: 

“A very good trader has a very deep understanding of the products” ¹ 

While there are many differences between traders on Wall Street and traders working in  media, both must have similar skill sets necessary to be successful in their “trade”. Like  financial traders, the best media traders have an extremely clear understanding of what it is  that they are trading and its entire ecosystem. Financial traders and RTB traders who work to  buy and sell advertisement space, both need to have sound analytical and data mining skills. An excellent media trader must excel in all four responsibilities associated with executing an  RTB campaign: 

1. Market Education 
2. Ad Ops: Launch Documentation / Tag Management / Pixel Management  
3. Trading: Campaign Management & Optimization (Managed Service) 
4. Reporting: Standard and Customized 

Market Education 

A large part of media trading is the ability to describe the mechanics and market color  surrounding the real-time interactions and cookie targeting to the principal who controls the  budget. Whether the principal is an advertiser, agency or a buyer, a great media trader will be  able to explain, in as much depth as necessary, to the buyer how each strategy works.  Strategies continually evolve across channels, but can be categorized or more of the following:

• Retargeting (use of 1^st party cookie data to drive online actions) 
• Prospecting (use of algorithm to drive online actions) 
• Branding (use of media to drive offline & long term actions) 
• Third party data targeting (use of 3^rd party cookie) 
• Contextual targeting (use of web crawler / semantic targeting) 
• Hyper geo targeting (zip code, IP based, or lat/long based targeting)  

Each one of these targeting solutions define the way data will be manipulated to execute the  strategy: through one or more demand side platforms, across mobile, display, video, & social  advertising channels. There are advanced integrations that should be used for each set up,  including post-back URL reporting and attribution alignment. As strategies are introduced, the media  trader often assists the advertisers in defining and refining the specific goals and methods by  which the campaign will be executed.

Ad Operations 

The technical side of launching a campaign include ad operations duties, that consist of three  distinct categories:  

1. Launch Documentation – A successful transition from the hands of a media planner to a trader  must convey specific information. It must clearly delineate the restrictions on the placement of  the ads and the campaign goals in concrete terms. Generally, agencies need some assistance in  putting their plans into executable trading instructions. Much of any campaign’s success can be  attributed to clean, strong, and detailed launch documents. The iterative process of creating this document has a positive and direct effect on the quality of the campaign.  
2. Tag Management – Most campaigns have 3^rd party creative assets. Uploading, debugging, click  tracking and implementing macros is referred to as ‘tag management’. Tag management  includes confirming the creative serves correctly, is eligible for all applicable publishers, and  tracking components on the 3^rd party ad serving sites are setup correctly. This also includes  troubleshooting all creative issues. 
3. Attribution – Most large-scale digital advertising campaigns have some type of key performance  indicators (KPI) which define the success of the campaign. In display, this is generally cookie  based. In mobile, this is usually a post-back URL integration. In the case of managing attribution,  this includes any troubleshooting or implementation of pixels, post-back URLs, macros, and click  trackers.  

Trading 

Depending on the strategy associated with the campaign, managing a campaign’s day to day is  referred to as ‘trading’. Trading is the core of intellectual property in regards to driving successful results. Trading includes setting the campaign line items and strategies up in the  demand side platform via UI or API. The trader, or person who executes the campaign is the  party that ‘green lights’ the campaign. Once the campaign is ‘spending,’ or purchasing the ads,  the trader is responsible for the feedback cycle that is running reports and making changes to  the targeting. These changes are in the interest of achieving the highest return on investment  (ROI) as well as monitoring for fraud. The trader is the person responsible for ensuring best  possible outcomes.

Reporting 

Reporting is the process of pulling the big data results of a campaign and presenting them in an  understandable and clear fashion. Most reporting in the industry is done in Microsoft Excel or  via Dashboards. These are all web interfaces or applications that aggregate data on the  backend. A talented media trader is able to pull all of the reporting and highlight not only the  performance of the campaign, but the key drivers of that performance. The reporting also  includes details where the media ran and attribution goals. 

Ultimately, a media trader can drive results through the RTB exchanges and explain how  mechanics were utilized to achieve success. They can integrate any form of attribution into a  group of DSP partners, set up detailed conversions reporting and determine fraud. They should  be able to recommend budgets and forecast on the scope of each campaign before launch. An  excellent media trader, much like financial traders, can clearly bring the efficiencies and value  proposition from the market directly to their clients.

Download this White Paper:

REFERENCE 

¹ Justin Weinberg