Is The Future Programmatic?
This commentary was originally written in April 2016 for an MBA course at the Schulich School of Business called “Strategic Solutions for Digital Media”. This is an emerging topic and was quite a challenge to research. The version below has been slightly edited.
Programmatic advertising is everywhere. Currently most prevalent in the digital space, programmatic is making its way into out-of-home and television ads. On the surface, Programmatic TV seems both intriguing and impossible, a collision of the future with the past. But it is very real, it is happening now, and it presents several strategic issues that managers in and around the television and advertising industries should be aware of.
The report begins with an overview of programmatic advertising in the digital context: providing a breakdown of the four types of programmatic ads, the size of the market, and three major issues facing digital programmatic. It goes on to examine the state of the TV advertising industry. Finally, it explores Programmatic TV: the three distinct concepts that are often combined into the term “programmatic TV”, the size of the market and recent developments, and the current challenges and opportunities. The report concludes by discussing four strategic repercussions of programmatic TV that managers should be prepared for.
Programmatic Advertising Overview
Ask a room of six digital marketing pros to define programmatic advertising and you’ll get six slightly different answers. In fact, there are countless variations of the definition published online and used in conversation. The lack of a standard definition tends to cloud the discourse on programmatic so for the purposes of this report, I will define programmatic advertising quite broadly, as the use of technology to automate advertising transactions.
4 Types of Programmatic Advertising
Many people associate all programmatic advertising with one specific type of programmatic, the open-to-any-bidder real time bidding (RTB) auction popularized by ad exchanges such as DoubleClick by Google. While this type of programmatic, called Open Auction or Open Exchange Buy, is by far the most prevalent, accounting for around 70% of all programmatic revenues, it is not the only type of programmatic advertising. There are four main types of programmatic ads, differentiated by factors such as pricing, inventory, and buyer participation.
Programmatic Advertising Industry Growing Rapidly
The exact size of the programmatic market can be tricky to estimate due to the differing definitions of what constitutes programmatic. For example, some estimates include social media and search as part of programmatic while other estimates exclude one or both of these types of ads. Despite the differences, there is near unanimous consensus that, no matter how you define it, programmatic is experiencing tremendous growth.
In its Programmatic Revenue Report, PwC estimated that in 2014, total programmatic revenues in the US reached $10.1bn, representing 52% of the $19.6bn of total display advertising revenues and about 20% of total Internet advertising revenues ($49.5bn in 2014). PwC figures include all non-search ads.
eMarketer pegs 2015 US programmatic digital display ad spending at $15.43bn and estimates that figure will reach $26.78bn in 2019. If its prediction for 2019 holds, programmatic would account for 72% of all digital display ad spending that year. eMarketer’s definition of digital display ads include social media ads, native ads, sponsorships, and several other types of ads.
Globally, the programmatic market is estimated to grow from $14.2bn in 2015 to $36.8bn in 2019, according to Magna Global Intelligence. Magna Global counts digital and video as programmatic, but does not include social media or search ads. While the pace of growth is rapid and programmatic is poised to dominate certain industry segments, it’s important to point out that programmatic is still a tiny fraction of overall digital and media spending. At $14.2bn in 2015, programmatic accounts for only 9% of total global digital spend ($159bn in 2015) and less than 3% of total global media spend ($513bn in 2015). On a country level, the US is the biggest market for programmatic in the world, representing well over 50% of global spend. The next biggest markets are the UK, Japan, China, and Germany. Combined, these top five markets account for over 75% of total global programmatic spend.
Canada lags its international counterparts in programmatic. According to eMarketer analyst Paul Briggs, “programmatic advertising in Canada trails other comparable markets in terms of its share of total digital ad spending due primarily to the structure of the market—relatively few publishers offering advertisers options for programmatic inventory”. eMarketer estimates programmatic digital display ad spending in Canada to be worth $881.8m in 2015, rising to $1.51bn by 2017. Magna Global estimates Canada to be the 10th largest programmatic market globally, with spend totalling roughly $300m in 2015. Despite the limitations in the market, Canadian brands and media buyers have adopted programmatic en masse, especially when it comes to digital video. 94% of Canadian buyers are purchasing digital video programmatically and 8 of 10 brands have either invested in bringing programmatic video buying in-house or plan to do so over the next year.
3 Issues Holding Back Programmatic
Programmatic is attractive to both advertisers and publishers. Buyers want it because it increases their access to inventory, improves efficiency, provides deeper targeting opportunities, and allows real time adjustment to ads. Publishers can benefit from operational efficiencies and, by increasing ad effectiveness through targeting, can better monetize certain parts of their inventory based on ad ROI. However, several issues still hold programmatic back from more widespread adoption. The three main issues are a lack of quality premium inventory, fraudulent transactions, and a complicated and fragmented path from advertiser to viewer that leads to increased costs.
1/ Lack of Premium Inventory: When programmatic advertising was first launched, it was used as a platform to trade remnant digital display ad inventory at prices so low that nobody cared about ad quality. A lot has changed in programmatic since those early days, with an increasing amount of higher end inventory being made available via programmatic (primary via non-Open Auction programmatic). While part of the early reputation for low quality remnant inventory persists, some in the business have been acknowledging the shift to better inventory. A 2014 Google survey of buyers found 62.7% already seeing improved programmatic inventory and 86.4% expecting to see higher-quality inventory over the next 2 years.
However, despite the rosy outlook, inventory issues continue to persist, especially in banner ads. Facebook’s experiment with programmatic selling of digital ads through its Atlas ad services platform ended abruptly in March 2016. According to Dave Jakubowski, Facebook’s Head of Ad Tech, “we were able to deliver ads to real people with unprecedented accuracy, but came up against many bad ads and fraud (like bots). While we were fortunately able to root out the bad actors and only buy quality ads, we were amazed by the volume of valueless inventory”. He went on to say that low quality ads accounted for over 75% of the volume coming into its system and ultimately, the company pulled the plug on banner ads because “only two ad formats delivered significant value: native & video”. While native and video can also be transacted programmatically, issues around the quality of the inventory of digital banner ads, the bulk of ads available programmatically, will continue to hound all programmatic advertising for some time.
2/ Fraudulent Transactions: As alluded to by Facebook’s Jakubowski, poor-quality ad inventory and fraud go hand in hand. Programmatic fraud can take three forms: non-human traffic (bots), zero viewability, and intentional misrepresentation. Non-human traffic, or bot traffic, is the most widely known form of fraud. It is used to generate everything from fake page views and ad clicks to fake form submissions. Zero chance of viewability fraud occurs when ads are hidden or obscured so that real human visitors have a near 0% chance of actually seeing them (but the advertiser is still charged). Intentional misrepresentation can take many forms from arbitrage (where a website buys traffic to its site for less than it sells ad impressions for) to click farms (real people paid to “work from home” who click on ads and fill out submission forms). Fraud used to be endemic in programmatic advertising, with as much as 70% of 2010 traffic suspected of being fraudulent.
Today, fraud is much less prevalent. In Q4 2015, 10.5% of display traffic and 10% of video traffic was fraudulent. However, in terms of overall economic value, fraud still has a significant impact. The Association of National Advertisers estimates that, globally, fraud will cost advertisers an estimated $7.2bn in 2016. Fraudsters tend to target higher-value programmatic opportunities, such as premium video content, or highly sought-after demographic groups. For example, video ads with CPMs above $15 had a 173% higher incidence of fraud than average and Hispanic-targeted programmatic media attracted 70% more bots than its non-Hispanic counterpart. With so much money at stake, the cat and mouse game between fraudsters and the ad industry is likely to continue for the foreseeable future.
3/ A Fragmented Industry: Programmatic advertising involves many intermediary ad tech companies who operate in the space between advertiser and publisher. On the buyer side, agencies, trading desks, demand size platforms (DSPs), 3rd party data/targeting, and verification platforms are just some of the players. Ad exchanges, supply side platforms (SSPs), and ad servers are some of players on the publisher side. The ad tech used in a programmatic campaign can vary widely depending on the nature of the buy. In aggregate, ad tech fees, also known as the “ad tech tax”, are substantial. Ad tech accounts for 55% of total programmatic revenues. The following hypothetical programmatic campaign demonstrates where value is captured in the execution of a campaign.
With so much of the value captured by intermediaries, there will be some opportunity for market players to consolidate the fragmented middle of programmatic as the platform matures in the future.
TV Advertising Overview
In today’s media landscape, “TV” comes in many incarnations. Broadcast, Cable, Satellite, On-Demand (SVOD and AVOD), TV Everywhere, and TV Channel Web Sites can all be thought of as TV. Since this report looks at programmatic TV, I’ll keep the focus on advertising supported TV models.
Traditional TV: Maybe Down, Definitely Not Out
Traditional TV, or linear TV, consisting of broadcast networks, cable and satellite programmers, and pay TV operators, remains the dominant medium for consuming video content. In 2015, traditional TV captured 72% of all video viewing hours in the US. Traditional TV also captures almost all the value in the ecosystem. Of the $185bn of US video revenues earned in 2015, 55.6% ($103.1bn) came from pay TV subscriptions and 36.2% ($67bn) from TV advertising.
Much has been written about the pending demise of traditional TV, but research and data don’t tend to support the doom and gloom predictions. For example, a 2016 survey of advertisers, agencies, publishers, and media companies conducted by Forrester Research found that 72% of respondents thought that, over the next three years, consumers’ time spent watching TV at the time it is broadcast would significantly improve, moderately improve, or stay the same. 71% of respondents thought that time watching television delivered via antenna or a cable/satellite provider would increase or stay the same.
There are numerous other reports and surveys that refute the TV is dying narrative. TV is, and will continue to be, the best medium for reaching large audiences at scale. From an advertising perspective, television’s advantage cannot be overstated. According to RBC Capital Markets’ David Bank, for advertisers, one first-run 30-minute episode of The Big Bang Theory is as valuable as ALL YouTube content for a week. However, the media landscape is changing. People are consuming more video overall, and the consumption growth is coming from non-traditional TV platforms. Advertising dollars, which have historically been heavily skewed towards TV and which are not keeping up with growth in digital media consumption, are being re-allocated to better reflect current audience media consumption patterns. In fact, digital media revenues are expected to overtake TV ad revenues in the US for the first time in 2016.
Changing consumption patterns are contributing to the fragmentation of traditional TV. The days of 3 big networks dominating the landscape in the US are long gone. Modern-day TV is spread across multiple channels. The fragmented viewership presents advertisers with new challenges by making it harder to measure who is watching and making the desired target audience more difficult to reach. In 2014, 50% of all TV ad dollars were spent on 11 networks. Those networks only reached 27% of available TV viewers. As Yin Rani, VP of Marketing Activation at Campbell’s puts it, “it’s the golden age of TV content but not the golden age of TV advertising”.
The combination of an unrivaled large-scale audience that is becoming harder for advertisers to reach, increased fragmentation among publishers/broadcasters, and a need for improved and innovative technology are exactly why programmatic TV is so intriguing.
Programmatic TV combines various aspects of programmatic advertising with the ongoing evolution of what constitutes “TV” in the modern era. Both these topics are subject to debate. For the sake of brevity, I’ll adopt TubeMogul’s definition; programmatic TV is the use of software-based systems to plan, buy, measure, and optimize television advertising.
Contextual vs. Addressable vs. Automated
The conversation about programmatic TV involves at least three rather distinct concepts: contextual targeting, addressability, and automated processes.
Contextual targeting, or audience-based buying, involves buying ads on a combination of networks, geographic locations, and dayparts which make it more likely that the ads will be seen by a specific target audience. Contextual targeting is usually done by buying ads on a certain group of TV shows more likely to be watched by that audience. While this has been happening for quite some time on a basic level (ie. an advertiser buys spots during sports programming to reach 18-49 males), using additional data sources lets advertisers target shows to a more granular audience. For example, those shows whose audience is more likely to visit a restaurant or retailer over the weekend. It’s estimated that 95m US households can be reached through contextual targeting.
Addressable TV is the ability to target ads at the household level through cable and satellite providers’ set-top boxes. Using a combination of cable company consumer data and third party data, advertisers can target ads to specific households matching a sought-after profile. In 2015, about 40m set-top boxes, or about a third of US households, were addressable. Although significant investment will be required by operators to substantially grow the addressable base, some providers are already pushing forward with the opportunity, with AT&T Adworks hosting addressable upfronts instead of a traditional TV upfront in 2016.
Automated refers to the computerized automation of some aspect of the advertising transaction. Neither contextual targeting nor addressable TV are inherently automated. Advertising utilizing both approaches can be planned and bought manually the old fashioned way, at upfronts and scatters. Automation, as it exists today in the context of TV, is used in planning, pricing, or targeting, but the ability to actually fully execute a purchase through software does not yet exist. The automated ecosystem that’s seen in the digital world – real time purchasing, open auctions, and machine-handled ad-insertion – is not happening in the TV space. For various reasons, many in the business argue full automation will never happen. “TV doesn’t really need automation, because it’s not time consuming to buy. It’s fairly easy to buy a billion dollars of TV time,” says Starcom Mediavest Group EVP Tracey Scheppach. But programmatic TV, defined as some combination of contextual targeting, addressable TV, and automated transaction, is already happening.
Small Size, Big Potential
One characteristic that programmatic TV shares with its digital namesake is the variation in market sizing estimates. eMarketer estimates that programmatic accounted for 0.1% of TV ad buys in the US in 2014, or roughly $68.5m. International Data Corporation, which includes TV ad exchanges, TV ad-targeting services, and connected TV-related services in its definition of programmatic TV, forecasts the segment to be worth close to $1bn in the US in 2016, which is still less than 2% of total TV ad revenue. Magna Global, whose definition is all spending transacted through a technology platform rather than a traditional insertion order, says programmatic TV was a $2.5bn business in 2015, a figure that’s expected to quadruple to $10bn by 2019. Programmatic TV remains a small part of total TV ad spend, but one that broadcasters and advertisers are embracing at an accelerated pace. The following chart highlights some recent programmatic TV developments.
Challenges & Opportunities
The recent programmatic developments by some of the major players in the TV industry are certainly encouraging. However, programmatic TV is in its early days and still faces major structural barriers and challenges to widespread adoption. These can be grouped into Technology and Human factors.
Technology factors revolve around a lack of, or the unsuitability of, infrastructure. Current programmatic solutions are fragmented and difficult to integrate into existing works flows and systems. Additionally, in order for programmatic to reach more TV viewers, there is a need for significant investment in technology to facilitate the increased reach of contextual targeting and addressable TV. Investment will also be required to further develop process automation when it comes to purchase execution, and eventually, machine-based ad insertion. These investments will require significant financial resources and will not happen overnight.
The main human factor is that programmatic TV is currently perceived as a threat by some major players in the TV industry. Some content owners and broadcasters in the ecosystem are not fully on board with programmatic TV due to concerns that the technology will disrupt a business model they are currently content with. There are also concerns that programmatic TV will lead down the path of devaluation and commoditization of ad inventory, a common criticism of digital programmatic. The result is a lack of premium inventory available programmatically. Finally, and slightly paradoxically, programmatic TV is very labour intensive. The integration of various data sources used for contextual and addressable buying involves significantly more manpower than a traditional purchase based on age/gender demographics. The lack of human resources with the required skills in this domain compounds this issue.
While the challenges may appear substantial, the opportunities of programmatic TV should prove attractive enough to spur wider adoption in the industry because both publishers and advertisers stand to benefit. There are three main benefits. First, although this is not the case today, the advancement in programmatic TV with respect to the automation of tasks and processes will eventually lead to operational and transaction efficiencies for both sides. Second, improved audience targeting is something both advertisers and publishers stand to benefit from. Advertisers can optimize ad targeting and focus ad spending on specific audiences, leading to increased returns on ad dollars. In turn, publishers will be able to extract extra value for high-value and niche inventory. Third, with the ongoing shift in media consumption patterns, advertisers are increasingly looking at campaigns from a holistic perspective, instead of separate silos for each media type. In the video world, this means cross-screen targeting and advertising. Programmatic TV solutions would help align TV buying, measurement, and results tracking with current practices in the digital world.
Strategic Repercussions of Programmatic TV
Programmatic TV is already happening in contextual, addressable, and automated ways. The continuing adoption and evolution of programmatic TV will present numerous strategic issues for the entire industry. The four strategic implications of programmatic TV that industry managers should be preparing for are the shift away from ratings and towards audience profile; the futility of trying to stop automation; the entry of new players into the ecosystem; and a future with programmatic content.
A Change in Currency: Programmatic TV will usher in a change in the currency of television. As Videology’s North American Managing Director Tim Castree puts it, “you can say the shift is from traditional TV to programmatic TV, but the underlying economic value creation is about the shift from demographically transacted inventory to audience-transacted inventory, regardless of the screen type”. Put another way, programmatic TV changes the currency of TV from ratings to audience profile. This will have significant repercussions on the way the entire TV content ecosystem operates. Currently, shows are judged on ratings and ads are sold based on those ratings. While some demographics are more valuable than others (viewers aged 18-49 for example), the demographic segments remain quite broad. A prime-time show popular with the 18-49 demographic may be most valuable now. However, for some advertisers, an even more valuable demographic is wealthy individuals concerned about global economics, a demographic who may be better reached at 6am on a business network. While that 6am show’s current ratings don’t support the premium value of the ad inventory, sales based on audience profile would change that equation. The currency shift from ratings to audience profile will impact everything from the type of programming that gets created, to how schedules are made and what stays on air, to the profitability and sustainability of channels. Managers should start thinking now about what the shift means for their firms.
The Futility of Trying to Stop Automation: One of the concerns the industry has with programmatic TV revolves around the belief that programmatic leads to the commoditization of ad inventory. That is one reason many of the programmatic TV implementations to date have tried to compartmentalize or segregate the type of inventory available through programmatic means. This has led to a lack of supply of premium inventory available programmatically and held back the potential of programmatic TV. Attempting to stop the move to automation is futile. As more and more ad dollars are shifting to digital channels, buyers will increasingly expect the technology-driven work flows, audience targeting options, measurement capabilities, and reporting systems that are present in the digital world. Buyers will demand TV advertising adopt these features. We are already seeing more and more technology and inventory making its way into programmatic. The attempt to prevent automation from reaching some segment of ad inventory will not be successful. It is akin to telling your customers that they can search for and book domestic flights online, but have to go to the travel agent to do the same for international flights. Instead of fighting a losing fight, publishers/broadcasters should instead make a push to innovate and embrace technology and automation.
Programmatic Attracts a Crowd: Programmatic advertising in the digital world meant a substantial rise in intermediation. A large number of ad tech companies entered the ecosystem and, as a whole, the ad tech firms now capture more than 50% of the value from digital programmatic advertising. While the story will not likely play out in exactly the same manner for programmatic TV, new players will (and have) entered the ecosystem and will take their share of the pie. Additionally, apart from Google’s failed bid to push into the TV ad space, the TV industry has largely escaped any attempts by tech-titans to invade its advertising turf. That may not always be the case. If the future of advertising is a future of user-data backed transactions that are audience-profile targeted and platform agnostic, then the competition of tomorrow could just as well be Amazon and Facebook as CBS or NBCUniversal. The established firms in the industry would be well advised to try to acquire as much of the future intermediary and critical capabilities as possible now, in order to keep as much of the value capture in house in the future.
Programmatic Content: Programmatic content is the moonshot. Assume for a minute that a few years in the future, the technological constraints of programmatic advertising have been solved. Almost all homes are addressable, real-time ad insertion has been implemented, and companies have been able to make significant strides in pairing offline and online IDs (withholding comment on the privacy impacts of that for the sake of argument). Given that environment, is it that far of a reach to think of TV in terms of mass customization? If we can target ads at audiences most receptive to them, why not product placements within shows, or native ads, or even parts of the shows themselves. What if producers knew that one joke resonated well with certain audience segments and a different joke did better with other segments. Programmers could make sure you would see the joke that would make you laugh. Programmatic content is a distant vision for what could become of TV. It may never be technically possible or sufficiently monetize-able. It may not even warrant much management attention at this time. But it does sound like an intriguing future vision for a medium from the past.