Monthly Archives: August 2012

Why Media Companies Are Being Eaten By Tech Companies

By Eric Picard (Originally published on AdExchanger.com, August 20, 2012)

My friend and colleague Todd Herman (LinkedIn) once wrote a strategy paper about video content when we worked together at Microsoft. Called “Don’t be food,” it was a brilliant paper that laid out a strategy for effectively competing in a world where content is distributed everywhere by anyone.  I love the concept of “Don’t be food.”  It applies to so many existing business models, but clearly where Todd initiated it – Media – it applies incredibly well.

The media business is being forceably evolved through massive disruptions in content distribution. In the past, control over distribution was the primary driver of the media model. Printed material, radio and television content required a complex distribution model. Printing presses and distribution are expensive. Radio and television spectrum is limited, and cable infrastructure is expensive. Most media theory and practices have been deeply influenced by these long term distribution issues, to the point that the media industry is quite rigid in its thinking and cannot easily move forward.

One of my favorite business case studies is the Railroads.  Railroad companies missed massive opportunities as new technologies such as the automobile and airplanes began to be adopted. They saw themselves as being in the “railroad business,” and not the “transportation business.”  Because of this they lost significant opportunities and very few of the powerhouse companies from the rail era continue to exist.

In media, new technologies have been massively disrupting the space for more than a decade. And there is an ongoing debate about technology companies stepping in and disrupting the media companies. Google is a prominent example, and its recent acquisition of Frommer’s is yet another case where it has eaten a content company and continued to expand from pure technology into media.  But Google isn’t moving into media based on the existing rules that the media companies play by – it is approaching media through the lens of technology.

But this issue doesn’t only pertain to the oft-vilified Google: Amazon continues to disrupt the book industry by changing the distribution model through the use of technology, and is clearly gunning for magazine, radio and video content as well.  Microsoft is changing the engagement model and subsequently the distribution of content to the living room via its ever-expanding Xbox footprint, and is broadly expanding toward media with Windows 8, its new Surface tablet devices and smartphones – again using technology.  Apple has turned distribution models on their ears by creating a curated walled garden of myriad distribution vehicles (apps on devices), but charges a toll to the distributors – again using technology to disrupt the media space.  Facebook, Twitter and social media are now beginning to disrupt discovery and distribution in their own ways – barely understood, but again based on technology.

Existing media models are functionally broken – and will continue to be disrupted.  Distribution is always a key facet of the overall media landscape, and will continue to be.  But as distribution channels fragment, and become more open, the role that distribution plays will radically change. Distribution is no longer the key to media – it is inherently important to the overall model of media – but it isn’t the key.

Technology is the key to the future of media. Technology can and has profoundly changed the way content is distributed, and will continue to do so. The future of media is wrapped up in technology, and this is an indicator of why technology companies are eating media companies’ lunches, if not actually consuming them in their entirety.

Media companies don’t understand technology because they are not run by technologists. And there is a vast gulf between the executive leadership of media companies and the needs to understand technology. Every media company should be running significant education efforts to pass along the concepts needed to compete in the technology space, but I’m not convinced even that would be enough to fix the problems they face.

At Microsoft I once had an executive explain to me why most of the executives running businesses at the company were from a software background.  He said something along the lines of, “A super smart engineer who can wrap his or her head around platforms and technology issues can probably learn business concepts and issues faster than a super smart business person can learn technology.”  And he was right – it’s that simple.

Business schools should have requirements today for anyone graduating with an undergraduate or graduate degree to learn how to write software, and most importantly to develop a modern understanding of platforms. These platform models are the future of distribution, and are barely understood even among many technologists. The modern platform models used broadly on the Internet and to create software on devices that drive content distribution are relatively new, and are frequently not understood by people with technical backgrounds who haven’t spent time working with them.

Bad business decisions continue to be made by media companies because of the significant lack of technology leadership in both executive and middle management. As technology evolved, the model for many years was that business people figured out “Why and What” to build and “Where” to distribute it, and engineers figured out “How and When” something could be delivered.  Great technology companies break down the walls between Why, What, How, When and Where. Engineers have just as much say in all of those things as the business people. Great technology companies don’t treat engineers and technologists like “back room nerds.”  They recognize that engineering brilliance can be applied to the business problems facing them, and that technology innovation will drive their businesses to disrupt themselves toward future success.

Media companies must evolve away from their historical strengths based on distribution control, and must embrace technology as a key principal.  And they need great engineers to do so. The problem is, great engineers won’t work for mediocre engineers. Great engineers won’t take bad direction from people they don’t respect, especially business people. And many media companies have treated their existing engineering organizations as an extension of traditional IT models, with mediocre engineering talent endemic in their organizations – frequently top to bottom.

Let me say again; great engineers will not work for mediocre engineers.  This means that the existing CTO and entire engineering infrastructure within a media company will not solve this problem. Before moving forward, executive leadership has to recognize that it is likely that their existing technology organization will fight, block and actively try to sabotage any efforts created outside of their own infrastructure. But it is very clear that without a significant change here, these companies are doomed.

For a traditional media company to compete effectively with Google, Amazon, Apple, Microsoft, Facebook, and the thousands of hot startups now competing with them, they must build world-class engineering organizations. This isn’t a light fuzzy requirement, it’s a core fundamental of their ability to survive for the next century.  These companies must evolve forward.  They must find ways to empower internal disruption.

Media companies must build startup organizations within their own internal structures that are isolated from the existing IT leadership and given bold broad empowered charters with the leeway to disrupt other teams’ businesses.  They must build a new technology driven culture within these large media companies that is separate from the existing groups, and then embrace those internal startups as the future of the company.  This isn’t easy.  It’s nearly impossible.  And this very likely will not work the first time it’s tried. But if media companies don’t commit to this kind of change, they are going to be eaten.

Advertisements

How publishers sell ad inventory

By Eric Picard (Originally published on iMediaConnection, August 09, 2012)

Ad inventory is typically broken down into four buckets: sponsorships, premium guaranteed, audience targeted, and remnant. Each of these buckets can be sold through a variety of sales channels.

Revenue distribution across this “layer-cake” inventory model flows downward — with the vast majority of inventory coming from premium and a significantly lower amount of revenue coming from the remainder:

The process of an advertising sale begins with the media buyer, who sends a request for proposal (RFP) document to numerous publishers. These RFPs typically are written in prose and define the overall goals of the advertiser in question, and of the specific campaign being executed. A typical RFP has between 50 and 100 elements that are laid before the publisher as acceptable or desirable outcomes, and these elements (attributes or attributes of the buy) are generally descriptors of the audience, of the media the advertiser is looking to run on, of the acceptable (and unacceptable) content to be associated with, etc.

Advertising inventory is the base unit sold by a publisher to an advertiser. It is measured in “impressions,” which are defined as an opportunity to show an advertisement to a person. Impressions at their most basic are blank vessels made up of opportunity. Inventory is generally defined in advance by the seller based on a variety of factors, and it is these predefined impressions that are contractually agreed up on between buyer and seller.

Nearly all impressions sold are made up initially of one or two media attributes based on content association (e.g., MSN>Entertainment or MSN>Entertainment>Celebrities; Yahoo>Autos, or Yahoo>Autos>News). Or they’re sold just based on category — in some cases blind, meaning without the knowledge of which publisher the impression ran on. Further refinement of the inventory is based on other attributes such as above the fold, rich media units, or a variety of quality scores. Additional media attributes included in the definition of a piece of sold inventory include various types of targeting and other types of intelligence and filtering such as inventory quality scores and contextual targeting.

Beyond media attributes, there are numerous audience-based targeting attributes available for the buyer to request, or for the seller to offer. These include such attributes as geographic, demographic, psychographic, behavioral, etc.

It is the combination of these various attributes that define the inventory that is sold. Inventory is sold in a number of ways, including on a guaranteed basis (a buyer contracts with a seller for a fixed volume of inventory between specific dates) and on a non-guaranteed basis (if inventory is available that matches, it will be sold, but the seller doesn’t make any guarantees on volume).

In order to predict how much inventory will be available, publisher ad platforms need to look at historical data with seasonality and apply some very sophisticated algorithms to make a guess as to how much inventory will be available during specific date ranges. These “avails,” as they are called, become the basis for how all guaranteed ad sales are done.

But ad inventory has many very complex and difficult-to-predict issues that are endemic to the problem — the problem of predicting how many impressions will exist in a specific month is sort of like imagining how many cars will cross the Golden Gate Bridge in a given week. Predicting this based on historical data isn’t too hard. And predicting the color of the various cars that might cross the bridge is probably feasible with some degree of accuracy. Maybe even predicting the general destinations of the cars crossing the bridge is possible. But trying to predict how many red Toyotas driven by women with an infant in the car who have red hair and who make more than $125,000 annually is probably not a solvable problem.

This is akin to the requests given on a daily basis regarding ad targeting. This type of prediction is extremely technically challenging; nobody has been able to accurately predict how much ad inventory will be available in advance for more than three to four targeting attributes in advance. Therefore, publishers rarely will sell inventory that contains more than three to four attributes because this causes an immense amount of work during the live ad campaign for the publisher’s ad operations team. (They must monitor ad delivery carefully and adjust numerous settings in order to ensure delivery of the campaign.)

Inventory is sold within a contract called an insertion order (I/O), and each sold element is typically called a “line item” on the I/O. Line items correspond to a variety of attributes within the publisher’s inventory management systems. A simple example would be MSN>Entertainment. But a more complex example would be MSN>Entertainment>Women>18-34.

Beyond a typical guaranteed media buy, there are several other mechanisms for selling ads. Some ads are re-sold by a third party such as an ad network (examples include Collective Media, ValueClick, Advertising.com, etc.). Some ads are sold through an automated channel such as a supply-side platform, or SSP (examples include Rubicon, Admeld, PubMatic, etc.). There are also ad exchanges that can sit in the middle of all the transactions, and as the industry has matured, the difference between an exchange and an SSP has become less clear. These exchanges and SSPs then create a marketplace that allows ad networks and various demand-side platforms (DSPs) to compete for the inventory in real time. We’ll refer to this as real-time bidding (RTB) even though in some cases this term doesn’t apply exactly.

The management systems for buying RTB inventory are generally called demand-side platforms (DSPs). In RTB media buys, it is extremely rare to have more than three to four targeting attributes (just like in guaranteed media buys), not because of prediction but because inventory that exists for each campaign or line item that contains more than three to four attributes delivers with extremely low volume. In fact, the amount of inventory available on a per-impression basis as you layer on more targeting attributes generally drops significantly with each new attribute.  This means that a typical line item for an RTB campaign would look very much like the one for a guaranteed buy: Entertainment>Women>18-34.

For a DSP to spend an entire media buy at more than four targeting attributes, the buyer would have to manually create hundreds or thousands of ad campaigns that each would then be manually optimized and managed. It isn’t actually feasible to do this at scale manually.

In summary
In a perfect world, advertisers would be able to find all available ad inventory that matches their goals, with as many attributes as exist on all impressions. The problem is that existing inventory management and ad serving systems are not designed to deal well with more than two to three concurrent targeting attributes, whether for guaranteed media buys or RTB.

So why do advertisers and publishers prefer to sell ads on a guaranteed basis?

Inventory guarantees serve several purposes. The most critical is predictability; media buyers have agreed with the advertiser on a set advertising budget to be spent on a monthly basis throughout the year. They are contractually obligated to spend that budget, and it is one of their primary key performance indicators. Publishers like to have revenue predictability as well, which is solved by selling a guarantee on volumes for a fixed budget.

For all the innovation in the ad-tech space over the last decade, it’s fairly impressive that very few of the core problems of a publisher have been solved. At the end of the day, 60-80 percent of the revenue that publishers bring in comes from their premium inventory, sold on a guaranteed basis — which represents generally less than half of all their available inventory. Nearly all the ad technology innovation in the last decade has focused on what to do with that other half in order to raise the median price of that revenue from nearly zero to a bit more than zero.

It seems to me that there is an opportunity to focus on something else. (And you might imagine that I’m doing just that.)

How real-time bidding works

By Eric Picard (First published on July 19, 2012 on iMediaConnection.com)

The real-time bidding ecosystem is still fairly new, and for many in our industry, there are a lot of misconceptions about how all the different parts of the ecosystem fit together. I’ve had a lot of requests from folks in the industry to explain how RTB works, and how the different players in our space fit together.

The biggest concept to get your head around with real-time bidding is the concept of programmatic buying and selling. The idea here is to streamline the buying and selling process by removing humans from the transaction. Now this is a very important thing to understand: By “the transaction,” I don’t mean that buyers and sellers no longer interact, or that there’s no role for sales in the equation. I simply mean that the act of booking the buy — let’s call it “order taking” — is completely automated. Ultimately this is a good thing for sales teams, as it lets them focus on building the relationship and selling the buyer on the value of their publisher brands. It lets the seller step away from the order-taking process.

Programmatic buying and selling is absolutely the future of this industry; it’s just a question of how long that transition will take. The lower cost of sales for publishers and more efficient buying for media agencies absolutely will make up for any hit to average CPM. And many (myself included) believe that we’ll actually see higher CPMs as a result of all this streamlining. Today most of the inventory that’s available is remnant, and it’s not the high-quality premium inventory currently handled by sales teams. Ultimately, all inventory will transact programmatically. But, like I said, sales will still play a very important role.

At the center of the RTB ecosystem are the ad exchanges. These platforms allow all the various players in the ecosystem to share supply and demand and create liquidity in the market. Examples of ad exchanges include Right Media, the DoubleClick exchange, AppNexus, and many others. On top of these pure-play ad exchanges are many ad networks and supply-side platforms that have essentially built ad exchanges on top of their existing products. The lines get very blurry between the “pure play” ad networks and the other aggregators of inventory that make that inventory available programmatically.

Similar to how stock or commodities exchanges allow inventory to be transacted upon at high volumes with maximum liquidity, the advertising exchanges play that central role. But it’s very important to understand that, just like in the financial services world, the big revenue opportunity is not with the exchange; it’s with brokers representing buyers or sellers. It’s these brokerages that represent the bulk of the value and that pull away the highest percentage of the transaction costs.

The equivalents of brokers in the RTB space are the ad networks, the supply-side platforms (SSPs), and the demand-side platforms (DSPs.) All of these ecosystem players have important roles and provide value. However, it should be noted that the lines are beginning to blur throughout the ecosystem. I predict that in the next few years, many DSPs will roll out SSP services, and many SSPs will become full-fledged ad exchanges. (But more on this in another article.)

So let’s follow the ecosystem participants from start to finish:

The impression starts with the consumer and runs through a web browser. (I didn’t put device in here, but note that even on mobile and tablets, there’s a browser involved.) The impression moves over to the publisher, through some SSP (or ad network) to the ad exchange, and then through to the DSP that is managed by an agency trading desk team on behalf of an advertiser.

There’s nothing very sophisticated about what I’ve drawn here — but note that this is the simplest way I could draw the RTB ecosystem. Here’s another view:

Note that even this is a simplified view, and that many of the various partners can service numerous blocks in the ecosystem. At the end of the day, the RTB ecosystem is made up of dozens of players (possibly hundreds), and they’re all scrambling to figure out their business models. This new ecosystem is definitely the future, but how all the pieces will ultimately fit together is still being determined.

The important thing to note in the RTB space is that from the moment consumers visit a web page, the entire transaction of selling and delivering the advertisements to them takes only a few hundred milliseconds. And this is where the revolution plays out; the competition over those impressions plays out in real time. The best ad for monetizing that user is theoretically shown, and the highest yield for the publisher is achieved. Thus, it should make the ad ecosystem function much better.

But there are many changes that have to take place, and I believe we’ll see it happen. First is that publishers need to push more and more of their premium inventory into the RTB environments. Publishers can make use of almost any ad exchange or SSP to create a private exchange where they can define advertiser-specific or agency-specific terms that are negotiated in advance, and the transaction simply makes use of the RTB infrastructure. Terms with specific advertisers can be reached in pre-decided negotiations, and the transaction takes place through the RTB infrastructure.

In a nuts-and-bolts summary article like this, I’ve glossed over a lot of the nuance and details, and I’m sure we’ll hear from a few parties about what I’ve missed or how I’ve not quite explained this correctly. But I welcome the dialogue. In the RTB space, I think there’s a lot of focus on the details, and not a lot of high-level framing going on — which alienates some of the industry folks who are looking to participate but haven’t dived in yet.