Ad Tech's One-At-A-Time Architecture Is Missing Its Other Half

Anywhere there is advertising, there is a story that needs to be told to a group of people. The revolution and evolution of digital media since the early days of the internet was driven by a mantra of “One At A Time.” No longer would advertisers be bound to large blocks of audience in take-it-or-leave-it transactions with television broadcasters, cable companies, newspapers, magazines, and radio. This was a true revolution for advertising.

Real-time bidding enabled buyers to individually select and price every pair of eyeballs they buy. To achieve this new vision, an entire infrastructure was put in place that dwarfed technology for transacting traditional media. We have seen multiple unicorn IPOs in digital, while technology for traditional media languished. But for many years, the secret everyone knew was that the money flowing through the traditional systems was much larger than the money flowing through digital systems. But, with nearly a decade where all media spending growth was happening in digital, the divide closed. Today, ad tech is waking up to the green field opportunities in transacting traditional media.

Annotation 2019-02-25 111052.jpg

Unfortunately, the green field of traditional media is on the other side of a vast canyon. Getting to the other side of that canyon requires an ability to transact blocks of impressions, before they show up. This is a key reversal of how these markets and technologies for addressable media work. A one-at-a-time market requires impressions (supply) to exist before the buyer presents their demand. Conversely in traditional media, where blocks of impressions are sold, buyers first present a large block of impression demand which the seller then answers.

For those that have attempted to use the one-at-a-time architecture to sell large blocks of audience, Google TV Ads and Google Radio Ads proved exemplars that no amount of money will let you fit that square peg in that round hole. In a one-at-a-time environment the way demand overlaps inventory is explicit. Impressions actually exist and all the demand (bids) for each are collected in an auction. A one-sided second-price auction. In this type of auction demand only exists if supply is offered.

When large blocks of audience are offered, inventories must be forecast, and the way demand overlaps inventory must be computed. That means sellers must be able to understand how an individual bid is applied to inventory as well as how that inventory is impacted by all other bids. In a one-at-a-time market the auction knows all the bids and the inventory being considered is only a single impression, so we know exactly how the bids overlap, they all do. In a one-at-a-time market there is no chance that there could have been a better deal. Conversely, when large blocks of impressions are sold, the seller must always consider what other deals are being forgone so that this deal can be made. These sellers are always questioning if they took the right deal. They are constantly evaluating inventories, current deals, historical deal activity , and pricing.

The technology to answer these questions and respond to demand is not offered by any company other than us. This is the missing other half of the ad tech stack and the reason that today’s ad tech is still restricted to media that is device addressable.


Linear TV Audience Standards For Supply And Demand Are The Wrong Solution

Originally published on Adexchanger

Unlike addressable media, linear TV publishers cannot sell single impressions or single units of attention. Due to its linear nature, TV publishers must sell mixed groups of audiences while buyers only target a specific segment.

The way buyers and sellers define value is different. Therefore, linear TV transactions are inherently more complicated than addressable media. There is no common language for defining value.

Buyers and sellers also do not speak the same language. Linear inventory sellers’ workflow focuses on what they can traffic and what they can sell: spots. Buyers want to tell their story to a specific group of people. This is not addressable media – this is linear media. As such, the trafficking systems “think” in terms of breaks and spots, but that is not exactly what buyers want to buy.

This language-difference problem is why some advocate for a standard definition of audiences. But I would argue that creating audience standards is not the solution that should be implemented by a modern market. A standard definition of audience is essentially a fixed, logical system.

Nearly all buyers already have very good processes and technology for audience definitions, as they have been buying digital audiences for nearly a decade. Sellers are further behind: They only have audience forecasting but lack audience-based inventory management, audience-based pricing and audience-based trafficking.

This reminds me of the early days of Yahoo vs. Google. Yahoo was a manually curated portal of topics, a standard organization created by humans. In comparison, Google realized that the way to organize the internet is by allowing it to define its own organization based on an internal voting mechanism. Google search results constantly changed as the internet changed. Like the Yahoo portal of the 1990s, a curated set of slow-changing and limited definitions will quickly hinder scale.

Standards mean centralized and external control of an inventory definition, which sellers must use to value and price their inventory. Linear trafficking systems can’t manage audience delivery. In other words, sellers would be expected to shoulder nearly all the new risk. Linear sellers are years away from the processes and systems required to enable linear and audience-based deals simultaneously, with any real automation.

If we try to approach the problem from a completely different angle, we may see a solution already exists. What linear TV buyers need is a great “audience search engine,” so they can find the linear media they want to buy without being forced into specific audience segment definitions. The translation of audience to linear demand is precisely what platforms such as 4C and VideoAmp are designed to do: to be that “audience search engine.” Buyers want to control how their audience data is translated into linear spots and sellers don’t want to manage linear trafficking by audience.

The Five Problems Holding Back Programmatic Linear TV

Liquidity flow.jpg

Over the past few months, we have been fortunate to strike deals with virtually every major Programmatic TV vendor. Some, like 4C Insights, and VideoAmp, we can mention, and some we can’t, yet.

While lighting up liner TV liquidity, we have identified key issues that are critical to creating a vibrant market. These issues focus on the market scaling challenges facing liquidity providers, buyers and sellers, that create obstacles to market liquidity and deal flow.

Inventory availability:  Access to granular inventory

Today’s sell-side programmatic linear technology is generally restricted to a small number of inventory products. Solutions in the market in 2018 have achieved a minimal level of automation by restricting transactions to a very small number of products.

For example, one network family, with eight networks, is only selling 40 daypart products programmatically but they have thousands more offered by the sales team. This issue is far more pronounced for large network families that have tens of thousands of products that are sold at differentiated pricing, based on agency and client. Such a product catalog translates to millions of price and product combinations. So, here is a simple question: is a system that was built to only handle dayparts and one rate card scalable to support an entire network family’s deal flow? We think not.

For most buyers of TV media, the inventory they want to buy is simply not available programmatically. If it is, the inventory is likely priced using someone else’s rate card, at a higher price. Most of the programmatic linear technology for publishers in the market today is just a proof of concept. It works, but everything will likely have to be rebuilt from the ground up, to handle real scale.

To increase the availability of inventory, publishers need a transactional platform that will allow them to present any sell offer they choose, to any private market or client, at the price they choose, validated against real avails. For publishers, this eliminates any fear of arbitrage or overselling, and creates an incentive for sellers to maximize the inventory they expose to buyers.

Programmatic Non-Preemptible Inventory: Deciding which deals to take or walk away from

Trafficking systems only understand inventory in the context of the deals that have already been closed. These systems don’t see or accept buy offers for non-preemptible deals. Trafficking systems use their placer algorithms as a preemptible inventory auction. In reality, placers are only designed to maximize revenue from existing deals. So, they are incapable of helping to decide which non-preemptible deal to take. For upfront and scatter deals, the trafficking system does not “see demand” until the deal is already done, and the inventory has been committed.

Linear TV publishers need a system to help to automate the answer to “should I take this deal?” (digital publishers have SSPs for that) A system that can access all the available inventory of spots and audience, access preemptible and non-preemptible deals, and understand deal terms (like pod exclusivity, advertiser category, and separation), to produce a single view of all existing demand.

An even greater challenge is yet to come, programmatically managing inventory on an audience basis, where sold audience deals in one segment are accounted for when offering a different segment that has an audience overlap.

Selling Audiences:  Guarantee liabilities that are hard to quantify

While some networks offer a managed service solution for audience buyers, the process is still very manual. People take in an order, run audience forecasts, generate a plan, validate the plan against actual available inventory, update the plan to accommodate booked inventory, and respond to the buy order. In the end, audience demand is translated to linear supply, so the trafficking system knows how to book the deal.

The key challenge in automating the selling of linear TV audiences is understanding the unsold inventory in the context of the existing audience and traditional linear deals. Current trafficking systems don’t understand audiences at all, they are built on a break and slot framework, e.g. 2:02 break with four 30 second slots.

 Dynamic Pricing and Price Discrimination: Matching offer prices to market conditions

Today, linear TV publishers offer inventory programmatically with very little price discrimination. There are no specific prices by advertiser category, no automated discounts for larger buys, no ability to dynamically update prices as inventory sells out, and no automated way to price inventory based on its audience target. 

Today, sale teams review market conditions and advertiser needs to manually adjust prices. Programmatic systems are needed to help sales teams do that at scale.

 Real-time responses to buy orders: Wait days to find out what you bought

If you have had the pleasure of using a DSP for linear TV, you already know that once you send out your buy order, it can take up to 72 hours before you get a response or confirmation. Why? It’s not your DSP’s fault. It’s because linear TV publishers simply don’t have technology that can compute an answer to your buy order and respond.

For media buyers and planners, this means that a programmatic buy, which includes multiple revision cycles, can still take weeks to finalize. It’s “programmanual.” Linear TV publishers need systems that can understand the demand, understand deal terms and pricing rules, to compute responses to buy offers.

Having said all that, it’s simple. Linear TV publishers need a platform that activates their TV inventory with DSPs that are standing in line and waiting to buy. The MASS Exchange platform solves these issues in ways no other platform does. If you would like to learn more, reach out to us at

Scaleable Media Markets - TV's Biggest Challenge

It's time to throw some shade: markets that can't support transactions for both individual units and packages will never present accurate liquidity, will never be tradeable, and will never truly scale. In TV terms, that means supporting upfront, scatter, pre-emptible, local, and addressable transactions, all through a single market.

Markets create value when buyers and sellers can easily find each other and make deals.  In 2018, TV sellers are challenged by setting open market unit rates they are willing to offer to all comers, while buyers who drive the majority of revenue pay very different rates. A small retailer will pay a lot more for a 30 second spot than Coca Cola, because one of them buys a lot more ads.

In the context of a market, big buyers may never find inventory at prices they are accustomed to paying, knowing full well that they can procure the inventory through existing relationships. In traditional TV deal making, sellers sell big blocks of inventory on the basis of a guarantee and left overs are cleared, one unit at a time, by accepting resting buy orders.

Markets that can not support transactions of both individual units and packages simply can't present liquidity accurately. In markets where transactions are priced by the 30 second spot, market data creates an illusion of differences in price. This illusion exists because market data is missing a key component, the transaction size. In reality, a buyer that gets a unit rate 'discount' is spending much more money to get that rate. Moreover, other buyers who are willing to spend just as much would likely get that same rate. The problem is that some platforms lack this type of context (transaction size), causing network families and cable operators to balk at showing the market the unit rates big buyers actually clear at. And you know what, they are right! Coca Cola and Citi Bank get those unit rates because they spend tens or even hundreds of millions of dollars.

It's not that sellers don't want to sell inventory at lower unit rates, it's just that the rate big buyers pay is not available to à la carte buyers. It would be like walking into a super market and demanding the Costco price. Yes, Costco unit rates are lower for napkins and dog food, but you have to buy four giant packs of napkins or 50 lbs. of dog food.  When the size of the transaction is part of the context of transaction data, people know that there are lower unit rates and know that they require large transactions. This can be seen in action on the supermarket shelf. Every buyer knows that the bigger the box of cereal they buy the less they will pay per serving. Unit rates are not a problem for the super market because the inventory is packaged correctly. Anybody can get the lowest unit rate, you just need to buy the biggest package.

Without the ability to address these pricing issues in the market, sellers rightfully look to their technology providers and balk at showing big package rates to à la carte buyers. TV sellers provide volume discounts that existing market platforms cannot accurately handle. Current technologies are predicated on constructing and managing a dizzying array of private deals and  rates that create the kind of complexity we see in real-time display. This complexity has allowed middlemen to come in and extract a massive amount of value from the transaction. None of the TV sellers we have spoken to have any appetite to manage hundreds of private markets and individual rate cards.

An ideal solution allows sellers to present large packages, at lower unit rates, in the market side-by-side with à la carte buys at higher unit rates. In such a market, buyers can choose to buy at the à la carte rate, the fifty unit package rate, or the one hundred unit package rate; it is clear to buyers that lower rates come with larger packages. Further, the à la carte buyer does not get the false impression that there is a better deal out there that he or she is missing. In the same way a supermarket shows the cost per ounce of cereal for each box size, the à la carte buyer knows that there is a better deal to be had if he or she commits to a larger deal. In other words, sellers don't have to artificially segregate buyers into private markets because the market does not create the false impression that that small brand can get their buy order filled at the Coca Cola price.

Current competitor platforms either support a pre-negotiated rate card for each buyer, so that higher paying customers don't see a lower unit price, or supply prices are set unnaturally high. This means that existing platforms fail to capture the reality that demand from a buyer of 10 units and a buyer of 1000 units is fundamentally different. This demand overlaps, but it is not the same. Likewise, a buyer buying a prime rotator and one buying the same rotator with a show exception and pod exclusivity are also not the same demand.

At MASS Exchange, we have the solution: A platform capable of transacting both individual units and packages using a bid and ask open limit order book.

Understanding Programmatic TV - Part 3: Markets

Agencies, network families, and cable operators generally seek programmatic solutions because they believe programmatic implies competitive markets. The primary role of markets is to bring buyers and sellers together, to determine the price of a transaction. That said, if you survey the current crop of technologies branded as programmatic solutions for linear television, you would find they use automation as a cover while implying a competitive market.

In the first two parts of our series, we explored the process and automation of programmatic linear TV. In the third part of our series we discuss how real markets can be supported with the automation programmatic enables.

The primary obstacle to creating competitive markets for linear TV is that buyers' demand is defined in reach and frequency while sellers' supply is defined in spots. Moreover, the management of 'rate cards',which are predetermined prices, perpetuates the illusion of a market.  If markets bring buyers and sellers together to determine the price of a transaction, but all the deals are already priced, then what is the market doing?


If prices are pre-negotiated before they get to the "market" then anything called programmatic is only performing audience optimization on existing deals. The whole point of using markets is to strike the deal in a competitive environment that is good for both buyers and sellers.

There are currently four types of programmatic TV solutions in the market:

Buy-side optimization: allocates spend across media products offered by seller(s).
Sell-side optimization: calculates least amount of inventory to use for each deal.
Deal communications: platforms for people to send deals to other people.
Exchanges: allows sellers to offer and buyers to bid with automated order matching.

Buyers and sellers want to explore supply and demand in the market. If buyers and sellers can see the other side of their deal, set their prices, search supply, and demand, then you have a market. A black box that stores rate cards from sellers or sends buy orders to individual sellers is not a market.

To enable real markets, sellers need to package inventory, price it, and keep orders in sync with inventory. That functionality is missing in all but one of the four types of programmatic solutions. On our platform, the inventory, catalog, and order management tools are all woven together.

Markets for linear TV already work this way, there is much more time to buy and sell the media. A vast majority of TV impressions are not served or filled in real time. eMarketer estimates that addressable TV will reach 4% of overall spend in 2019.  For linear, buyers and sellers need a different market. If the seller has no bids, they can wait. If a buyer sees no inventory they can leave a bid. This trading only works in two-sided markets.

TV media is a negotiated transaction, there is no negotiation in rate card driven systems. In a negotiated environment, each side has the power to change the price at which they are willing to do the deal and the amount they are willing to buy or sell. In two-sided market transactions prices, are set by both sides. Neither side has to do the deal right now. Negotiation is all about dynamic systems. The solution for programmatic linear TV is a two-sided market. Here's why:

Matches bids and asks, the market doesn't determine a deal's price.
Buy and sell orders stay open until matched or canceled.
Buyers can search for supply and Sellers can search for demand.
Depth of supply and demand liquidity can be measured.
Supply competes for demand and demand competes for supply, in the same market.
Market data is deterministic.

Most other technology providers are really not programmatic markets. As an industry we need to have an honest conversation about technology for linear TV. 



Understanding Programmatic TV - Part 2: Automation

This is the second of a multi-part series of posts to dispel some of the confusion and hopefully have a better common understanding of what is going on in the programmatic TV space. In our first post, we compared the differences between the programmatic TV process and the programmatic digital process. In this post we focus on how the process differences discussed in part one drive  radically different process automation needs. Thus the difference in the use of the term 'programmatic' in linear media (TV) vs. addressable media (video and digital.)

While technology providers in multiple media verticals now use 'programmatic' as a description of automation provided by technology, the business challenges that are solved by this automation are very different for linear and addressable television as compared with digital verticals, such as display and video. The easiest way to describe the difference is that in television, inventory is waiting for buyers, while in digital buyers are waiting for inventory. This difference is the result of selling future media placements or audiences as opposed to real-time impressions.


What ‘programmatic’ means to TV media buyers

In practical terms, televisions buyers are buying promises of placements that meet a set of parameters (a package), as opposed to specific individual insertions or impressions. From a workflow perspective, this means that once a deal is done, sellers provide an initial report of how the media may be delivered (a prelog). Following the actual placement, sellers provide a final report of how the media was delivered, to satisfy the contract (a postlog). Sellers can change the specified units the deliver on a contract without breaking the ‘promise’ they made the buyer. In combination with the fact that some orders are guaranteed to run (non-preemptible) and some are not guaranteed to run (preemptible), managing how or if promises are fulfilled is one of the primary areas where ‘programmatic’ creates value for buyers of television media. This is not a problem in programmatic display and video, as each impression is run through an auction and supply outpaces demand.

Since buyers are buying on a forward basis, answering the following questions at scale is critical:

  1. What is available to buy?
  2. How much inventory is available?
  3. At what price is the inventory available?
  4. Can I present my demand if it is at a different price than the supply?
  5. Are deals automatically pushed into my workflow platform?
  6. Once I have a deal, what can I expect to be delivered?
  7. Once the deal was satisfied, what exact inventory was delivered?

What programmatic means to TV media sellers

For sellers of television media ‘programmatic’ solutions are needed to solve the challenges associated with what inventory promises they can make buyers (sell orders), how those promises would be priced by P&I, and how those promises should be managed given the ever-changing landscape of available inventory. In programmatic display and video, ‘programmatic’ solutions focus on maximizing monetization of individual impressions as a minuscule portion of the inventory is sold on a guaranteed delivery basis.

Since sellers are making promises for inventory, answering the following questions at scale is critical:

  1. What is my unsold inventory?
  2. What inventory promises can I make to buyers?
  3. How do I charge the right price for this inventory given this advertiser and agency?
  4. Which demand in the market can I satisfy?
  5. Are deals automatically pushed into my trafficking system?
  6. Once I have a deal, what am I reporting might be delivered?
  7. Once the deal is satisfied, what exact inventory do I report was delivered?
more colors.jpg

Programmatic vs. programmatic

As illustrated in the questions above, the challenges of transacting linear media are very different. While programmatic represents a set of automated functions in video and display, programmatic in linear media represents a whole different set of functions that need automation.

While several technology providers in today's market use the term programmatic about their products, current programmatic solutions only focus on audience optimization of linear buys. Mass Exchange is the only programmatic solution integrated through-and-through from the sell side trafficking system all the way to the buyers’ platform.

The Mass Exchange programmatic solution enables transactions on an audience or unit basis and can support all types of TV media buying on a single platform including national, scatter, direct response, and local. 

Understanding Programmatic TV - Part 1: Process

Recent advancements in the programmatic TV space, and growth of the number of vendors, shows platforms in linear television are finding new ways to add value.  Finding the right terms to describe these new services is tough. So, many have adhered to what they learned over the last few years, which is RTB-speak. While a good starting point, RTB-speak is used to describe non-linear real-time transactions. But, nearly all TV media is linear and non-real-time.

So, there's lots of confusion.

This is the beginning of a new multi-part series of posts to dispel some of the confusion and hopefully have a better common understanding of what is going on in the programmatic TV space. In this first post, we will compare the differences between the programmatic TV process and the programmatic digital process.

Understanding the Term “Programmatic” in TV

Across many professional publications and industry bloggers the term “programmatic” is used to describe technologies that create efficiencies for buyers and sellers of media. The problem is that so many folks have different definitions that a great deal of confusion surrounds this term of art.

The most basic use of the term programmatic is as a kinder and gentler way to say automation. Some don’t like the term automation, because it sounds like replacing people and eliminating jobs. So, many marketing folks in adtech firms use programmatic as a substitute. That is why people in the know call most of the linear TV solutions in the market today “progra-manual.” While some of the process may be facilitated by technology, it is not end-to-end. What’s worse, many programmatic solutions in linear TV simply push the manual work over to the other side of the deal table.

Real programmatic means that inventory availability, pricing, supply offers, demand bids, and transactions are all technology enabled. In other words, inventory is analyzed and priced programmatically, avails presented to buyers are managed programmatically, buyers search the market programmatically, bids are matched to sell offers programmatically, and deals are trafficked into agency workflow systems and TV trafficking systems programmatically.

Linear TV is not the same as digital

The Non-Linear Process

The Non-Linear Process

Let’s start with the digital process. While this may be less familiar to the folks who have done traditional linear media transactions, for those with some familiarity to programmatic digital media, this should look straight forward.

The most important aspect of programmatic digital media, is that the transaction happens after the impression is created.

Once an impression is created, buyers are asked to place a bid on the impression to determine who will show their advertising. When buyers receive the bid request, two key questions must be answered 1. Is the audience someone the ad is targeting? 2. If so, what should I offer the seller to buy the media? If both of those questions are answered, the buyer bids on the seller’s media. The bid enters an auction, along with other bids, and the winning buyer is determined.

With the auction winner chosen, an ad must be served and shown to the individual representing the impression.

Along with starting at the impression, the digital programmatic process also possesses the inherent characteristic of being non-iterative; the process always moves forward in a strict order. This means iterative processes like negotiations cannot be handled.

So, using terms from programmatic digital media like ad server, DSP, and SSP, muddies the waters. While the steps for programmatic digital and programmatic linear TV may seem similar, the way in which the process operates requires a vastly different approach.

The Linear Process

The Linear Process

Changing gears, let’s look at the process for programmatic linear TV. Since individual impressions are not available for evaluation, buyers rely on historical and forecast data to understand where the audience they are targeting will be consuming media. Based on that work, media providers are selected and a “bid” (aka request for proposal, RFP) is sent.

Sellers, linear TV providers, respond to bids with offers of inventory products and pricing. There is no auction and the buyer and seller iteratively negotiate until they reach agreement. It is this call and response process that needs to be accommodated by programmatic solutions for linear TV.

Once agreement is reached and the deal is accepted, buyers and sellers must wait until the media airs for the ads to be served (that’s why that little clock is  there)

While all of this is probably not news to anyone who is on either side of deals in the media industry, what it means is that “programmatic” in linear media is vastly different from “programmatic” in digital media. In reality though, most media buyers and sellers could care less about these differences as long as they are getting good ROI and good revenues.

On the other hand, for those in management roles in charge of selecting vendors and technologies that will impact their ability to generate revenue or buy targeted media efficiently, these differences are critical to understand. Without a full understanding, some folks will choose solutions that sound good, but in actuality will not deliver enough value.

In our next post we will explore some of the ways in which "programmatic" has been used and abused by vendors, to close deals that deliver far less automation than expected.


What we’ve been up to

As you may have noticed it has been quite a while since we last shared a discussion on our blog. On our end, we have been working hard to expand our platform to handle some of the fundamental differences required to facilitate the trading of television inventory.

The fundamental difference is that TV can be transacted in both linear units and addressable units. From a system’s perspective, those are apples and oranges. You could just transact on impressions across both, but doing that in the context of inventory management, yield optimization, and pricing along with the transaction make the impression based approach untenable.  That means systems need to be designed so that the way buyers and sellers transact is abstracted from how the actual inventory management, yield optimization, and pricing happen. For example, in linear TV, yield optimization happens at the insertion level, while in display media it happens at the impression level.

What presents additional complication is that many linear buyers do not buy units one at a time. Buyer price is generally a function of the size of the spend. The more you spend the bigger your unit-rate discount.  That means that packages. No matter how you slice it, by unit or by impression.

In short, there are three fundamental transactional units, instead of one, that all represent the same inventory; the critical difference from digital media.  Buyers are interested in measuring their units of trade in impressions, insertions, or packages. This critical difference is just the starting point.

All three types of transactional units need to live side by side in a centralized platform capable of unified inventory management, yield optimization, and pricing. From day one, our platform was designed in anticipation of this functionality. So, when we expanded to supporting TV clients, our product road map quickly accelerated this development. Also, we have been working on all of those things in the guts of our product that are designed to make our product so “it just works.”

Now, back to our blog…

So, much of the work that inspired our previous blog posts has been overtaken by the unglamerous work of making trains run on time. We are now at a stage where the work we have done yields learning from its application. Our core functionality was based on innovating functionality and creating new ways of solving old problems. Much of our recent unglamerous work has focused on problems of operationalizing our innovation, which is not necessarily innovative in and of itself. For example, QA automation or the automation of scaling instances and spinning up environments.

Moving forward, our topics will be shifting from explaining how media could be transacted more efficiently and the problems with existing technologies. We will be focusing on how media companies and agencies are leveraging our platform to transform their media trading practices, to usher in audience-based and addressable buying, while still having one foot in the linear unit-based and upfront world.

Building On-Ramps to Media Futures Markets

It's been a long time since technologists started building electronic systems for trading forward media . To be fair, it has been at least 20 years since the first companies gave it a shot. In fact, we have a little symbolic graveyard in our office to commemorate our fallen comrades. Some pivoted, some got acquired, some went out of business, and some were divisions inside big tech firms that faded into the background. All made valiant efforts. All failed to build a real exchange for forward media. Let's pour one out in their honor.

Fallen Comrads.jpg

Our fallen comrades who had taken on forward media buying and failed, could not effectively prove to both sides that their platforms and technologies drove greater efficiency and led to better deals.  

So how is it that so much media is bought on a guaranteed basis, but none of it actually trades. As far as we know, no one in the market provides a biddable environment for forward media. To have real markets supply must compete with other supply and demand must compete with other demand in the same market. But, competition is not just the application of a single bid to a single inventory opportunities, it also means the application of multiple bids to multiple inventory opportunities simultaneously. Without this, there is no real exchange, without a real exchange, none of these technologies were able to reach critical mass.

We can ask ourselves if there was a theme among all of these heroic attempts. Is there a common reason that we can point to? Is there one thing we can identify that tripped up silicon valley greats like Google, Yahoo, and Microsoft? We think there was. 

When we look at any of these "marketplaces", some of which still exist in one form or another, we see that they all actually provided ecommerce platforms for media, not exchanges. While that was true, most were marketing their products as if they were exchanges. If a searchable list of products that can be purchased, from multiple buyers,  through a website is a marketplace, than why is doing the same thing over the phone with a sales rep not a marketplace? (Because automated sales wasn't sexy or VC fundable)

Media buying is much more like portfolio management than buying electronics from Amazon. Setting aside the fact that none had an environment in which anything can be traded, none provided the types of solutions that enabled demand from other channels to flow into these forward media marketplaces. In real-time impression markets, pricing and analytics tools facilitated the movement of budget to increasingly targetable audiences. This in turn enabled the measurement of greater efficiency from real-time impressions markets than traditional insertion orders.  In short, buyers and sellers could measure how good of a deal they were getting and simply followed the money. And so, many left the on-ramps to their highways unbuilt. It was too hard to get into the markets, so none got traction. 


I Got 99 NP-Hard Marketing Problems...

We have spent a fair amount of time thinking about why some problems faced by data driven marketers are so much harder than others, advertising's P vs. NP problems. We were looking to find a reason why these hard problems have not been solved.

In short, we realized that there are two ways of thinking about solving data problems, the first is by modeling (think algorithms) and the second was a new idea. The new idea is that there are problems in marketing whose solutions are emergent from the data itself. In other words it can be measured but not modeled. Effectively it means that some problems can only be solved by working out every possibility and measuring which work and which don't. It means that the way we design our data models really matters.

Emergent solutions arise when new properties and behaviors "emerge," with no one directing and no one able to foresee the new characteristics from knowledge of the constituents alone.

In chemistry, for example, the taste of saltiness is a property of salt, but that does not mean that it is also a property of sodium and chlorine, the two elements which make up salt. Thus, saltiness is an emergent property of salt.

17th century economist Adam Smith described an "invisible hand", an emergent property of markets, that guides markets to produce just the amount and variety of goods that the public needs. The stock market has its own "invisible hand." The purely self-interested actions of thousands of buyers and sellers result in the purely blind workings of the stock market—the sudden shifts in activity and valuations, the bubbles and crashes—as well as the market's notorious properties of stupendous intricacy and frustrating unpredictability.

There are some problems in media whose solutions will only emerge from data. The key challenge is usually measuring and identifying the source of the issue. Once a problem can be identified and measured, it’s solution emerges from the data. Patterns can be identified and changes can be made.

Solutions for marketing problems such as attribution are data-emergent in nature. As evidence, look to the current models used for attribution. Frankly, the market has just begun measuring attribution and the models in the market kind of suck. We can’t even model how and why people click on ads, just that they did, let alone understanding why they made a purchase.

To do that we need to change the way we understand attribution data. Books are a good analog to attribution. What if we stored the data for a library in such a way that every book's data contained its translation to every language on earth. Fundamentally, War and Peace is the same story in every language, but understanding all translations at the same time will surely yield understanding that we couldn't get from just the English version. This exemplifies how storing the same basic data in different ways unlocks knowledge that was simply inaccessible before. The old way of storing that information made it invisible. Every language has its own unique way of capturing and communicating ideas. If a story can be told in every language, the underlying fidelity of the ideas in the story are far more well defined. The improved definition is the emergent property in this case.

Many of the problems we face in marketing are as difficult as they are to solve because the way we describe the underlying information needs to be improved. We are simply unable to see a solution because we are unable to identify the source of the problem.

The Dangers Of Scaling Today’s Addressable TV Buying Methods

Originally Published on AdExchanger

For more than a decade, some of the biggest names in television have been hard at work on achieving household-level addressability across their footprints.

Leaders in the space, such as Cablevision, Comcast and Dish, have deployed a great deal of hardware and software to ensure the right creative can be delivered to the right household. But what about the way in which the media is transacted?

Many in the television space have girded themselves for the onslaught of technology they fear will undermine their positions the way display publishers believe early open-market RTB eroded their pricing power.

The problem is that even though there is no real-time bidding in addressable television, the forces that weakened publishers’ pricing power in RTB are rearing their heads in addressable television. But it is not the programmatic or real-time technology that hurts a seller’s pricing power – it is the flow of information.

So, what is it about open-market RTB information flow that is similar to the information flow in today’s manual methods of buying household addressable television? Something is causing this asymmetry of the information flow that will undermine seller’s pricing power, even without any programmatic markets in place. When I shared this insight with some in the industry, the initial response was disbelief. After more explanation, they came to see the trends.

Sellers Left In The Dark

The root problem of the information flow asymmetry is a lack of shared understanding of what is being bought and sold. In the traditional model of linear TV buying, both the buyers and sellers understand the audience and placement being purchased. Both sides know how Nielsen defines the audience and both sides have a general understanding of what is being bought and sold. Therefore, the buyer and seller agreement represents a real market price.

Today’s addressable TV buyers are very different. The buyer leverages data from Rentrak and other third parties to identify a list of target households, which defines the media buy. This is where the information asymmetry begins. A seller that receives a list of households has no idea why the buyer targeted those households or what drove the value of the transaction.

If sellers have no idea why their products are valuable, how can they possibly determine a fair price? But that is how addressable TV media buys are like open-market RTB in display. The buyer knows everything about the product and what defines the value of the deal and the seller knows nothing.

In RTB markets, the publisher has no idea why any impression cleared at the price it did. A publisher has no access to the data the buyer is using to bid. This information asymmetry created a huge advantage to buyers through the use of open RTB markets. It is not the technology that undermined publishers’ pricing power – the technology just leveraged a specific type of deal model, which most onlookers conflated with the technology itself.

Repeating History?

So what does this mean for today’s sellers of addressable TV media? It means that sellers who relinquish control over defining products and packages will be doomed to repeat the long-feared loss of pricing power experienced by display publishers. For addressable TV markets to flourish, sellers will need to use first- and third-party data to certify packages and make them discoverable.

A buyer unwilling to define the target of their buy with the seller knows full well that the value of what they are buying is much higher than what they want to pay. It is a signal to the seller that buyer is only willing to buy the inventory if the publisher sells it below its fair value.

While this may be beneficial in the short term, transaction models that strongly benefit one side of the transaction become overly complex and fail to scale to meet market needs. This has been proven by the movement of display media liquidity from open market RTB to private marketplaces, header bidding and programmatic direct. Display publishers have sought alternative channels to regain lost control.

Continuing down this path for addressable TV media transactions will result in yet another technology and vendor map that is as confusing and complex as display. That’s something I’m pretty sure none of the TV media buyers and sellers want.

A Lack Of Market Forecasting Is Holding Back TV Media Trading

Originally published on AdExchanger

This year will be a great one for digital media trading. Digital media trading has jumped the shark and landed safely. Transparency will increase, fraud will decrease and market structures will only get better. Moving forward, the real challenge is to bring what we have learned to drive the growth and expansion of media trading into all other media and create unified platforms.

At this point, TV media is purchased and not traded. But that could change with the creation of better market-forecasting tools, which would improve our understanding of media markets, enhance decisioning and enable faster, effective and efficient technology-driven TV media trading. The inability to model and understand media markets is a leading obstacle holding back the birth of TV media trading.

Sellers of TV cannot reliably predict how trading may impact revenue flows, while traditional channels are reliable with forecastable revenue. If trading is not provably more reliable and revenue-accretive, there is no reason to switch from purchasing to trading. When sellers forecast higher yields via trading, there will be incentive to actively trade.

Unfortunately, display media is the petri dish where the earliest experimentation happens. Once technologies are created in display, companies try to cram those technologies into other media with little success. The trading technology developed in display was designed for machine-based trading, not people.

Discoverable Pricing, Secret Strategies

The challenge to developing appropriate TV market-forecasting technology is two-fold. The transaction space is highly fragmented, and the asymmetry between buyers and sellers makes the collection of granular data only possible on the buy side. Current technology makes market pricing and trading strategies inseparable, so everyone in the markets wants all of their data secret. Markets operate best when pricing is discoverable and strategies are secret.

As demand aggregators, agency buyers have a significant advantage. Buyers see upfronts, scatter, cable, satellite and local supply, while the suppliers only see the demand for their own inventory.

Broadcasters and networks have few tools to help them understand how to use packaging and pricing techniques to address viewership and market-demand fluctuations. Sellers can see the scatter budgets coming their way and what their sellout is, so they see both supply and demand, but only for the tiny part of the market that their inventory represents.

There is little that data sellers can use to understand granular demand across the market. Given this environment, people use their anecdotal experience and intuition to make up for the lack of data, which reduces the use of machines to add value.

No Solution In Sight

What is even more disturbing is that while the asymmetry exists, the lack of an organized market generating this historical transaction data means that no one is even working on this problem. Buyer and seller transaction information is so closely guarded that general trends are the best we can hope for.

No existing technology accounts for campaign goals, budget, available inventory, market prices or historical performance metrics. There is also no technology that offers recommendations for an optimal media plan.

On the other side, media companies lack the sales tools that enable what-if scenario modeling to understand how the demand they expect should be most effectively allocated across the media. In reality, media plans and media packages span multiple media and publications. That is what the deal conversations are all about, not a single impression or a line item for a placement. Buyers and sellers need a better way to forecast what those conversations will be like and decide the best course of action.

In Use Now

TV media buyers and sellers need better tools for making decisions. For real-time display, the tools are more straightforward, as people could never pull the trigger fast enough, so machines bear the brunt of the control.

For forward-media buying, in which TV is the 900-pound gorilla, people bear the brunt of control, and technology plays more of a support role. While there are some systems that help manage yield after the deal is made, today, the selection and decisioning of what to offer a potential client and the price at which to offer it is nearly completely manual.

With everything we have accomplished, we still face obstacles that will challenge this momentum moving forward. First and foremost among those is the revolution of market forecasting. Unified platforms require unified – not homogenized – tools and analytics to support omnimedia trading.

While media mix modeling has been around for decades, buyers’ and sellers’ ability to transact an ever-increasing granularity of content and audience programmatically is driving an increased need for a revolution of market-forecasting technology.

Momentum is moving in the right direction to help buyers determine what they can buy, how they can buy it quickly and easily and whether it turned out as expected. The technology for modeling what TV media to buy and sell has not kept up.

Everyone Needs To Understand These Media Trading Models (But Few Do)

Originally published on AdExchanger

I have a confession to make: Before writing this, I couldn’t really say that I, or anyone I know in ad tech, fully understood how different ways to trade media compared with each other.

There are all these different ways that technology brings buyers and sellers together to trade media. This stuff is so confusing and hard to wrap your mind around that I never really felt like I had a solid grip on it.

Buyers and sellers choose the trading models they want to use based on their objectives, technology and access to data. Each media trading model fills a niche in our ecosystem. Understanding the different media buying models is not about picking winners or declaring one better than another. That is up to you, given your business needs.

I set out to map these media trading models so that everyone can pick what works best for them and have a far better understanding of what is going on. To understand the different types of media trading models, I spent a few days talking to folks throughout the industry, which made it clear that it was time to decipher this enigma.

I decided to focus the comparison on three areas of a transaction where key business decisions are made by buyers and sellers. These are the handling of inventory before the trade, managing of the negotiation and how the deal’s final price is determined. The outcome of the effort to understand media trading models resulted in a visual landscape for media trading models that I call a “Media Tradescape.”

I designed it to compare 10 key areas where we can differentiate each of the nine trading models used in display media; I give more detail on each trading model below the tradescape. I’ve also included a breakdown of each trading model with corresponding characteristics as part of my “Table of Trading Models.”

What did I learn from my investigation? There is significant upward price pressure from demand within media trading environments. Since the advent of real-time bidding, an increasing number of media trading models have risen to allow buyers to access inventory at increasingly higher prices. Trading models developed in the last couple of years focused on unlocking trades for more effective and higher-quality inventory, at higher prices. The environment for transacting high-quality and high price-point inventory via technology is very much in flux.

You can download a PDF version here.

Inventory availability: Control by sellers to allow buyers to see forward supply. Some models empower sellers to give buyers a searchable supply, others empower sellers to answer requests for specific supply and, in some, sellers can't expose forward supply at all.

Inventory priority: The ability to guarantee inventory amount and price to a seller. Some models focus on reserved inventory, others on unreserved and some on both.

Inventory allocation: The way in which supply is allocated to demand. Once an impression is received by the publisher's ad server it attempts to allocate the supply to demand, based on the outcome of its yield optimization. If no desirable outcome is found by yield optimization, an outside auction is used to determine the outcome.

Inventory units: The definition of what is supplied and demanded. Units of audience are a mix representing all of the consumers of the supplied media, and “impressions” are units of supply that all share a common set of attributes.

Inventory targeting: The ability to apply granular definitions to the supply using a specific set of agreed-upon attributes. Some trading models enable targeting definitions from the supply side or enable use of definitions from the demand side, while others enable both definitions for filtering.

Pricing model: The units of price used by buyers and sellers. Most trading models are based on CPM pricing, while two support multiple pricing models, including CPM, CPC and CPA.

Seller offers: The ways in which supply is represented in each trading model. Some use hidden floors in an auction to set the lowest price a seller will accept. Others use a set of rules to represent offers to buyers and some use manual processes to set the sell offers used for trading.

Buyer bids: The ways in which demand is represented in each trading model. Trading models can be divided into environments where buyers can bid or buy using a take-it-or-leave-it price. Rules-based bidding allows buyers to set the price at which they are willing to purchase the supply. No-bidding environments provide an ecommercelike experience where buyers can only buy at the offered price.

Negotiation: The ability to consider both the seller’s offer price and buyer’s bids in determining the clearing price. Some models do not account for any negotiation since the price is determined manually by humans before technology is involved. Other trading models support a negotiation and the final price is determined by a technology.

Deal priced by: The layer of technology that determines the clearing price. Regardless of how the negotiation occurred, the final price at which each impression clears is either determined by the publisher's ad server, or that decision is outsourced to a technology outside the publisher's ad server.

This is my interpretation of the landscape of media trading models. I want to have a shared understanding of these trading models so let’s have a conversation. Please add your perspective to our shared understanding of the space. I and many others would be very happy if we could clear this up.

Below is the "Table of Trading Models," which breaks down this information by trading model. You can download a PDF version here.

Four reasons why ‘Infinite’ media is a lie

The idea of infinite digital inventory has been batted around as a concept among colleagues for a long time. Every time I hear this, I get pissed. The reason for this reaction is that I know some folks are turning a blind eye to what is actually happening in advertising markets. There are four ways in which the lie of infinite media is created.

In reality, there is too little real supply.  If real supply and demand are in balance, there is very little room for shenanigans.

First, let’s prove media is not infinite. Then, let’s examine why people think it is. Lastly, let’s contrast the idea of infinite inventory with the idea of finite attention. I’ll use the analogy of data transfer. To illustrate, let’s think of inventory as the volume of data and attention as bandwidth. The idea of infinite inventory is like saying you have unlimited data on your internet service. Everyone can effectively download an infinite amount of data, the question is how fast.  If you try to watch 4K video over DSL, you will get 10 seconds of buffering for every 1 second of video. In this same way, the idea of infinite media ruins the attention experience. If a single user has limited attention bandwidth than the sum of all users also has a limited attention bandwidth and therefore media is finite. The essence of media is the content consumed and attention opportunities it creates not the number of ad placements served.

Now, let’s examine the lie of infinite inventory. Let’s start by assuming it is correct, there is infinite inventory. If true, it is important to consider some possible causes.

1.      There are so many users and so many ads served that there is not enough demand to meet the supply. Supply is growing faster than demand making supply effectively economically infinite.

2.      Infinite supply is produced to syphon off known demand: bots and non-viewable inventory.

3.      The accuracy of and transparency of 3rd party audience derived from lookalike models.  Overly broad segment definitions can create the illusion of far more inventory than actually exists.

4.      The way in which the amount of inventory is measured yields a false count.

Now, let’s address these one by one.

In terms of users, yes, the number of users and the amount of time spent online has grown significantly over the last decade. But let’s be real, finding the right audience, in the right environment, at the right time is pretty darn hard, let alone finding an infinite amount of it. Further, supply can be exponentially increase by flooding ads on a page, turning an article into a ten page slide show, and many other tricks that provide for lots more ad inventory crammed into the same attention bandwidth.

In relation to bots and non-viewable inventory, tomes have been written. In reality, this is inventory that provides zero real audience attention. If media really was infinite, than fraud and ad blocking would not be real issues. Blocking ads, in other words removal of supply, is not a problem when supply is infinite. Further, adding supply via bots, should also have no influence on the market as we could simply replace the lost and fraudulent inventory with real inventory. Further, if $7.2 billion of fraud will take place in 2016 as the ANA study finds, it should leave a pretty big ‘mark’ in the marketplace. It is really hard to hide $7.2 billion. If all programmatic spend in 2015 was $14.2 billion, according to Magna Global, there are a lot of folks making a significant portion of their revenue via fraud.  I’m not a tin-foil-hat conspiracy kind of guy, but if some players in the industry are profiting from this fraud there must be many other players who are happy to look the other way. Said otherwise, if you are willing to buy fraudulent traffic, there is an infinite amount.

Overly broad audience segmentation models are also perpetuators of this lie. For example, a couple of hours of research shows that about 6% of US households replace their car every year. Add to that the fact that the buying consideration period is about a month, then only 0.5% of the population is actually in market for a car at any given point in time. Said otherwise, any data provider that tells you that a general interest or news site can sell you auto intenders that represent more than 0.5% of their audience over any period of time, is likely broadening the model to ‘create’ audience inventory. Using lookalikes may yield lots of opportunities, but do you know how many of those auto intender lookalikes are really auto intenders and not just people with similar behavior? What behaviors drive the model? Just because I visited and does that mean I’m going to buy a car?

Last is the way in which we measure inventory. Some buyers and sellers have overcome this measurement problem by transacting on share of voice. Using this model, deals are a commitment for a portion of the overall attention instead of the number of impressions. Measurement by impressions is very easy to manipulate, measuring by attention or share of voice is nearly impossible. If you were choosing between two inventory sources at the same price and you knew that one had 20 ads on the page and one had 4, you would choose to run your ad in the media that was less cluttered. Share of voice or share of attention based deals remove a publisher’s incentive to cram ads and guarantee the buyer is buying the attention that they bargained for.

It’s time to open our eyes. It’s time that legitimate advertisers and publishers take control of the market. Because in the end the lie of infinite inventory only hurts the good guys and makes money for the bad guys.

Analogy Makes It Obvious: The RFP Is Bad News

Technological tillers teach us that sticking too much to old perspectives and ideas is a surefire way to fail. We need to stop thinking of the automated RFP as the product itself, but rather a framework allowing buyers and sellers to access media transactions. If the framework can be avoided while simultaneously creating a better experience for media transactions, then that framework should be avoided.

If technological tillers are ignored new products and technologies fail. For those who acknowledge and solve the problem, it generally  results in a revolution and massive success.

I recently watched a great keynote that I thought applies directly to transacting media and specifically media trading. In  short, the main point is that sticking too much to old perspectives and ideas is a surefire way to fail. The term  for this behavior is 'technological tiller.'Scott Jensen, product strategist at Google, gave this great talk on this subject and came up with this great new term, the  'technological tiller.'  Scott defines a technological tillers as happening when we sticking an old design onto a new technology wrongly thinking it will work.  He derived the term from a tool called a boat tiller, which was, for a long time, the main navigation tool known to man. Hence, when the first cars were invented, rather than having steering wheels as a means of navigation, they had boat tillers.

Cars that used boat tillers to steer were horribly hard to control and prone to crash. Cars could only get widely adopted after the steering wheel was invented and added to the design.

With an understanding of technological tillers, the two main types of media buying, real time and forward,  look very different. For everything that was built using the RTB protocol, there was pretty much no old perspective to stick to when building the technologies and processes, so many problems were avoided and there were no tillers to speak of. For forward media trading, this lens of the past has driven most technology to be developed using the old perspectives. RFPs are a technology tiller. The context and the technology have changed dramatically in recent years.

Worse still, most of the language of media buying and selling technologies was created in the realm of inventory management. Before media could be bought and sold effectively, the very first pioneers of media technologies created technologies such as cookies and ad servers. This first generation of technologies was not developed to support media buying and selling, it was designed to facilitate an understanding of what, and how much, inventory would be available at some time in the future, to track the activity of users for the media owner to provide a good experience, or other non-transactional reasons.

For media trading, this is a valuable lesson:  context or technology changes most often require a different approach. In our car example, the new technology that added a motor engine to a horse carriage needed the new design of the steering wheel to make the resulting technology, the car, reach its full potential.

For forward media trading to reach its full potential we need something different than the RFP. Linda Boff, chief marketing officer at GE, recently wrote a great article Marketers: It's Time to Say RIP to the Media RFP

"...when I think about the many ring-around-the-rosy conversations that characterize the standard RFP process, I literally want to cry. In a post-RFP world, agencies and publishers don't waste time on proposals that will never go anywhere, and brands don't devote resources to sifting through cookie-cutter submissions. Instead, time and talent can be invested where it matters: in identifying breakthrough experiences that are good for users and drive attention -- the only metric that really matters."

The only thing that Linda does not address in this great article is what the solution is. In other words, if automating the RFP process is a technological tiller, then what is the steering wheel solution for media buying? And that my friends is why media futures technology is so very necessary.

How To Fix Media Asset Standards So They Don't Suck

Standardization can seem like a technical topic but it is simply the domain model (the rules that 'govern' ). In the past standardization seemed difficult because the individuality of media assets created an illusion that a standardized domain model for media assets was impossible.

Some think that applying a standard to media would mean forcing a set of descriptions that everything should fit within. Like the way the IAB created a set of context standards that seek to capture all contexts. But what if the standard is not about restricting what descriptions are acceptable but rather about standardizing the way descriptions themselves are created. In other words, any type of descriptor can be added to a media asset, so long as it is done in a way that the system, can understand. Like Mendeleev’s periodic table, the system has a place for newly discovered elements, even though we never knew they existed.

In Chemistry, 19th century researchers faced the same problem. ‘We have all these elements, but how do we standardize our understanding of their respective relationships?’ To solve this problem in chemistry, a new domain model had to be created to bring rational order. While it may have seemed impossible before the problem was solved, afterwards it seemed completely obvious.  Dmitri Mendeleev solved the problem in 1869. His genius was in building a model that simultaneously defined how the elements are similar to each other and how they are different from each other, in more than one way.

Like Mendeleev, media solutions need to conform to a standard that defines how we order the information in the domain, not limit what information can be contained in the domain. On a side note, some of the marketers making periodic tables of marketing are doing it wrong. They don't actually understand why the periodic table is so brilliant.

Most folks in media that we have talked to question the ability to standardize media assets because they can only imagine domain models based on similarity or difference.  The beauty of Mendeleev’s arrangement is that the relationship between and element and its neighbors to the left and right are always the same. The relationship to the elements above it and below it are also always the same. In other words, the elements are arranged such that they show how they are different from each other by showing how they are like each other. Media standards need the same type of domain model.

Periodic Table trends
Periodic Table trends

To build this domain model we have to understand the relationship between two different media assets (two ‘atoms’) and figure out where they belong in the domain relative to each other (their ‘location’ on the ‘periodic table’).  Most media technology was not build to handle transactions, so it did not need to standardize the domain model for media assets, it standardized the communication of transaction requests, e.g. RTB protocol.  For the exchanges and other technologies that were designed for managing transactions, that part of the system was simply ostriched. Why? Because the second price auction does not need to know what is being auctioned to be successful, it simply manages bids and price floors. This was a design feature, not a bug. Since second price auctions are being used to sell an impression in real time, it doesn’t much matter what the impression is, it only mattered who will pay the highest price.

When we want to trade avails (read media futures) we and the auction itself need to understand what is being bought and sold. Since there is so much uniqueness in the domain, we decided to reverse the perspective. Instead of trying to define what makes one piece of inventory unique, we define how it is different from everything else. We can work by inclusion or exclusion, the results are the same. So in technical terms, each piece of information describing a media asset is a vector. Each media asset is a collection of vectors.

In practical terms the ‘standard’ is a way of ordering the information submitted to the system. This means the system incentivizes conformity without demanding it. For example, there can be two competing methods of defining context, but both buyers and sellers have strong incentives to choose what is best for the market. So, if two competing standardization systems will yield the best outcomes, the domain must support both. If one standard will yield the best outcomes, the domain must support that as well. The standards create an incentive to find the optimal solution, the domain does not define the optimal solution.

Like Mendeleev, media solutions need to conform to a standard that defines how we order the information in the domain, not limit what information can be contained in the domain.

PII issues will never go away with real time bidding

Houston, PII has a problem
Houston, PII has a problem

PII issues have long been a point of discussion among us all.  In all that talking and discussing, we never uncovered the root cause of why PII issues are such a dominant force in the current real time bidding market architecture. I propose that taking another point of view at the problem reveals that it is a direct outcome of the market architecture and not a side-effect of some other economic inefficiency. The current market architecture in real time bidding is a ‘call-and-response’ system. One side, the seller calls, and the other side, the buyer responds. This means that the entire market is dominated by the way sellers define their demand. In simple terms, if no one is selling what I am specifically looking to buy in the market, how do I market my demand?

This means sellers need to express their supply so that buyers will bid. Economics teaches us that in this situation, the seller is best served by providing as much information as possible on this impression, so that the maximum number of potential buyers is achieved. In other words, there is an economic incentive to say as much as possible about the impression.

The problem with this market architecture is that sellers can’t search for buyers before the inventory shows up. If a seller could search for demand and elect to meet some of that demand with their supply, the only information transferred during the transaction is that this audience member and ad placement unit meet the criteria of the buy order. So, if a seller never meets demand that violates PII standards, all transactions will be free of PII issues. In a market structure where demand can be transparent, the incentive is to share as little information as possible. This is the opposite market structure incentive from that of the real time market architecture.

The real time market architecture segregates supply and demand to the ‘call’ side or ‘response’ side, the market self-defines itself as an asymmetric market. For some inventory acquisition strategies like retargeting, this is a great, and most probably the most optimal, market structure. But, for big brands this market asymmetry is bad. We all know that they buy huge swaths of audiences across all media to build their brand. For these buyers, the real size of the transactions ($) they want to make is not accounted for in the second price auction. That auction does not know or care that you have a $25K budget for this line item.

This is a problem. By leaving this demand out of the price calculations we are effectively only looking at the tips of icebergs; and we still have tons of PII issues. If you are a marketer or a publisher navigating your boat through these treacherous waters, no wonder you’re fed up.

The #1 Way To Improve Yield Optimization, Thanks Dr. Laffer

Most yield systems 'think' about a graph of their revenue versus impressions andsee a line going up from left to right. The idea is that every single impression has the opportunity to generate income.

This view can only be true if each additional sellable impression is a real person that gives just as much attention to each new ad impression. With the exception of a microscopic minority, this is impossible. We all know that overloading the user with ads means they actually wind up ignoring all of the ads.

In reality, ethical publishers know full well that jamming your pages full of ads, and playing games and inflating page loads doesn’t last. Increasing supply undermines future pricing power and increases the opportunity for others to arbitrage the publisher.

In part, yield optimization is to blame. The real culprit is the second price auction. It’s baked into that auction method and it can’t be removed.

Let’s take another view. If we apply what we know, we know that the publisher’s revenue curve actually shows that if we keep pumping impression production up artificially, the total revenue we can generate through the market goes down.

Graph 2
Graph 2

So, the real question is what is the minimal amount of impressions that will maximize the publisher’s revenue. The curve illustrates that if there are zero ad impressions, there is no revenue—obviously—to the publisher. But if a publisher had no content and just ads, that won't generate revenue either, as there is no longer any incentive for a person to consume that publisher’s media.

If you’re a bit of a policy nerd like me, that sounds like a theory that came in to legend on a napkin “…officials Dick Cheney and Donald Rumsfeld in 1974 in which he reportedly sketched the curve on a napkin to illustrate his argument[1]” This is the legend of the Laffer Curve, an idea that made a big impact on tax policy throughout the 1980s.

To answer the revenue maximization question above we have to figure out the shape of the curve and how close we are to the top. Sounds simple enough right? Now, let’s think about any technologies in advertising that do that? Is there a yield optimization system that does that?

If you are managing media inventory today, can you tell me what dot on the curve above best represents your organization? If you can’t, your organization is managing the world through the lens of the line graph and not the curve.


The Untapped Potential Of All This Data

Originally Published in AdExchanger We’ve all marveled at the new technology solutions entering the programmatic marketing and advertising ecosystems, along with the vast quantities of data produced. Everywhere a business problem could be quantified, in terms of something that can be counted or measured, a solution has sprung up, specializing in things like behavioral modeling, viewability and attribution.

I believe the untapped potential of all this data is to predict the future. If we took the 100,000-foot view across both marketing technology and advertising technology, we find a very interesting pattern. The land of mar tech and ad tech data is divided in to three main areas, but only one is populated with almost zero technology.

Some technologies can be found in the area I call “The Past.” There are lots of technologies that are in “The Present.” There are hardly any in “The Future” – there are very few forecasting technologies.

I predict the third generation of advanced mar tech and ad tech solutions will focus on telling publishers and marketers what will probably happen in the future. Predicting the future is an exercise of understanding how past data about what happened, when it happened and why it happened can be used as the colors to paint a picture of the future.

The Past: What Happened?

Today’s technology for data capture and storage is like a 100-megapixel camera – it provides a super accurate picture of what happened in the past. We have 100% certainty that what we measured happened. It’s not like there are other possible outcomes in the past.

Data storage solutions provide this vast understanding of anything that we chose to measure. If a data point was created and saved, it can exist forever. The evolution of this area of technology is focused on the expansion of what data is measured and captured. The ever-growing sea of data is a beautiful sight to behold for the analytical among us.

The Present: What’s Happening?

The last massive wave of technology innovation in mar tech and ad tech happened in this category, which focuses on the collecting and disseminating data about the present. The technology that collects data about what the present looks like is less sharp than data about the past. To understand the present, we need to pull a lot of data together really fast so we can act on it.

For data about the past, the effectiveness of analytics that bring data together is not limited by the amount of time that it takes. For that reason, the present is a little less sharp. We don’t have time to look at all the data together.

What’s more, as the sea of data being collected grows, the amount we can actually act on becomes an ever-decreasing portion of what we actually have. It’s more like a 20-megapixel camera.

Technologies that work to understand what is happening and take action include yield management, creative optimization and supply-side platforms. This area of technology is evolving with a focus on the expansion of delivering and processing an ever-growing data set to answer a question in less than a second.

The Future: What’s Going To Happen?

In this category of mar tech and ad tech, the fewest solutions exist. There are no companies on the LUMAscape dedicated to forecasting; I only know of one startup. Forecasting features in current technologies are treated the way municipal politicians treat sewage infrastructure: Nobody wants to talk about it, it’s hard and dirty work, but no one can live without it.

The future will never be as clear as the present or the past, but in this space of mar tech and ad tech technologies, innovation and investment have significantly lagged the market. Predicting the future is hard. It’s never like the past – it’s fuzzy and out of focus. Our current tools for predicting the future are, at best, like a 0.5-megapixel camera. It’s really hard to tell what will happen.

This is where a ton of untapped potential exists. Leveraging all this data being collected everywhere to build better modeling tools will help bring the future into focus. No one can predict when this market shift will gain significant traction, but I think we will see the future as an increasingly important topic of conversation for industry innovation and thought leadership in the next few years.