Category MZ (Machine Zone)

An Alternative to the Order Book as the Market Design of a Crypto-Economic Trading Platform

In a crypto-economic trading platform:

  • “The network becomes the exchange”
  • Snapchat (ephemeral) bid-asks
  • User-defined smart contracts

The order book is a market design for the exchange of goods and assets.  It dates back to the European coffee houses of the late 1600s.  In London, Jonathan’s Coffee House was a significant meeting place for traders in London in the 1700s. It later became the site of the first London Stock Exchange.

In the late 1700s, in what later became known as New York City, Dutch traders met at a Buttonwood tree in lower Manhattan island to buy and sell goods coming into the port.   Now know as Wall Street, this location became the center of financial asset exchange in the United States.

Until the 1970s, stock exchanges were characterized by a market design involving traders gathering around pits with specialists manually matching bids and asks in paper order books (see below).

The great financial economist Fisher Black wrote a prophetic article in 1971 called “Toward A Fully Automatic Stock Exchange”   where he laid out the implications of the coming automation of the manual order book.  He speculated on what the computerization of the order book would mean for bidding mechanisms, liquidity and overall stock market efficiency.

Screenshot of Bid-Ask Order Book of Poloniex

Market design, indeed all design related to computers, is coupled tightly to the computer technology itself.  Just because one design is associated with a particular computer technology does not mean that the same design should be mindlessly carried over when the computer technology changes.

We recall the mindless carry over of the 80 character line limit established by IBM for punch cards in the 1920s to cathode ray tube (CRT) terminals in the 1970s.

There is a whole host of other instances of mindless carry over of designs when the technology changes.  One notable example is the organization of the factory floor after the conversion of machine power from a centralized shaft driven by water to decentralized electric power.

In the last several years, there has emerged a new decentralized, peer-to-peer (p2p) paradigm in computer architecture propelled by several trends — Internet of Things (IoT), autonomous vehicle-to-device (v2x) communication, and crypto.

This change demands a rethinking of the appropriateness of the centralized client-server order book market design in a crypto-economic platform.

The trend away from client server architecture is driven by a need to do more raw compute “at the edge” before sending data to the server for storage and higher order analytics.  This is known as “edge computing.”  The use cases for edge computing are Internet of Things (IoT) and autonomous vehicle-to-device (V2X) communication.

The trend away from client server architecture is also driven by the tremendous interest in Bitcoin, Blockchain and Ethereum.  Interest in crypto could be the start of a paradigm shift away client server financial intermediaries earning opaque rents and toward decentralized, trustless p2p protocols for validating and accounting for the exchange of financial assets.

A true true, decentralized crypto-economy involves not only a DLT layer but also high speed transaction layer. 

The thesis of this paper is that the time is now to consider a transaction layer with a true decentralized market design.

We believe that publish-subscribe will be the leading protocol of the transaction layer as it has already been deployed at scale an the platform behind several MMO games (from MZ) and chat platforms (WhatApp from Facebook, WeChat from TenCent).

 What is needed is an innovative p2p market design.  It could be along the lines a many-to-many, high frequency “take it or leave it” (TIOLI) publish-subscribe mechanism which could also be described as a discrete time, many-to-many, high frequency version of the Myerson auction.

Value Proposition:

  • user-defined contracts ( e.g. options with odd expiration dates, long-short pairs, straddles)
  • tokens earned by peers supplying liquidity spread contracts
  • elimination of latency rents going to HFT and server co-location fees going to exchange
  • elimination of “data ownership” rents earned by exchange

Specification suggestions:

  • high frequency, many-to-many, pub-sub protocol
  • messages in form of  Myerson “take it or leave it” (TIOLI) bid-asks
  • “serverless” with ephemeral matching with-in Redis-like in-memory data structure store, used as a database, cache and message broker.
  • ephemeral bid-ask data, only data “owned” is history of matches.
  • discrete time, batch process (i.e. events) following  Eric Budish’s work on continuous time design flaw in  HFT platforms 
  • third party AI bid bots
  • third party custodial services
  • settlement a function of DTL layer

Companies with pub-sub platforms

  • Satori (formerly MZ)
  • Facebook (WhatsApp)
  • TenCent (WeChat)
  • Google (Cloud Pub/Sub)

Some relevant URLs

Gabe Leydon, CEO Satori (MZ) TokenPost Interview During Korea Blockchain Open Forum,  July, 2018 https://www.youtube.com/watch?

Satori’s “AI Mesh network” transaction layer  stats — 500 Million “messages” per second or 1 million publishers sending 100 bytes a second 

Hadera Hashgraph’s DLT stats — 500,000 transactions per second with 100% consensus based on a “gossip of gossip protocol” and a consensus latency of 3.5 seconds.

Eric Budish, The Design of Financial Exchange, Some Open Questions at the Intersection of Econ and CS.  Simons Institute of Computing UCB, November 2015 https://www.youtube.com/watch?v=Rilv2AJ1TWM

Eric Budish, “Will the Market Fix the Market?”, AEA/AFA Joint Luncheon Talk, January 2017 https://www.aeaweb.org/webcasts/2017/luncheon

Albert “Pete” Kyle, “Continuous Auctions and Inside Trading”, Econometrica, November 1985   https://www.rhsmith.umd.edu/files/Documents/Centers/CFP/research/kyle1985.pdf

Albert “Pete” Kyle, “The Changing Nature of Trading Markets”, U of Maryland Conference,  May 2017 https://www.rhsmith.umd.edu/files/Documents/Centers/CFP/2017/kyle.pdf

Fisher Black, Toward A Fully Automatic Stock Exchange, 1971  http://17mj9yvb9fl2p5m872gtgax5.wpengine.netdna-cdn.com/wp-content/uploads/2017/07/Towards-a-fully-automated-stock-exhchange-part-1.pdf

 


An Outline of an Decentralized Alternative to the Order Book

In a crypto-economic trading platform:

  • “The network becomes the exchange”
  • Snapchat (ephemeral) bid-asks
  • User-defined smart contracts

The order book is a market design for the exchange of goods and assets. It dates back to the European coffee houses of the late 1600s.  In London, Jonathan’s Coffee House was a significant meeting place for traders in London in the 1700s. It later became the site of the first London Stock Exchange.

In the late 1700s, in what later became known as New York City, Dutch traders met at a Buttonwood tree in lower Manhattan island to buy and sell goods.   Now known as Wall Street, this location became the center of financial asset exchange in the United States.

Until the 1970s, stock exchanges were characterized by a market design involving traders gathering around pits with specialists manually matching bids and asks in paper order books (see below).

 

The great financial economist Fisher Black wrote a prophetic article in 1971 called “Toward A Fully Automatic Stock Exchange”   where he laid out the implications of the coming automation of the manual order book.  He speculated on what the computerization of the order book would mean for bidding mechanisms, liquidity and overall stock market efficiency.

Screenshot of Bid-Ask Order Book of Poloniex

Market design, indeed all design related to computers, is coupled tightly to the computer technology itself.  Just because one design is associated with a particular computer technology does not mean that the same design should be mindlessly carried over when the computer technology changes.

We recall the mindless carry over of the 80 character line limit established by IBM for punch cards in the 1920s to cathode ray tube (CRT) terminals in the 1970s.

There is a whole host of other instances of mindless carry over of design when the technology changes.  One notable example is the organization of the factory floor after the conversion of machine power from a centralized shaft driven by water to decentralized electric power.

In the last several years, there has emerged a new decentralized, serverless, peer-to-peer (p2p) paradigm in computer architecture propelled by several trends: Internet of Things (IoT), autonomous vehicle-to device communication (V2X), and crypto.

This technological change demands a rethinking of the appropriateness of the centralized client-server order book market design as the core of a transaction layer in a crypto-economic platform.

The trend away from client server architecture is driven by a need to do more raw compute “at the edge” before sending data to the server for storage and higher order analytics.  This is known as “edge computing.”  The use cases for edge computing are Internet of Things (IoT) and autonomous vehicle-to-device (V2X) communication.

The trend away from client server architecture is also driven by the tremendous interest in Bitcoin, Blockchain and Ethereum.  Interest in crypto could be the start of a paradigm shift away from client server financial intermediaries earning opaque rents and toward decentralized, trustless p2p protocols for validating and accounting for the exchange of financial assets.

A true decentralized crypto-economy involves not only a DLT layer but also a high speed transaction layer. 

The thesis of this paper is that the time is now to consider the possibility of pairing a transaction layer with a true decentralized market design with a decentralized distributed ledger technology (DLT).

We believe that publish-subscribe  currently is the leading protocol for the transaction layer as it has already been deployed at scale an the platform behind several MMO games (from MZ) and chat platforms (WhatApp from Facebook, WeChat from TenCent).

We believe that MZ’s recently spun-off subsidiary Satori is leading the integration of a pub-sub transaction layer with a DLT called Hedera Hashgraph. The question is what will be the market design for the transaction layer?

Some URLs relevant to Satori’s plans:

Gabe Leydon, CEO Satori, TokenPost Interview During Korean Blockchain Open Forum, July 2018

Gabe Leydon, CEO Satori (MZ) Fireside Chat Crypto Invest Summit, May 2018

CEO Gabe Laydon leaves MZ to focus on crypto — Venturebeat June 1, 2018

Gabe Leydon video at Hedera Hashgraph NY announcement April 18, 2018

Satori’s “AI Mesh network” transaction layer  stats — 500 Million “messages” per second or 1 million publishers sending 100 bytes a second 

Hedera Hashgraph’s DLT stats — 500,000 transactions per second with 100% consensus based on a “gossip of gossip protocol” and a consensus latency of a 3.5 seconds.

 

 

 

 

 

 

 

 

In a crypto-economic trading platform:

  • “The network becomes the exchange”
  • Snapchat (ephemeral) bid-asks
  • User-defined smart contracts

Value Proposition:

  • user-defined contracts ( e.g. options with odd expiration dates, long-short pairs, straddles)
  • tokens earned by peers supplying liquidity spread contracts
  • elimination of latency rents going to HFT and server co-location fees going to exchange
  • elimination of “data ownership” rents earned by exchange

Specification suggestions:

  • high frequency, many-to-many, pub-sub protocol
  • messages in form of  Myerson “take it or leave it” (TIOLI) bid-asks
  • “serverless” with ephemeral matching with-in Redis-like in-memory data structure store, used as a database, cache and message broker.
  • ephemeral bid-ask data, only data “owned” is history of matches.
  • discrete time, batch process (i.e. events) following  Eric Budish’s work on continuous time design flaw in  HFT platforms 
  • third party AI bid bots
  • third party custodial services
  • settlement a function of DTL layer

 

Companies with pub-sub platforms

  • Satori (formerly MZ)
  • Facebook (WhatsApp)
  • TenCent (WeChat)
  • Google (Cloud Pub/Sub)

Register with the SEC as an ATS or ECN not an exchange.

Targets — continuous time order-processing client-server exchanges with massive multi-million dollar rents going to server owners and HFT snipers.

  • Pseudo-crypto DEX with client server order books
  • FOREX with tokenized fiat money
  • Swaps
  • Options
  • Dark Pools
  • Replace “book-maker” gambling with p2p gambling

 

Abrams tweets on the need for decentralized market design as part of a true decentralized crypto-economics transaction layer

 

 


MZ (Machine Zone) and Its Satori Platform

Satori® – Towards a P2P Crypto-Economic Platform

 

Towards A Crypto-Economic Market Design: Discrete-Time, High Frequency, P2P

 

Machine Zone (MZ): A $10 Billion Dollar Unicorn in the Making

 

Edge Computing Use Cases for MZ’s (Machine Zone) Platform

Machine Zone and The Perversity of Unicorn Lists

 

Machine Zone (MZ): A $4 Billion Unicorn That Walks the Walk

 

Machine Zone: IPO or What?


The Facebook – Google Duopoly — Where They Differ?

Summary:

Facebook and Google differ fundamentally in the degree of adtech vertical integration.  Facebook only has a significant internal supply side platform (SSP).  Google has a significant SSP, ad exchange, and a demand side platform (DSP).

Facebook recognizes that its supply of ad impressions has reached a ceiling due to user annoyance of ads in their feeds.  With supply now inelastic within Facebook, there are three options open to increasing revenue — quantity times unit price.

  1. Get into selling impressions outside its “walled garden” which it has done through a retargeting business.
  2. Assert its pricing power as a duopolist and just “shift up the supply curve” (i.e. limit ad impressions).
  3. Work to “shift up the demand curve” of advertisers by sharing user data with independent DSPs to improve ad buy ROI resulting in a willingness to pay higher prices.

Option 2 exposes Facebook to antitrust lawsuits.  Option 1 and 3 are promising, but it relies on a sharing of user data with independent demand side platforms (DSPs) which has become problematic due to the Cambridge Analytica debacle.

We make the case for Option 3.

 

The Facebook – Google Duopoly

The intermediate market for digital ad impressions is now dominated by the Google – Facebook duopoly.  In 2017, the duopoly’s total share of ad impression sales was estimated at 60.4% (see below).

In the fast growing digital native ad subset market, the duopoly’s share rose to 91.9%.

The data above largely reflects ads embedded in mobile and PC content.  The market for real time ad impressions embedded in ad-supported streaming TV — as opposed to subscription-supported Netflix and Amazon Prime — is just beginning.

 

The Intermediate Market for Digital Ad Impressions

On the extreme demand-side of the digital ad impression market are advertisers wishing to buy ad impressions and on the extreme sell-side are content publishers selling ad impressions.

Between advertisers and publishers is an adtech intermediate market. It consists of supply-side platforms (SSPs) also dominated by Google and Facebook with Amazon being a fast riser in the SSP space.

Next comes ad exchanges where transactions and pricing takes place increasingly via real-time auctions.  Google dominates with DoubleClick AdExchange. Facebook launched its own FBX in 2011 but let it slowly die over the next three years.

Next comes demand side platforms (DSPs), which is the most competitive segment.  Google has its own internal DSP called DoubleClick Bid Manager which dominates ad buying both on Google’s “walled garden” of search and on its own subsidiary YouTube.

Facebook has an internal DSP called Facebook Ad Manager useful to small and medium business wishing to buy ads within its eponymous “walled garden.” But, Facebook seems to be holding back from customizing its internal DSP to cater to large advertisers with unique needs.

Because of the publicity surrounding Cambridge Analytica, Facebook is under enormous pressure to pull back all the ways its allows third parties to access its user data.  We think it would be “throwing out the baby with the bathwater” if Facebook were to pullback of all adtech — SSP, AdEx, and DSP — within its “walled garden”.

Rather, it needs all the help it can in filtering out “lemons” ala George Akerlof’s iconic economics paper “A Market for Lemons”.  It is in Facebook’s own interest to foster independent DSP’s with “clean room” access to data (see below).

 

The Focus of Antitrust Concerns

The sheer scale of the Google – Facebook duopoly is the current focus of antitrust concerns. But, the extent of the duopoly’s vertical integration —  owning the full vertical stack of businesses from content publishing to a SSP to an ad exchange to a DSP — needs to be analyzed in much greater detail by adtech experts for potential antitrust violations.

Antitrust concerns so far has focused on the supply side with the most recent flare-up being Facebook’s decision to limit third-party news feeds inserted into the social graphs of Facebook users.

This supply-side focus on content is understandable as such developments hurt the job prospects of the very paid  tech and business writers [ not us 😉 ] who write about Google and Facebook.

In contrast, we think a recent development on the demand-side  deserves more attention. This is because it signals that Facebook is seeking to deflect antitrust concerns by actively assisting in the development of strong, independent DSPs.

 

Facebook’s Effort to Foster Strong Independent DSPs

In December 2017, TechCrunch published an article describing how an internal Facebook team of 100 engineers has been working with big advertisers to develop their own customized DSPs.

Right off the bat, this revelation is a clear sign that Facebook today does not intend on repeating the anticompetitive tactics used by Microsoft in the mid 1990s.  Back then, Microsoft used its control over the dominant Windows PC operating system to throttle the ability of users to replace Microsoft’s own default browser with a popular third-party browser developed by Netscape.

Notwithstanding the antitrust motivation for supporting independent DSPs, Facebook’s initiative is good from a pure business perspective.  Facebook realizes that improving ROI on purchased ad impressions via reduced information asymmetry translates into a willingness by advertisers to make higher bids for “peaches” instead of “lemons” ala George Akerlof’s iconic economics paper “The Market for Lemons”.

Indeed, we see the current digital ad impression market as the “mother of all markets for lemons”  including the following list of extreme conditions for markets with information asymmetry between sellers and buyers:

  • zero marginal cost on both sides creating lemons and traps
  • dueling artificial intelligence (AI) on both sides morphing sell-side lemons and buy-side traps
  • real-time auctions
  • low latency rendering and fill of ad after purchase of impression
  • double interrelated information asymmetry (see below)

 

Facebook’s Options For Increasing Revenue

Facebook recognizes that its supply of ad impressions has reached a ceiling due to user annoyance of ads in their feeds.  With supply now inelastic within Facebook, there are three options open to increasing revenue — quantity times unit price.

  • Get into selling impressions outside its “walled garden” which it has done through a retargeting business.
  • Assert its pricing power as a duopolist and just “shift up the supply curve” (i.e. limit ad impressions) which it in effect has done by limiting outside news feeds.

Facebook Option 2: Shift up the Supply Curve

  • Work to “shift up the demand curve” of advertisers by working with independent DSPs to improve ad buy ROI resulting in a willingness to pay higher prices.

Facebook Option 3: Shift up the Demand Curve

Facebook is signaling that it is giving option 3 a try.

Here is a great quote from an April 2017 adExchanger article on digital ad prices as a reflection of quality and the opportunity for Facebook to receive higher prices — shift up the demand curve  — from advertisers and agencies using more discriminating DSPs.:

We buy it cheaper” used to be the lead differentiator in a pitch. Today, agencies that lead with “We can buy digital cheaper” have a sign taped to their back that says, “We buy lots of fraud.” Low prices in digital media are not only no longer a badge of honor, they’re a warning sign.

Here is a quote from the TechCrunch article on the ROI improvement coming from DSPs built with the assistance of the Facebook engineering team:

Facebook says that on average, clients working with the solutions engineering team see their return on ad spend improve by 100 percent.

The article mentioned that after working with the Facebook team to improve the performance of its own internal DSP,  the mobile game company MZ (formerly Machine Zone), has spun off its internal DSP as an independent business called Cognant ®.

It should be noted that even before the spin-off, MZ was already the largest “direct response” advertiser in the world and likely on Facebook itself.

As Facebook’s largest direct response advertiser, MZ was the likely first recipient of access to Facebook user data located in “clean rooms” on Facebook servers.  Here is a February 28, 2017 description by AdExchanger of the linkage:

Google and Facebook are each responding to advertiser demands for more data. Facebook does data-sharing deals on the DL with large marketers that push for it.

In so-called “clean rooms,” for example, advertisers can compare their first-party data with impression-level Facebook campaign delivery data using laptops that have never touched the internet. Facebook also allows certain large advertisers to create a private instance on its server to run advanced analytics.

We would expect Google to lag behind Facebook as Google’s supply of ad impressions is more elastic.  Google can increase revenue via increasing the supply of impressions especially on its YouTube subsidiary.

Facebook has no room in its “wall-garden” for more ad impressions. It will be interesting to see how much Facebook derives revenue from its retargeting business outside its “walled garden.”  Otherwise, the only way Facebook can increase revenue is by working to improve ROI on the demand side and “shift up the demand curve.”

 Facebook antitrust lawsuits will inevitably dwell on ad price trends as measured by cost-per-click (CPC).  Consider the following graph showing that Facebook’s CPC  rose 136% in the first six months of 2017.

How much of the above trend was due to Facebook asserting it’s pricing power and how much of that trend was due to other factors?

For example, the upward trend could be due in part to a secular improvement in ad ROI delivered by independent DSPs with help of Facebook supplied application programming interfaces (APIs), thus reducing information asymmetry on the part of buyers.

Of course, it takes more than API hooks for a DSP to deliver significant improvements in ad ROI.  It takes a DSP that can build a sophisticated real-time programmatic bid engine and a real-time predictive analytics platform that feeds off Facebook-supplied user data and spits out bids with improved click-through rates.

In sum, Facebook has deflected the antitrust case against it by assisting independent demand side platforms (DSPs) like MZ’s (Machine Zone’s) Cognant ® to build strong countervailing platforms.  

 At the same time, working with independent DSPs to improve ad buy ROI and a willingness to pay higher price is a way out of its conundrum of growing revenue while limiting ad impressions in its “walled garden”.

 


“Lemons” and Antitrust: Why Facebook and MZ Are AdTech Frenemies — DSP Cognant

(Postscript: July 11, 2018)

Adexchanger reported that MZ has shut down its independent DSP business Cognant and laid off all 125 employees as well as half of its in-house media buyers.

(Published: March 1, 2018)

Summary: 

Facebook has deflected the antitrust case against it by assisting independent demand side platforms (DSPs) like MZ’s (Machine Zone’s) Cognant ® to build strong countervailing platforms.

Facebook is signaling that it does not intend on being “the Microsoft of the 1990s” by throttling competition on the demand side of the ad impression market.

Furthermore,  this cooperation makes sense from a pure business perspective as Facebook is “shifting up the demand curve” of advertisers by working with independent DSPs to improve ROI on ad buys resulting in a willingness to pay higher prices.

Facebook wants to reduce information asymmetry in the ad impression market. It is in its own interest to help make it a market of “peaches” not “lemons” ala George Akerlof’s iconic economics paper “A Market for Lemons”

Indeed, we see the current digital ad impression market as the “mother of all markets for lemons”  accompanied by  a whole new lexicon — clickbait, click farms, fake news,  brand safety, and walled gardens.

 

The Intermediate Market for Digital Ad Impressions

The intermediate market for digital ad impressions is now dominated by the Google – Facebook duopoly.  In 2017, the duopoly’s total share of ad impression sales was estimated at 60.4% (see below) .

In the fast growing digital native ad subset market, the duopoly’s share rose to 91.9%.

 

 

On the extreme demand-side of the digital ad impression market are advertisers wishing to buy ad impressions and on the extreme sell-side are content publishers selling ad impressions.  In between  is a complex web of software intermediaries that would dumbfound (and still does) Madison Avenue.

The data above largely reflects ads embedded in mobile and PC content.  The market for real time ad impressions embedded in ad-supported streaming TV — as opposed to subscription-supported Netflix and Amazon Prime — is just beginning.

The value of the impressions to advertisers is a function of viewer eyeballs, associated data, and engagement as measured by time spent on the site.  Sheer eyeballs without data is of little value to advertisers as exemplified by Twitter.  Facebook and Google Search/YouTube dominate because of all three — eyeballs, data, and engagement.

If you ask people in the tech world what is the business of Facebook and Google, they would not say social networking or internet search or video delivery.  They would say the two dominate the performance marketing business — selling ads with quantifiable (as opposed to estimable ala Nielsen) results as in costs per click or impression.

Between advertisers and publishers is an adtech intermediate market. It consists of supply-side platforms (SSPs) also dominated by Google and Facebook with Amazon being a fast riser in the SSP space.

Next comes ad exchanges where transactions and pricing takes place increasingly via real-time auctions.  Google dominates with DoubleClick AdExchange.  Facebook launched its own FBX in 2011 but let it slowly die over the next three years.

Next comes demand side platforms (DSPs), which is the most competitive segment.  Google has its own internal DSP called DoubleClick Bid Manager which dominates ad buying both on Google’s “walled garden” of search and on its own subsidiary YouTube.

Facebook has an internal DSP called Facebook Ad Manager useful to small and medium business wishing to buy ads within its eponymous “walled garden.” But, Facebook seems to be holding back from customizing its internal DSP to cater to large advertisers with unique needs.

Because of the publicity surrounding Cambridge Analytica, Facebook is under enormous pressure to pull back all the ways its allows third parties to access its user data.  We think it would be “throwing out the baby with the bathwater” if Facebook were to  pullback of all adtech — SSP, AdEx, and DSP — within its “walled garden”.

Rather, it needs all the help it can in filtering out “lemons”.  It is in Facebook’s own interest to foster independent DSP’s with “clean room” access (see below) to data.

The Focus of Antitrust Concerns

The sheer scale of the Google – Facebook duopoly is the current focus of antitrust concerns. But, the extent of the duopoly’s vertical integration —  owning the full vertical stack of businesses from content publishing to a SSP to an ad exchange to a DSP — needs to be analyzed in much greater detail by adtech experts for potential antitrust violations.

Antitrust concerns so far has focused on the supply side with the most recent flare-up being Facebook’s decision to limit third-party news feeds inserted into the social graphs of Facebook users.

This supply-side focus on content is understandable as such developments hurt the job prospects of the very paid  tech and business writers [ not us 😉 ] who write about Google and Facebook.

In contrast, this paper focuses on the demand-side and DSPs.  We think a recent development on the demand-side  deserves more attention. This is because it signals that Facebook is seeking to deflect antitrust concerns by actively assisting in the development of strong, independent DSPs.

Facebook’s Effort to Foster Strong Independent DSPs

In December 2017, Tech Crunch published an article describing how an internal Facebook team of 100 engineers has been working with big advertisers to develop their own customized DSPs.

Right off the bat, this revelation is a clear sign that Facebook today does not intend on repeating the anticompetitive tactics used by Microsoft in the mid 1990s.  Back then, Microsoft used its control over the dominant Windows PC operating system to throttle the ability of users to replace Microsoft’s own default browser with a popular third-party browser developed by Netscape.

Notwithstanding the antitrust motivation for supporting independent DSPs, Facebook’s initiative is good from a pure business perspective.  Facebook realizes that improving ROI on purchased ad impressions via reduced information asymmetry translates into a willingness by advertisers to make higher bids for “peaches” instead of “lemons” ala George Akerlof’s iconic economics paper “The Market for Lemons”.

Indeed, we see the current digital ad impression market as the “mother of all markets for lemons”  including the following list of extreme conditions for markets with information asymmetry between sellers and buyers:

  • zero marginal cost on both sides creating lemons and traps
  • dueling artificial intelligence (AI) on both sides morphing sell-side lemons and buy-side traps
  • real-time auctions
  • low latency rendering and fill of ad after purchase of impression
  • double interrelated information asymmetry (see below)

Facebook recognizes that its supply of ad impressions has reached a ceiling due to user annoyance of ads in their feeds.  With supply now inelastic within Facebook, there are three options open to increasing revenue — quantity times unit price.

  • Get into selling impressions outside its “walled garden” which it has done through a retargeting business.
  • Assert its pricing power as a duopolist and just “shift up the supply curve” (i.e. limit ad impressions) which it in effect has done by limiting outside news feeds.

Facebook Option 2: Shift up the Supply Curve

  • Work to “shift up the demand curve” of advertisers by working with independent DSPs to improve ad buy ROI resulting in a willingness to pay higher prices.

Facebook Option 3: Shift up the Demand Curve

Facebook is signaling that it is giving option 3 a try.

Here is a great quote from an April 2017 adExchanger article on digital ad prices as a reflection of quality and the opportunity for Facebook to receive higher prices — shift up the demand curve  — from advertisers and agencies using more discriminating DSPs.

“We buy it cheaper” used to be the lead differentiator in a pitch. Today, agencies that lead with “We can buy digital cheaper” have a sign taped to their back that says, “We buy lots of fraud.” Low prices in digital media are not only no longer a badge of honor, they’re a warning sign.

Here is a quote from the Tech Crunch article on the ROI improvement coming from DSPs built with the assistance of the Facebook engineering team:

Facebook says that on average, clients working with the solutions engineering team see their return on ad spend improve by 100 percent.

The article mentioned that after working with the Facebook team to improve the performance of its own internal DSP,  the mobile game company MZ (formerly Machine Zone),  has spun off its internal DSP as an independent business called Cognant ®.

It should be noted that even before the spin-off, MZ was already the largest “direct response” advertiser in the world and likely on Facebook itself.

As Facebook’s largest direct response advertiser, MZ was the likely first recipient of access to Facebook user data located in “clean rooms” on Facebook servers.  Here is a February 28, 2017 description by AdExchanger of the linkage:

Google and Facebook are each responding to advertiser demands for more data. Facebook does data-sharing deals on the DL with large marketers that push for it.

In so-called “clean rooms,” for example, advertisers can compare their first-party data with impression-level Facebook campaign delivery data using laptops that have never touched the internet. Facebook also allows certain large advertisers to create a private instance on its server to run advanced analytics.

MZ’ DSP is also likely an early adopter of the Unicorn startup Sprinklr for CRM and Martech:

Sprinklr is the most complete social media management system for the enterprise. We help the world’s largest brands do marketing, advertising, care, sales, research, and commerce on Facebook, Twitter, LinkedIn, and 21 other channels globally – all on one integrated platform.

A tight integration of Sprinklr + MZ’s Cognant, especially around real-time brand management (e.g., seeing the impact of localized Facebook ads for McDonald’s garlic fries on purchases at local outlets in real-time)  would certainly be a threat to an earlier generation of cloud-based CRM,  like salesforce.com and Oracle, that draws on dated information.

MZ’s DSP very likely has benefited from face-to-face meetings with the Facebook team located in Menlo Park not more than a 20 minute drive from the MZ’s headquarters in Palo Alto.  Indeed, MZ’s current HQ in Palo Alto on Page Mill Road across from Stanford was the former HQ of Facebook.

It should be noted that MZ is also located close to Google’s HQ  in Mountain View. It will be interesting to see if Google might offer similar assistance.

We would expect Google to lag behind Facebook as Google’s supply of ad impressions is more elastic.  Google can increase revenue via increasing the supply of impressions especially on its YouTube subsidiary.

Facebook has no room in its “wall-garden” for more ad impressions. It will be interesting to see how much Facebook derives revenue from its retargeting business outside its “walled garden.”  Otherwise, the only way Facebook can increase revenue is by working to improve ROI on the demand side and “shift up the demand curve.”

The question is can MZ’s DSP Cognant and maybe a few other DSPs scale sufficiently and demonstrate enough independence to be called true countervailing powers to the Facebook – Google duopoly?

Or will Cognant become a “front” of  independence “playing nice” with Facebook?  Will Cognant become some fake sign of competition to be trotted out by Facebook lawyers in some antitrust lawsuit down the road?

Such an antitrust lawsuit will inevitably dwell on ad price trends as measured by cost-per-click (CPC).  Consider the following graph showing that Facebook’s CPC  rose 136% in the first six months of 2017.

How much of the above trend was due to Facebook asserting it’s pricing power and how much of that trend was due to other factors?

For example, the upward trend could be due in part to a secular improvement in ad ROI delivered by independent DSPs with help of Facebook supplied application programming interfaces (APIs), thus reducing information asymmetry on the part of buyers.

Of course, it takes more than API hooks for a DSP to deliver significant improvements in ad ROI.  It takes a DSP that can build a sophisticated real-time programmatic bid engine and a real-time predictive analytics platform that feeds off Facebook-supplied user data and spits out bids with improved click-through rates.

Right now, we believe that the only independent DSP that has this capability is MZ ‘s Cognant.

In sum, Facebook has deflected the antitrust case against it by assisting independent demand side platforms (DSPs) like MZ’s (Machine Zone’s) Cognant ® to build strong countervailing platforms.