By Davis Rosborough
Davis is a technology-focused M&A advisor and investor based out of New York City. He is currently a Vice President at Progress Partners and is the co-founder of IndustryMaps.
February 19, 2018
Machine Learning & AI
Machine Intelligence, in the form of ML, Neural Networks, and self-adapting AI, will grow in importance and significance, impacting all industries in some ways that have yet to be determined. Is it ?an interest? or ?a necessity? for companies to replace loyal carbon elements with silicon and source code? It depends: McKinsey?s Nov-17 report indicated that over 800M jobs globally will be replaced by AI efficiencies by 2030 ? most of those in China, but 24-45% of the US labor market will need to upgrade skills or change occupations.
Impetus stirs not purely from opportunity to create operating leverage; yes, it is a matter of continuing to find ways to create shareholder value, but in many industries, it is about the capacity to ingest and digest huge volumes of data ? a necessity if a company hopes to create a sustained competitive advantage. To compete, some companies will build their own ? if they can recruit a team from the pool of 22K PhD-level AI engineers globally (note: Amazon and Google recruit at a 10x pace to IBM). But likely most companies will adapt rented AI solutions from providers like IBM Watson Cognitive Applications business, or Google?s open source offerings, and then build frameworks that accelerate their own business? ability to deploy evolving solutions that get smarter by increased frequency of inputs. The difficulty new market entrants face will not be developing new applications of AI to solve problems, but that incumbents will benefit from compounding value of larger, more exclusive data lakes.
Sundar Pichai, Google?s Chief, described the AI Renaissance as marginally ?more profound and revolutionary than the advent of fire or electricity.? Take him at his word.
In 2017, Facebook?s AI Research team quietly created two AI bots and taught them to communicate. Within hours from launch, the researchers noticed something odd. The bots had quickly developed a new language that its authors were unable to understand. Within 36 hours of the divergence away from rich expression of English, the bots were abruptly shut down. Perhaps in 2018, we will determine that the Corpus of our level of understanding is just the starting point on an exponential path toward more efficient AI problem solving. The bots will learn from us and move on.
The good news: as AI automates workflows, it will free up the humans to concentrate and speculate on strategy and engage in deeper white-space thinking. In the case of one our clients, their AI engine is outperforming automated human-driven buying by an average of 781% ? a certainly significant figure. The question in 2018 will be how to gauge varying AI tech head-to-head.
The next twelve months will be a very interesting time: Facebook for example is very close to building chat bots with true negotiation skills, while Google?s Translation services, built on a neural network, may quickly allow us to communicate in real-time in any language or dialect. This doesn?t mean we should cease learning new ones ? if we want our brains to be competitive, more rigor.
At what point does the subjectivity of the Author?s Source Code not meaningfully affect outcome? When the quality, integrity, and persistence of your data inputs reach marginal sustained significance over your competitors.
Amazon will charge the field on Facebook and Google
Consolidation within ad-tech and mar-tech thematically drove M&A events in 2017 ? there were fewer seed financings, but more tech IPOs (while institutional VC is increasingly moving to larger rounds, earlier, for proven founders).
Consolidation of market power among the Have?s ? particularly Facebook and Google, who each have volunteered first-party data moats, and Amazon, with its transaction-rich audience segments ? give these three firms tremendous market advantage and capacity to absorb most of every new dollar that enters the market.
Bret Brase, a Managing Director at Progress Partners, ran an analysis of the market stimulus that would arise in digital mid-market if Facebook and Google shrink by just 4%. Thousands of companies would become profitable, fueled the diversification of demand away from the Duopoly, enabling hiring and investment, subsequent mergers, founder and employee wealth creation, and the seeding of secondary and tertiary start up marketplaces outside of New York and Silicon Valley. Word is that the paper never reached Janet Yellen?s desk.
Respectively, Facebook and Google rely on 92% and 88% of revenues from Advertising, selling Audience-as-a-Product, and in Google?s case, owning end-to-end monetization, controlling the transaction, in addition to the content, and the audience it bears. While the market tries to get the pair to admit they are in fact media companies ? in reality, both are data companies, as data is the center of gravity business advantage, or fulcrum of their market control.
Amazon currently threatens that duopoly, owning two-thirds of Share-of-Voice, with more Amazon Prime households than voting households in the 2016 election, setting up a near-term likelihood of Amazon cutting a huge swath of Google?s search dominance away by way of Voice. Early adopters are enthusiastic or at least experimental, but for the youngest generation it is second nature ? a huge determinant of Google?s future success relies on Google Assistant and Home.
Scott Galloway a NYU Stern professor, and author of The Four: The Hidden DNA of Amazon, Apple, Facebook, and Google, has been the chief voice in the call to break up the the Big Four. It is most likely to happen to Amazon, which has the incredible capacity to borrow money more cheaply than China. The behemoth has six core markets it?s attacking beyond e-commerce, and astonishingly invests 100 cents on the dollar back into the consumer experience (like no competitor can), and continues to see its share value increase without the need to grow profits. Regulators in the next administration will likely take notice ? by then, Amazon will likely have become the first company with $1 trillion market cap.
Who can compete? Perhaps, AT&T?s acquisition of Time Warner, a mighty marriage of distribution from a 5G network with premium video content to push through the pipe. AT&T-TWX (or Company-To-Be-Named-Later) is primed to be the next Walled Garden, especially if they then acquire a leading Demand Side Platform (or direct Real Time Bidding infrastructure to tap into premium publishers). Doing so would start the inevitable consolidation of demand activation around four or five leading platforms ? which will give a chance for Value Chain Transparency for the first time in the digital media ecosystem.
Is there a Distributed Ledger behind that Buzzword?
There is huge potential for Distributed Ledger Technology (DLT) to streamline and rectify calcified financial services and data management firms ? inherently vulnerable and defensive on open TCP/IP to denial of service attacks. Blockchain will be the single biggest technological shift for the media industry ? characterized by a jointly incentivized distributed infrastructure, encrypted and secure identity management, and decentralized or ?trustless? consensus mechanisms ? media companies will be able to store and transfer value, under self-effectuating ?Smart Contracts? where rules encoded defining terms and the actions that stem from it ? with flexible compliance (adapting for changes in price and ownership). These fundamental components are why Blockchain is expected to be a $355 billion US market in the next eighteen months.
Few businesses wouldn?t benefit from exploring techniques for adoption or experimentation, and with low cost structures, there are implications for inaction. For example, a report cited that financial institutions could save over $20 billion in annual infrastructure and operational costs. Decentralized, forward and backward compliant ledgers could carve away [read:disrupt] the audit industry by making a painful process near instantaneous. These few examples help us understand why the World Economic Forum expects 10% of global GDP will be stored within blockchain by 2027 ? and perhaps that could be even sooner.
Long-tail cash for crypto assets ? over $19 billion in investor dollars has been poured into these ICOs which are fundamentally securities offerings, but amazingly do not include equity from the issuer. The SEC is behind, but in 2018 will move quickly to regulate ICOs under federal securities laws as these assets are being treated as securities by their promoters. The question remains: if you?re not one of the accredited few, will it be possible to bar participation? ?Bans? smack ridiculous to perceived ?real? threats. It is also important that Cryptocurrencies are fundamentally divested in our own mindshare from the larger potential of Blockchain ? whether you are in market or not (I am).
Market Making as Life Death and Rebirth
In one sense, retargeting has become an example of bad ad behavior, and on the other hand, the specificity and addressability of media offerings are dependent on the closer and more persistent understanding of identity than ever before (it?s necessary and retargeting has begun retargeting itself).
There is a huge greenfield race toward owning and controlling ID management and authentication of User ID: as you shift from environment to environment, device to device, moving across the globe, consumption habits and behaviors will be traced, and that data will then be mined for increasingly customized experiences wherever you go. Someday soon you may have a Personal Data Broker.
A big question to be raised is which entity (or entities) will own ID Authentication & Identity Management in Blockchain Protocol for Media? Who manifests the central Key Maker?
Location-based data sinks will gradually become commoditized ? but only if your method or process lacks differentiation in compilation or maintenance [mindfully delineate (1) static, probabilistic, historical look-back from (2) dynamic, deterministic, predictive capabilities], while unique nodes of collection will be harvested and sold. Those with strong integrity in on-boarding and compiling (e.g., reducing discrepancy and error rates) will win.
We will also see increasing verticalization of mar-tech, with access to unique audience segments, delivering to vertical clientele performance based outcomes will allow the ecosystem to spawn new winners over generalist players. Adobe, Oracle, Salesforce, IBM, and Cision will continue to sop up market share as the dominant marketing clouds ? all with different approaches to the same ultimate prize.
So what will the future hold? The world is divided between those trying to protect the Past from the Future, and those actively working to protect the Future from being inhibited or delayed by those staunchly protecting the Past. It is already written a thousand times over that the former act in vain. I am among the latter.
For more musings from Davis, follow him on LinkedIn or on Twitter at @davisrosborough.