I’m more passionate about this topic than anything I’ve ever written about. As an eternal optimist I don’t see a world dominated by Google, Amazon, Chase and Walmart…. But rather a new economy where millions of small businesses thrive. Where every person can employ their creative talent in collaboration focused on value creation. A world where capital flows to great ideas across geographic boundaries lifting the stifling poverty of most emerging markets.
I have been working on this blog for 4 months and still not happy with it. My readers are probably used to my half-baked thoughts, so be prepared for cookie dough. This blog builds 6 previous efforts, so what is new here?
- How transaction costs and distributed intelligence impacts bespoke commercial constructs and networks (ie disrupting integrated companies).
- Investment premise – modularity and service standardization as key indicators of segment development
- Modularity as a mechanism to make “small” work and how it relates to networks and transaction costs (TCE).
- Data advantages of Google/FB/Amazon and new developing approaches of data federation
- Core services for orchestration networks (and Commerce Signals)
- Key challenge for collaboration: value measurement and dissemination
Environmental Forces Driving Structure
The forces that have driven scale, and the design of today’s economy, are atrophying (Transaction Cost Economics, asset intensity, information intensity, finance… etc). Paul Graham’s calls this change Refragmentation, I tend to view this as the transformation of networks or commercial entropy.
It’s as if the gravitational constant (the big G) is changing and new forces are driving formation of new networks, influenced by a rapidly evolving world of “weak links”. Information intensity has moved beyond “tweaking” 100 yr old business models to transform the design of industries, communities and people.
“Big” integrated supply chains have proven their ability to win through price, scale and the ability to attract capital. However small wins in nature where millions of specialists operate to create a beautiful quilt of thousands of unique sustainable environments. We see increasing entropy not only in the Third Law of Thermodynamics and the cosmos but also in biology and sociology (ex wiki on marxism). Commercially, each “value creator” struggles to be both free, and associated with other elements where it can survive.
As I outlined in Networks and Openness,
Scale-free distribution (completely open networks) is not always the optimal solution to the requirement of cost efficiency. .. in small world networks, building and maintaining links between network elements requires energy…. [in a world with limited resources] a transition will occur toward a star network [pg 75] where one of a very few mega hubs will dominate the whole system. The star network resembles dictatorships in social networks
The gravitational “G” is a metaphor for the environmental forces that shape networks. To change them both “G” + incremental energy is required in order to “break” existing links and create an environment where resources are open to new networks (see transaction costs below). I see four paths to organize resources:
- Orchestration Networks (ex Uber, Google Buy, AirBnb)
- Standardization/Modularity (Intel’s PCI, Europe’s SEPA)
- Market (eBay, NYSE, Mechanical Turk, …etc)
- Self Organization (Part 2 of this Blog – see UofM’s Micro Mote)
I postulate that the Discovery, Standards/Rules, Value Dissemination, Pricing and Access to Capital are the central differentiators to these different schemes. Too arcane?
Example. Today, Fortune 100 companies benefit from both low Transaction Cost Economics (TCE) and ability to raise capital based upon control of value creation. How can an investor trust a collection of 5 interdependent companies to execute on a finished product? It happened with PCs (standardization).
The success of start ups demonstrate how specialized providers attract the best talent and operate with tremendous efficiency. Talent is rewarded for contributions and risk toward the end objective. The challenge for start ups is that they can only focus in areas where they are in control of end value creation (think FB, Twitter, Pinterest). How can capital flow to resources in new networks? For example, could a specialist obtain a loan to purchase apartments in NYC to lease them on AirBnB (legality not withstanding).
I’ve covered orchestration in Transformation of Commercial Networks and 6-8 other blogs. The brilliance of Uber is obvious, but just the start of a very big trend. If you were to take apart Uber and assess its core services you would find:
- Message Exchange
- Service Standardization (ie What are you buying?)
- Service Provision
- Trust (Operating rules for drivers, rating system, ..)
- Clearing and Settlement
These same services are present in other successful networks (ie Visa, Verizon, Rail, United Airlines, … ). Per my blog on Network Transformation,
Networks allow for semi-closed counterparty transactions, where markets are defined by their pricing and open-ness. Nobel Laureates Oliver Williamson and Ronald Coase outlined how transaction costs impact the design of firms and small teams. The focus of their innovative insight was the to address the gap between theory of the “pricing instrument” as the central coordinating instrument to a market (macro) and the role of specialist/entrepreneur in defining how companies are organized.
From a broader view, the internet was designed for reliability, anonymity and fault tolerance, not for commerce, value exchange and trust. Orchestration networks allow free flowing resources to publish and discover value interactions to align and form structure. In the last 20 years, resources were large structures formed to build trust and value exchange (eBay, Paypal, Amazon) as well as discovery (Yahoo, Google). Today Trust, Reputation, and Discovery are portable within in these environments (not across them). Your interaction within them enables them to serve you better, and your community reputation keeps you from leaving (information intensity). The separation of identity (ex. authentication), reputation (ex FICO Score) and discovery (below) separate from the network is transformational.
The founder of the internet (Tim Berners-Lee) is seeking to reclaim it from Google and Facebook (see this excellent article in Digital Trends). Tim’s new project, underway at his MIT lab, is called Solid (“social linked data”), a way for you to own your own data while making it available to the applications that you want to be able to use it. This federated data design (Tim’s SOLID/Data Pods) are also core to what we have built at Commerce Signals.
In summation the internet provided connectivity, not the STRUCTURE for resources to discover or transact. The services needed to enable commerce are moving from closed platforms toward new networks AND across networks (ie Blockchain, authentication, identity, discovery) realigning individual resources toward new value equations. In the future resources can be amorphous, adapting to environmental needs AND OPPORTUNITIES.
Orchestration – Value Concentration
Mobile has taken the role of “bridging” the virtual and physical world. Value is shifting from platforms executing transactions to entities coordinating interactions. This interaction of entities is what I refer to as Value Orchestration, certainly not a concept I developed. A January 2001 Harvard Business Review Article: Where Value Lives in a Networked World put it this way:
In more general terms, modern high-speed networks push back-end intelligence and front-end intelligence in two different directions, toward opposite ends of the network. Back-end intelligence becomes embedded into a shared infrastructure at the core of the network (cloud), while front-end intelligence fragments into many different forms at the periphery of the network, where the users are. And since value follows intelligence, the two ends of the network become the major sources of potential profits. The middle of the network gets hollowed out; it becomes a dumb conduit, with little potential for value creation. Moreover, as value diverges, so do companies and competition. …. In a connected world, intelligence becomes fluid and modular. Small units of intelligence float freely like molecules in the ether, coalescing into temporary bundles whenever and wherever necessary to solve problems.
Value orchestration is dependent on data. The HRB quote above painted a picture where “small units of intelligence float freely like molecules in the ether, coalescing into temporary bundles whenever and wherever necessary to solve problems”. If it floats freely …how on earth can anyone organize it? Doesn’t someone need a directory? for at least one side? How can intelligence be “self assembling”?
Story – Google Buy Now is an Orchestration Platform (see blog Google’s Strategy) for advertisers and physical retailers. Local retailer uploads store inventory, google helps consumers discover and buy online.. and ships the goods to the customer’s door twice a day with an Uber like delivery services (FOR FREE). This enables local retailers to compete for online business. Hypothetically, this story shows how the network value of 1M small retailers could exceeds that of an “integrated” Amazon. With each component delivering measurable value (thus accountable to perform) and distributed based upon performance.
As intelligence develops, it will initially aggregate (ex Google/Facebook), then subsequently disseminate. I covered this topic back my December post Building Networks
“The network forms around a function and other entities are attracted to this network (affinity) because of the function of both the central orchestrator and the other participants”.
Given that each node and cluster is resource constained .. each maintains connections to a finite number of “efficient” orchestrators/networks. Early networks build very substantial momentum.
Story 1 - Software as a Service In 1997 I was fortunate to work with Roy Schulte and Yafim Natis at Gartner, the authors: “service oriented architecture” (SOA – 1997). Although this wasn’t a completely new concept (ex. software through contracts,) it seemed to capture the core elements of distributed software development and operation. In the late 90s we had a burst of activity in: software connection (RPC/Sockets -- > RMI-IIOP, SOAP, ..etc), discovery/ registration (UDDI, ASP.NET, MOM pub/sub, ..etc), structure (WSDL, JAX, JSON, …), security (SAML, tokens, …etc). While these services were successful within a company, they largely failed in the vision of distributed applications across company boundaries. Failure? The failure was not technical, but rather a commercial one: there was no path to price a service based upon value. Take a look at If This Than That (IFTTT), there are many services that you can obtain (temperature in Phoenix, price of an xBox at Best Buy, … ) however this is public data, certainly not something that requires exchange of economic value or trust of the counterparty. The problem “web services” is that there is no structure for shared value exchange, which impacts how participants invest AND collaborate (a MUST for any commercial network). This is the reason that universal service directories like UDDI are dead.. (there is no incentive for publishing a web service beyond public data). Imagine if Verizon created a location service whose input was phone number and output was LAT/LONG and uncertainty. This service exists today, but the users are highly restricted (banks are able to ask if consumer’s phone is near the location of the payment transaction). This information not only must be permissioned, it must be priced. Why should Verizon give away a service that has value to many parties. Today most carriers just sell their location data in bulk to SAP, who then resells it to buyers.
SOA was called software through contracts. What we missed in the evolution between the two was the economic aspect of contracts. There will be no flow of information between economic boundaries without a firm understanding of the value that flows as a result of the interaction. The modularity section below provides more economic background. Distributing economic value within collaboration is our core focus at Commerce Signals (and focus of our IP).
Phase 1 Orchestration – Connect existing capability with intent/needs
Because mobile is at the intersection of both virtual and physical, the network is larger.. it touches every consumer, every business and every “cluster”… it is therefore many orders of magnitude more complex. In this dynamic environment, small companies are much better positioned to deliver “focused” if they can collaborate to form a greater combined value (ex dinner and a movie).
New orchestration networks (ie data flows) enable market forces to operate at a much smaller scale. Swarms of specialized communities are hyper meritocracies which react to serve areas that are: #1 inefficient, #2 opaque (ie wiki leaks, banking, advertising…).
The logical first phase of orchestration is in connecting existing resources with demand. In Google Creating Platform for Mobile Economy, and Collaboration and the Sharing Economy I laid out several consumer focused value propositions of connecting existing resources (ex. paint store + painter, dinner + movie) with the primary services of discovery and coordination (ex lead generation, offer presentment).
Who is best able to execute this phase of orchestration? The entity with the best consumer relationship with the highest frequency of touch (Google, Facebook, Amazon). Take a look at the dramatic progress that both Facebook and Google have made in mobile advertising (see this NYT Article). The growth of mobile ad dollars follows consumer time spent, and mobile’s unique ability to target individual consumers based upon data/behavior is highly disruptive. This dynamic is obviously behind Verizon’s Yahoo and AOL acquisition strategy.
Phase 2 – Service Discovery and Measurement
In a normal market the price of a good is based upon supply and demand. Data is in infinite supply, and its value is based upon use. No market (or network) can operate without a pricing function. Measurement of value (based upon use) is thus critical to both resources and networks as it enable the pricing function of each entity (quantification and dissemination).
Within Google and Facebook measurement is a top 2 strategic priority. If advertisers agree on the methodology of measurement, there will be no need to bill on clicks or impressions but on sales.
In addition to solving the value exchange problem, measurement also allows automated discovery and correlation. For example, Google’s initial “pagerank algorithm” was a crude form of measurement. The value of any specific link is based upon the number of links that are associated with it. Obviously it is not possible to ascertain the economic value in this approach, only the relative value vs alternatives. Thus Google’s approach works well for public “free” information, not for private data. The value of private data is based upon use, as is its availability.
Measurement is thus a core function to pricing, discovery, modularity and capital investment. Per my first question “How can any investor trust a collection of 5 interdependent companies to execute on a finished product?” if the value each resource creates is measurable, each can obtain funding. Commerce Signals works to enables measurement, and thus the discovery and pricing of federated data (for the benefit of the data owners).
Modularity and Economics
Before beginning a new topic it’s good to have a baseline for the definition of the term technically and economically. Modularity is the key technical term describing how units interoperate (specifiability, measureability, predictability, interfaces). Per my Distributed Innovation blog, Legos® are a great metaphor for standardization (although not offer a great model for value sharing).
Story – During my 8 yrs at NASA modules were how we managed internal systems complexity. Every unit required only limited input Signals to operate (which is where we get the Signals in Commerce Signals). Through “modules” each system can be manufactured and tested individually AND also as an integrated whole. The precepts of modularity above remain consistent both here, PCs, phones and our human bodies.
Most engineers obtain this view within their systems and controls class (ie Laplace Transforms). For signals to be acted upon they must be consistent and understood within the operation of a system. There are signals with a high degree of uncertainty, for example navigational location and attitude, but action still must be taken.
Modules are also a wonderful construct to allow people to specialize within a field, evolving the performance of single components in complex systems (Intel and Moore’s Law). The evolution of modules follows an Innovators Dilemma (ie over performance) curve. Within complex commercial organizational systems the design of “boundaries” (within Transaction Cost Economics – TCE) conceptually mirrors many of the interfaces of modules. As opposed the the “technical” interaction and testing of modules, TCE is focused on the economic optimization that combination of functions have within an overall system.
TCE is a commercial construct with deep economic thought, Modularity is technical construct. Let me attempt to bridge. From the MIT press we have Design Rules: The Power of Modularity which covers BOTH Modularity and TCE. Directly from this book, I’ve given a quote on the economic view of modularity:
From pp 366-370,
What happens when a newly modular design emerges in an advanced market economy? A modularization, as we have explained, multiplies and decentralizes valuable design options. Market economies are decentralized complex adaptive systems in their own right: their members have the ability to act unilaterally in response to local calculations of value. Hence it is interesting to investigate what happens when a design laden with valuable, decentralized options “hits” an economy that permits and, indeed, rewards decentralized action and initiative.
A modular design makes possible decentralized design evolution. In the presence of advanced capital markets, a modular design also makes possible decentralized industry evolution. In other words, when an artifact with a modular design is created in an economy with advanced capital markets, subindustries of firms and markets organized around modules may emerge and evolve in parallel with the module designs themselves.
…. the “microstructure” of designs affects the economics of design processes in deep and unavoidable ways. In particular, the many coordinating links that are needed to implement an interdependent design process will have a profound impact on the costs of getting things done under different contractual regimes.
Tom Note: the modular design may be in the control of a network or platform that sets prices or controls sales channels (ex Boeing) impacting how each component can expand beyond a given supply chain. Also important to note that designs that reside within advanced capital markets (read exposed to market forces with access to capital) are held accountable for performance as competitors are funded.
For this reason, the interdependencies in a design and task structure will influence the optimal boundaries of firms and the location of markets in the surrounding industry. In this section, we will make this reasoning more precise, arguing as follows: In the presence of moderate transactions and agency costs, putting interconnected tasks within a single corporation is a uniquely efficient way to “package” the task structure. To make this argument, we must first define what we mean by “transactions and agency costs.”
For our purposes, transactions costs are the costs associated with a formal transfer of property. They can in turn be broken down into:
- information costs— the costs to either party of verifying the quality of the property being transferred, as well as the costs (also to either party) of verifying the value of the payment being made for the property;
- operational costs— the costs of effecting the transfer once the parties have agreed to it (including transport costs, clearing costs, etc.).
Agency costs are the costs of delegating tasks to other human beings, whose objectives are not identical with those of the delegator (called the “principal”). For their part, agency costs can be broken down as follows:
- direct costs— the cost of the work not being done as the principal would like it to be;
- bonding costs— the costs of changing the nature or timing of the agent’s incentives so as to reduce direct agency costs (e.g., the agent may post a bond,
…Since the publication of Ronald Coase’s seminal paper “The Nature of the Firm,” it has been recognized that firms (generally) and corporations (specifically) are prearranged contract structures, whose function in part is to reduce transactions and agency costs. By setting up a firm or corporation, it is possible, not to eliminate transactions or agency costs, but to reduce the amount of friction these costs introduce into the flow of day-to-day work. The reduction of frictional transactions and agency costs may be accomplished in two ways (which are not mutually exclusive):
- A group of transactions may be standardized and regularized, so that their variable cost approaches zero.
- Claims on the ending value of a collective effort may be created and used to compensate agents for their work.
For example, an entrepreneur might create a standardized set of employment contracts with designers that included a reward proportional to the value of the final product. By the act of standardizing the terms of employment, the entrepreneur would be using Coasian logic to create a firm….
Hence, for a Coasian firm to become a corporation, all that needs to happen is for that firm’s various transactions, contracts, and claims to be recognized as valid transactions, contracts, and claims within the larger economic system in which the firm operates.
Tom Note: Transactions and contracts across boundaries are difficult to manage because of trust. A contract can’t be trusted beyond the counterparties engaged. This is where commercial networks succeed. Visa enables millions of parties to interact through a common network agreement with standardized transactions. Thus a commercial network is a expands boundaries for standardized services with the core services of trust and enforceability. Whereas a platform is a service definition, a commercial network is an economic and contractual construct.
The contract-structure boundary of a corporation refers to an invisible but economically significant boundary that separates the “inside” from the “outside.” Property and actions that arise inside the boundary can be rearranged using the corporation’s internally sanctioned procedures. However, property and actions outside the boundary are not under the corporation’s direct control. When property (or action) crosses the corporate boundary, an event occurs in the larger economic system. For example, the sale of a product, the issuance of a security, and the hiring of an employee are all boundary-crossing events.
Thus a corporation is a social artifact that provides for a reduction in frictional transactions and agency costs within prearranged contractual boundaries. But what should those boundaries be? …. The economic advantages of having a single corporation obviously depend on the differentials in transactions and agency costs. If transactions or agency costs differentials are very low, then even with a high degree of interdependency, the advantage of having a single corporation will be slight. Baldwin, Carliss Y.; Clark, Kim B. (2000-03-02). Design Rules: The Power of Modularity: Volume 1 (MIT Press The MIT Press.
Technical Future – Modularity
Modularity is a mechanism for managing complexity. Each “module” must understand its boundaries and how it interacts with the environment, dependents and interfaces. Successful modules must display the core tenants of: predictability, specifiability, verifiability (measurement).
Perfectly modular designs do not spring fully formed from the minds of architects. When a design is first modularized, the designers’ knowledge of interdependencies is usually imperfect, and as a result the initial set of design rules will be incomplete. Incomplete design rules give rise to unforeseen interdependencies, which in turn will require consultations and iterations between the hidden-module designers and the architects of the system. The integration and testing of these designs will be fraught with difficulty and the designs themselves will have high risks of failure. As the properties of the system and the modules become better understood, the design rules will tend to become more complete. Then, as more of the innate interdependencies come to be addressed in the design rules, integration and testing of the system will become more cut-and-dried. Baldwin, Carliss Y.; Clark, Kim B. (2000-03-02). Design Rules: The Power of Modularity: Volume 1 (MIT Press The MIT Press.
Platform as a Form of Modularity
As I outlined in Sharing Economy, the standards you adopt and the services you support impact how you compete and collaborate. Platforms are a form of modularity where the interfaces, standards and testing are defined by a central party. In Platform Leadership: How Intel, Microsoft and Cisco Drive Industry Innovation. The authors outlined 4 Levers of Platform Leadership.
- Scope of Firm: What is done inside, how they encourage outside investment and focus
- Product Technology: Architecture, Interfaces, Modularity, What do they expose to partners?
- Relationship with Complimentors: Support of Complimentors, acting on ecosystem needs, path to consensus and standardization, profitability
- Internal Organization: What is the “core”, and how are resources allocated to core activities vs support for partners.
The PC Platform (defined by Intel Architecture Labs) is the most well understood platform for the economic discussion of modularity. Within this platform, modules which have differential performance capture higher margin (Intel’s chip vs the Power Supply). The assembler (PC integrator’s) margins are below 2% (Dell, IBM, …) where Intel’s (orchestrator) are over 60%.
Hypothesis: the entity which controls “the standard” within a modular network is the most profitable (Apple, Mastercard, SWIFT, …).
As most of you know, last month Google just killed project Ara (modular phone). Creating an “open spec” where anyone could assemble a mobile handset. This project could have driven a stake through the heart of Handset OEM vendors. Google’s economic incentive: driving consumers to Android and the Google Universe. Where Apple is the most wonderful integrated consumer product company in the world, Ara would be a nerd nirvana. Integration makes sense where there are significant risks, or costs in integrating components, or where integrated product significantly out performs best of breed, but not where all integrated products are good enough (… thus Apple’s current iPhone state).
Modularity in these two examples exposes commodities (ex power supply) to commodity pricing and allows specialists to reap margins aligned to performance (independent of integrator). Thus the technical construct of modularity changes the transaction cost economics of an integrated group of collaborators. Standards do not benefit poorly differentiated commodities, with low switching costs, as they have weak influence over demand or supply.
The problem I have with a “platform view” are: #1 closed ownership, #2 static nature of the boundaries and interfaces (resource creation/assignment) #3 few successes and natural metaphors. While static ‘platform’ models may serve as initial scaffolding in resource allocation, there needs to be a much more rapid way to discover capabilities and “assemble to value”. I have no desire for our future to look like the Borg (of Star Trek).
Investor Note – What to look for?
When assessing investment within this hypothesis, I recommend the following:
- What industries are: #1 inefficient, #2 opaque, #3 poorly serve the consumer or #4 at risk (ie Darwinian)
- What new standards are developing and who controls them? (Open – Blockchain, Mandated – SEPA, Controlled – Intel PCI)
- Who is investing in the network and what is the economic model?
- Where is top talent flowing?
- Network revenue trends (resources/sellers and buyers)?
Networks at Risk – Examples
What industries do you think of when answering question #1 above? Mine are healthcare, financial services and advertising. Retail banks in the EU are just the first to suffer through forced standardization (SEPA, PSD2, OBWG), with cool startups owning the front end and banks left to manage risk and regulatory compliance.
Bank economies of scale are atrophying due to structural changes in consumer behavior (ie mobile). Consumers discover “commodity” financial services differently today, per last month’s WSJ article non-banks have replaced banks as top mortgage originators. Per my blog Changing Economics, Payment Networks and Handset OEMs have reinvented risk management (ie Authentication) and distribution (ie issuance cost of 0) enabling smaller banks to compete at scale with larger issuers in payments. I’m not alone here, as Wellington (the largest institutional investor ~$1T AUM in the US) is also the largest holder of small bank equity.
Scale and Risk mean nothing in a world where Merchant is closer to the consumer and identity and risk can be separated from the specialist network (ie bank). In this world merchants are better positioned acquire customers for instant credit (acquire or securitize). The margin is in acquiring the consumer, not in delivering the loan. For example did you know that over 65% of Kohl’s transactions are on their private label card.
My primary investment test is #3. The beauty of Visa, Mastercard, Uber and AirBnb is that millions of businesses invest billions of dollars to make them work. Shared economic incentives are keys to a resource allocation, but network participants must have “influence” on pricing. Within payments issuers have it, but merchants do not.
NYU’s Thomas Philippon published jaw dropping research detailing how Payments and Banking are one of the few network businesses in the HISTORY OF MAN to grow less efficient (rail, telecom, energy, …).
Within Uber neither consumers or drivers do not have pricing control. These are not signs of an ineffective system, but one that could face competition.
How does Small Win?
- Data Federation and Ownership (Solid, Apple, Commerce Signals, …etc)
- Resource discovery and measurement
- Modularity and Semi Open Standards (to Participate)
- Right to Assemble – Regulatory, Resource Access
- Create Measureable Value
- Access to Capital – Financing based upon reputation and value measurement
As always your thoughts are appreciated. If you are energized about any of this, please let me know how we can work together.