General Causation of Antitrust Activities in the EDI Communications and Value Added Network Market:

Interconnections, private equity, and architectural deficiencies supply the leverage to dominate today’s critical supply chain communications markets.  

A.D. Wilensky, Analyst, BizQuirk Strategy

Abstract

Interconnection dependent markets are very different than product driven markets. The participants in interconnected marketplaces are influenced by so called ‘network effects’ - a phenomenon whereby interconnections between networks influence the subscribers residing on the networks, thereby creating a composite of a market’s total value for all of the participants.^[1] Stated simply, no interconnections between networks, no market for network services.^[2]

There has been a great deal of research on the topic of network effects^[3]. In the area of antitrust and telecommunications law, the courts and regulators^[4] have struggled to clarify the economic issues pertaining to networked markets, and how the policies of providing or withholding interconnections to one’s lateral competitors influence the market. When negative influences arise, and are magnified, as a result of interconnection abuse and network effects, there should be expedient mechanisms for resolving anti-competitive fallout. In unregulated markets, as in the present case, the courts have found it difficult to rectify damages inflicted upon competitors overcoming overt interconnection denials by the over-sized incumbents within their particular market sector.

In this briefing, we focus on the anticompetitive effects arising from (or reinforcing) the abuse of market power by a singular, large EDI transaction network (VAN or Value Added Network). By exploiting a large, influential subscriber population gained via acquisitions, GXS Inc. has come to leverage its inter-network connections (interconnects), making its dominance all but absolute; GXS has passed the point where econometrics or revenues adequately portray the company’s unchallenged ability to bar any competitor from the market, or its potential ability to inflict harm on its existing competitors at a time of its choosing. It is where and how GXS applies these market levers that bear on the current case.

Finally, we should acknowledge the positive aspects of network effects; in the present case of Value Added Networks, cooperative interconnection policies built the industry, fostered a diverse market, allowed subscribers freedom to choose among many providers (and the ability to switch providers), and finally, new ventures entering the market were able to innovate and offer creative solutions.

All of the attributes of a diverse, open, and healthy market, an EDI market offering the same types of network ubiquity that we as consumers take for granted in our phone and Internet services - is being hobbled by the interconnection practices of GXS Inc. The short-lived fecund stage of EDI’s renaissance lasted for a brief year or two in the late 1990’s, until GXS’ ascendancy was cemented during the 2005- 2009 years which witnessed the consummation of the key GXS acquisitions; the GXS campaign that led to its hegemony began much earlier, however^[5].      

Finally, all of the prized characteristics attractive to risk capital, i.e., end-user agility and the incentives to innovate are being denuded in today’s EDI market due to the interconnection foreclosures by GXS. Any market influenced by a monopoly is inhospitable to seed and mezzanine investors - a state describing the present EDI sector as dominated by GXS^[6].

Market Overview           

Markets that derive value from network effects differ immensely from those deriving their value from the sales of goods or services; widget bound markets may become subject to network effects, tying agreements, etc., and may also be affected by interconnected services, such as road, rail, freight, pipeline, and of course, networked communications services. These dynamics have simply become pervasive in our times.

In the case at hand, Value Added EDI networks, we have systems predating the Internet proper, that are loosely related to modern email systems, although there are important distinguishing factors that bear on the case. For the purpose of this brief, however, let us agree that EDI networks, called VANs, are simply private email systems used to convey standardized business documents destined for delivery to automated systems, i.e. inventory databases and retail systems, as distinguished from email communications, destined for delivery to people.      

Value Added Networks (VANs), are private messaging systems used to convey common business process documents in standardized formats: invoices, purchase orders, and various shipping acknowledgements are sent and received via approximately 40 domestic VANs supporting various supply chains of the major retailers, manufacturers, and package /courier/ freight common carriers. VAN services comprise a 1.5 billion dollar global services market - a figure that belies a system that supports the transactions representing more than a trillion dollars of annual domestic and international trade. EDI transactions are conveyed in real time, to and from real companies of all sizes. If any one of the major EDI networks (VANs) were to fail prior to the holiday season, there would be no cheer.

EDI communications are therefore not inconsequential, especially for small and medium enterprises forced to comply with strict EDI transactional specifications dictated by the mega retailers and manufacturers they conduct business with. Servicing the likes of Ford, GM, Target, or Sears, forces suppliers to accede to the realities of implementing the guidelines of these influential trading partners, the supply chain buyers.

The lone market competitor of similar scale to GXS is IBM-Sterling, whose interconnection practices stand in harmony with the EDI industry’s tradition of granting non-settlement interconnections to bona fide network operators. The practice has not compromised Sterling’s ability to generate new business or retain accounts, and comports with IBM’s philosophy of having its acquired business units act as stewards within their market niches. The author will extract a further lesson later in the monograph, regarding the cleavage points distinguishing these two companies.^[7]   

VAN politics in an unregulated market space - who may play, who shall decide?

It was previously explained that there is a substantial top down influence emanating from the large trading partners to their subordinate suppliers. The larger partners can, and often do, dictate the means by which suppliers communicate with them, and this might even include not using VANs at all  - an increasing option that is illustrative of the colloquially termed, VAN Exodus’. But, for better or worse, today’s EDI market is dominated by VANs, which means GXS and IBM-Sterling own the top 70% or more of the large purchaser market. Restated, Retailers and Manufacturers often dictate all sorts of terms to their suppliers, such as which VAN ID to transact with, the technical composition of transactions (transaction set), etc., and, (outrageous though it may be to the Spirit of Free Enterprise) that subordinate trading partners use the same VAN as the purchaser. Though his last requirement  is not widespread, it is also not unheard of in the industry. Such requirements by a major retailer dictating ‘same VAN terms’, sound as outrageous as your uncle telling you to, “use AT&T like I do, or you shall have no inheritance.

Resisting EDI requirements essentially results in suppliers forfeiting their trading relationship with major purchasers, making EDI crucially important to the survival of downstream members in the supply chains.

One interesting artifact of this lopsided influence arising out of VAN network effects, is that large VANs may avail themselves of impressive and influential ecosystems of clients that are sum total buyers ; and as borne out in practice, such VANs may in turn misuse the influence of such clients to the disadvantage of other EDI networks catering to the smaller and mid-sized enterprises making up the great majority of suppliers. So, in the EDI business, a VAN gains influence not only by the number of clients, but by the type of clients.  

Influence gained by a ‘weighted’ client type population, is further exacerbated by exploiting a VAN’s  routing access (interconnection arbitrage), decoupling the bad actor from typically reliable econometric measurements of simple percentages. For example: 1000 clients on VAN A, a mix of buyers and suppliers, is different in quality than VAN B, composed mostly of influential retailers and manufacturing hubs. And astounding though it is, these key clients are often blissfully unaware of the VAN’s intention to exploit their cachet. Let’s, however, not mince our words here -  the only VAN conducting this practice as to court legal and regulatory scrutiny is GXS, as its nearest competitor, IBM-Sterling, has never come close to such shenanigans.^[8]     

Simply recapitulated, major retailers (or manufacturers) are “must have” EDI message paths, and the suppliers need them to survive. Concomitantly, VAN interconnections to competing networks thereby become an ideal abuse mechanism to achieve the power to foreclose market entry to competitors, new ventures, and most certainly, technological innovators.

Market Effects: EDI providers wielding unequal Influence via their unwitting clients

In the absence of collegial interconnection policies, the concentration of a large cohort of purchasers on a particular VAN creates the potential for toxic network effects - as these large hubs tip the scales of influence towards a VAN acting with ill intent. Interconnections between VANs and other specialty commerce communications providers are the bulwark against such behaviors, keeping the market honest, prices competitive, and enhancing the climate of innovation and end-user choice.

Interconnections are the ultimate market lubricant; they expand the market reach of all participants while preserving the interests of the end-users. There is nothing else quite like interconnected markets. The specific practice of non-settlement peering between the VANs, an established and accepted practice for years, built this very market. The fight to preserve the policies and practices of agnostic, non-settlement interconnection the one concentrated incumbent seeking to convert peers into payers^[9], is absolutely crucial to the industry’s survival.  

Innovators into the Breach-  Electronic Commerce Service Providers and the EDI market’s lone Communications Specialist Network embark into hostile territory.

Not much earlier than the 1990’s (some say as late as 2005) it was hopeless for SMEs to engage in EDI automation with trading partners, due to the costs and technical requirements.   Accommodating the SME’s became the prevailing trend, beginning with a handful of innovative IT ventures seeking to take the sting out EDI. These New Age Electronic Commerce Service Providers (ECSPs) sought to democratize entry into automated supply chains using EDI documents that were mandated by the largest purchasing hubs, the great majority using the dominant VANs for conducting EDI communications. The Service Providers were faster and more innovative the classic VANs, taking a bite out their profits, as SPS Commerce shook up the industry by shrinking the profits of GXS. SPS, unwilling and quite correctly not eager to replicate a message routing infrastructure, took up with Loren Data, in order to prove out Todd Gould’s ECGrid, a message switching built for EDI multi-tenant operators, SPS being one of the first to offer an on-demand alternative to VANs.

Loren Data Corp gained respect in the industry by being one of the first to pass some of the most onerous Federal Purchasing systems EDI testing regimes. By directly serving Service Providers, rather than end users, Loren Data positioned itself in a market gap not served by VANs, who make their bread and butter on end users directly. Loren Data Corp specializes in delivering targeted and specialized communications, API’s, Directory Services, etc. - they delivered a better class of pure wholesale connectivity while the vast majority of the VANs offered basic FTP and AS2 connections.

Todd Gould, Loren Data Corp’s founder, engineered specialized, differentiated unbundled communications that every growth industry needs - and today EGridOS is still the only API for EDI network management. This must be profoundly vexing to the Executives at GXS, who are finding it difficult to compete with Loren Data Corp on a level playing field of technology, value, and performance.  
 

Interconnects in EDI Networks -  Using Routing policy as a Lever

Non-settlement interconnections between collegial networks created today’s EDI market. The history of AT&T’s rise to monopoly power mirrors the travails between Loren Data Corp and GXS, Inc. Though topologies may differ in minor details between telephone, Internet, and EDI networks, the early march towards a consolidated market dominated by one actor is a virtual repetition of actions that seeded the history of telecom antitrust case law. The corporate character of the litigants is also somewhat reminiscent of the parties populating landmark antitrust cases often mentioned in the post-divestiture era.

The issues are similar as well to the landmark MCI v. AT&T case, with Loren Data Corp playing advocate for liberal interconnection, technical modernization, and end-user freedom of choice, while GXS Inc. takes the role of consolidator-monopolist, and technological laggard. Although, AT&T has played the laggard at times, they have contributed substantially to the telecoms market via numerous research laboratories, academic sponsorship, etc.

However, we see in the case of GXS that private equity has facilitated the actions of a laggard who has bought its way into a position of dominance. We can posit therefore, that thought leadership and technological innovation are entirely dispensable when seeking to dominate a communications marketplace - this is not a revelation at all.                

The  present matter of Loren Data Corp v. GXS Inc. is an almost identical case of interconnection and market abuse via these interconnections, buying other networks along with their subscribers, leveraging private equity, and subsequently arbitraging these VAN interconnects. GXS’ buyouts of IBM IE and Inovis flew past FTC, DOJ, and UK Competition regulators - all of these agencies were perfectly uniform in misunderstanding the the particular deficiencies of EDI routing policies that are exploitable by toxic actors.          

A familiar history of a rapidly reconstituted monopoly

VANs, or more generically, the EDI communications industry, grew from a few major IT communications specialty providers in the late 1960’s, and provided that one specialty service - conveyance of electronic business documents in their early non-standard formats. We can compare EDI messages to the email of today, with a caveat that the recipients are not people, but are rather computers used in processing automated inventory, manufacturing supply chain, retail shelving management, and ever-popular Enterprise Resource Planning software, made ubiquitous by SAP.

VANs provided secure communications, connections to trading partners, and a modicum of expertise. The early technologies wedded communications to specialized professional services that came to be known as ‘The Value Add’. Thus, Value Added Networks were indeed private, wireline facilities, almost indistinguishable from the telegraph, telex, and TWX wire communications services that were well under the regulatory authority of FCC titles and tariffs. However, VANs escaped becoming subject to the regulations befitting a sector that handles the data representing billions in supply chain and durable goods / raw material data, not to mention crucial Federal and Military logistics data in those critical supply chains. Early VANs, amounting to a handful of highly specialized systems operating during the period of 1960-‘80’s, simply did not leave a trace on the regulatory radar.        

The most popular and compelling value added service of the early EDI era was the ability to connect any two large, influential companies in a supply chain relationship; a corporate purchasing department database, a complex mainframe, would generate and send orders via a VAN, and the top tier supplier(s), almost always Fortune 500 behemoths, would receive and subsequently ingest the transaction, thereby automatically updating their inventory or shipping status. The science of EDI reduced the human effort and errors that were common in the high volume ordering and fulfilment systems of the time.

The recipe for concocting a VAN monopoly is quite similar to other markets: 1) enter a market with existing cooperative interconnections, where 2) network operators participate in a fully interconnected communications marketplace, and 3) offer non-settlement terms (do not dun other networks for bi-directional traffic flows), 4) secure private capital, 5) buy the top two or three VANs and software assets, and 6) begin the process of interconnection foreclosures.

This formula has worked in other interconnected markets, and was often overlooked by regulators. The remedies that are eventually assessed are trivial, or take decades to fully resolve. The AT&T consent decree and subsequent divestiture took almost fifty years to complete, covering four Presidential Administrations, the creation of two new agencies, and the participation of one far seeing yet crafty AT&T VP, the much vaunted Kingsbury.

Future Remedies; the suitability of Courts of Law in Resolving Interconnection Disputes

In order to remain viable, competitors must overcome network access hurdles (placed by incumbents), acting with the utmost alacrity, while navigating legal processes that are routinely exploited by incumbents with superior legal resources. The incumbent often accomplishes as a defendant what it could never hope to attain as a competitor in an industry with rational interconnection policies.

Lawsuits require remedial explanations and evidence based metrics designed to spare the Jurist, and to an extent, the plaintiff’s counsel, who are both non-specialists. However, cases of interconnection denial are complex, and extant antitrust laws are simply ill-suited to correct the abuses in an expedient manner, in this case^[10].

The industry would much rather see expert administrative panels delivering binding directives. Until that time, the courts will have to resolve the most dire disputes in the EDI Communications market, a sector properly belonging to wire-line enforcement, one that fell prey to the Internet’s all consuming grasp in the course of the 1996 telecom reform act. There may be more direct ways of resolving the problems associated with interconnections in an unregulated market, and the author will include some thoughts on alternative legal approaches in the appendices of this monograph.      

     

Misnamed: Value Added Network Interconnections are, in truth, Message Forwarding and Traffic Exchange Agreements.

Once again with great feeling: VAN interconnects are not classic interconnects, they are agreements to exchange message traffic. As currently used, the soft configuration of a VAN interconnect is nothing more than the entry of technical parameters. After configuration, most VANs never touch the ‘interconnect’, unless there are client issues. Therefore, these agreements are between VANs that share no physical links, WAN hardware, or other special provisioning. A heavily used ‘VAN interconnect’ might require the most pedestrian accommodations, such as increased memory, or perhaps the addition of a commodity class Server.     

EDI messages are subject to bailment, or should be, as the message in its equivalent physical form would be treated as such by the common carrier. Although parties subject to the strictures of the taking of message bailment have not been put to a legal or regulatory test, this cannot be far in the future, and is in fact, overdue for adjudication by a competent authority.    

VANs currently operate within a hairsbreadth of the same regulations that governed Telex, TWX, and Telegraph Messaging Services. If the agencies would focus, we would see VANs regulated under the concept of taking of bailment, i.e. the transfer of responsibility to deliver a valuable EDI message.

On Interconnections, briefly.

Interconnections are special. They are more than essential features of networked marketplaces - Interconnections create the total value of our connected industries. Abusing interconnections in order to exploit and limit competition is a shortsighted tactic, a limiting and uninformed strategy that hobbles innovation.

Interconnection abuses led to AT&T’s multi-decade monopoly - an era of limited services and artificial scarcity. A similar mindset marks current B2B IT services marketplace as a pool of limited opportunities, where the largest trading partners are captives. Such thinking is simply outmoded in this era of ubiquitous connectivity, where the true costs of data transit is becoming vanishingly inexpensive.

The only antidote to the malady of interconnections abuse, or arbitrage is the establishment of open, equitable interconnections accessible to all bona fide network operators specializing in EDI communications, but more accurately termed “commerce communications’, not limited to  x12 and EDIFACT EDI.

So what are the effects of a refusal to carry the message traffic between two contractually bound trading parties?

1) A refusal to carry EDI messages between trading partners located on selected EDI networks’ (in the present instance, a refusal to interconnect the TGMS VAN with ECGrid®) on the terms granted to other commerce networks, is a fundamental distortion of the market.

2) Applying selective settlements to specific network providers dishonors the trading partner’s right to choose freely among EDI networks or service providers, as they see fit. Interconnection denial by one large network created the era of the telecoms monopoly, which left an enduring legacy of distorting the markets and limiting innovation.  

3) Minimizing the intrinsic value and importance of each and every EDI transaction. Every message in the supply chain is equally important, regardless of point of origin, buyer or seller.

4) The creation of a lopsided market dynamic is being accomplished by transforming interVAN connections into instruments of competitive leverage, often by exploiting the relationship between TGMS VAN clients and their trading partners.  

All of the foregoing are important and urgent in their own right, but the keystone issue is Interconnection Denial, and the perpetration of routing gateway interference, i.e., using the prestige of the (unwitting, unknowing) trading partners on the TGMS VAN, to influence and disadvantage a competitor.  

Appendices

Fact Sheets

Loren Data Corp Facts:

Founded: 1987 Founder: Todd Gould

Branded Service Offerings: ECGrid®, ECGridOSSM, FBO Daily, WorldWide EDISM

Staff, Full-time + Contract: 5

Key Persons: Todd Gould, President, CTO | Crystal Kuczynski, VP of Network Operations | Kristine Finlay,GM | Shelly Donkin, Network Operations | Alan Wilensky, Product Strategy + Industry Relations

Primary Line of Business: EDI Message Routing

Evolution of Lines of Business: Web Services Message Switching, Data and Partner Integration Management

Forward Lines of Business: Component Commerce Network as a Service,

Current Gross Revenues: ~$1mm Projected Gross Revenues end ‘011: ~1.5mm

Infrastructure Type: SAS-70 Collocation in Portland, OR and Los Angeles, CA

Architectures: Windows, SQL Server, .NET Framework Web Services, IPSec VPN

Featured ECGrid® Clients: SPS Commerce, Trubiquity, CovalentWorks, Energy Services Group, Radley, Activant, Pinnacle Data Services,

Incorporated in: California

General Offices: Indian Rocks Beach, FL

Network Operations: Colorado Springs, CO

ECGrid Network:

Network ID’s Total Active + Registered Inactive + Test = 19,000 Entries

Total Trading ID’s active in last 6 Months = 8-10,000

Transactions Avg. Daily: 30,000

GXS TGMS + Inovis + IBM IE (G Inc.) MegaVAN#1

GXS is the combined organizations of the former GE Information Services (GEIS), Inovis, and IBM’s IE VANs, plus the resources of their catalog of software and professional services acquisitions:
GXS acquires RollStream, a Enterpirse 2.0 B2B service.
GXS Merges with Inovis June 2010

GXS becomes the world’s largest network of integrated business communities. Verified by 3rd Parties Gartner, Forrester, others.

GXS Acquires Interchange Services December 2008

Interchange Services, a B2B integration services provider focused on the Brazilian market.

GXS Acquires UDEX November 2006

UDEX, was a for profit quasi-industry standards body, like a GS1, providing data cleaning, meta-data, and data quality services for the retail sector in the US and UK.

GXS Acquires G International April 2005

Acquired G International, a IBM’s former EDI VAN and Business Exchange Professional Services Group.

GXS Acquires HAHT Commerce February 2004

Acquired HAHT Commerce, a provider of Customer Order Management and Product Information Management applications for chemical, high tech, and consumer products industry.

GXS Acquires Celarix June 2003

Acquired Celarix, a market leading provider of Logistics Visibility applications offered in a Software-as-a-Service (SaaS) model.

Numbers: Analyst’s Say:

GXS has over 100,000 trading partners on network

GXS says 40,000 before the Inovis buyout

Two independent analyst bureau’s have called GXS undeniably the largest VAN

2200 employees

450MM++ annual on last SEC filing

They  are Big.

Alternative Legal Application of Laws Governing Telegraphy as Applied to Value Added Networks.

Alternatives to the Sherman Antitrust Act - Are VANs not properly classified as

The Community of Analysts Complicit in the Perpetuation of a GXS Hegemony

The Industry’s True Momentum