From Computer Business Review, a sister publication

Electronic data interchange (EDI) has always been a technology of promise. Can its marriage with the Internet finally create mass- market appeal? Dan Federman, president and CEO of Premenos Corporation, a company which might be described as the uncontested leader in an untested market – a place where the dull and worthy world of electronic data interchange (EDI) touches the emerging commercial frontiers of the Internet – admits It’s boom or bust time. For Federman and the many other suppliers and users which have invested heavily in EDI, a technology which enables fast, application-to-application data exchange across wide area networks, the commercialization of the Internet, along with the development of new security standards, is opening up huge possibilities. There has never really been any choice but to use a secure private network for EDI. But what are the possibilities, he asks, now that we are in a position to be able to make the EDI transactions themselves secure and self- authenticating? The suggestion is that the EDI marketplace is about to take off. It would not be before time. Only 60,000 to 80,000 of the 6.2 million registered businesses in the US have shown anything but a passing interest in EDI, a technology which was once viewed to be of enormous potential. Market penetration is still slight, even on the back of double-digit growth rates throughout the 1970s, 80s and 90s.

By Kevin White

All along, the possibilities of EDI have intrigued the business world. In theory it enables companies to speed up and simplify the processing of many routine commercial transactions. Mundane paper-shuffling and the manual tracking of order processing, invoicing and dispatch can be automated out of the operation. Many of these companies that have used it – large retailers, manufacturers, insurers and the like – have reported substantial benefits. For most, however, the technology and cost of EDI remains daunting. But here is a new hope: Could recasting the EDI transport mechanism from a private members-only value-added network (VAN) to the public access Internet do the trick and kick start the market? The EDI market to date has been driven by a small number of large corporates insisting that their trading partners implement EDI. Small companies just haven’t seen the value in making themselves EDI-ready and large corporates have had to persuade them with their purchasing clout, says Steve Youngblood of Harbinger, a company selling trading-partner implementation services to large corporates such as Mobil and Kmart. Harbinger has been doing well at this business, growing 50% a year for the past five years. But now Youngblood believes EDI can be sold to small companies as a useful technology in its own right. The objective is to capitalize on the Internet, says Youngblood. To this end, Harbinger has entered a $12 million joint venture with telecommunications services company Bell South. The first offering, called Trusted Link, is a blend of EDI products and services aimed squarely at the small business. Part of the product set is a $500 Windows suite which will Internet- enable any EDI application, acting as a kind of ‘EDI middleware’, handling encryption, message-logging and managing the audit- trail. EDI-server subscription services are also planned to provide smaller firms with a clearing house for their Internet- based EDI messages and a gateway to the world of the established private VANs, which are currently dominated by large corporations.

Software smorgasbord

Whether the ‘EDI for all’ approach will provide Harbinger with a means of super-charging or even maintaining its growth is debatable. Analysts believe that if EDI is to successfully combine with the Internet, then implementing it must be easy. For EDI, while pitched as a way of speeding up and simplifying daily business life, is notoriously complicated in practice. Much of the complexity of EDI messaging lies in the notion that it requires no manual interv

ention. But this requires preparation: in a straight transfer of EDI data the computer applications of both sender and recipient must exchange messages in identical file formats. In this sense, it is quite different from firing off electronic mail or sharing files across a network, and raises all sorts of problems. One of these is the ‘software smorgasbord’. In order to switch message formats back and forth between one of many available standard EDI formats and the local software environment, it is necessary to use EDI translation software. This software, in turn, needs to interact smoothly and efficiently with every type and flavor of off-the-shelf and custom-built business system. This is not easy: standard EDI interfaces are still sorely lacking, even among some of the more popular accounting suites. Another problem is ‘EDI interrupts’. A great many businesses are only prepared to meet their EDI- committed trading partners half-way, complain EDI evangelists. These organizations, most of which have been compelled by commercial pressures to use EDI, choose to add time and costs to their existing business processes rather than fit proper EDI hooks into their core applications. For example, some companies have blanked the automated approach preferring to print and then re-key EDI message receipts – line by line – into their back- office programs. A further problem is the lack of a digital lingua franca. EDI use is fragmented, comprising a myriad of vertical, private trading communities in retail, finance, construction, shipping and automotive sectors. Each of these groups exchange messages using their own preferred transaction sets based on any of over 200 ‘standard’ formats. All of these issues make installing EDI a significant challenge. A small company supplying several large retail chains, for example, may find itself having to set up multiple separate and often incompatible EDI systems. Many companies, which lack either the technical know-how or the trading volumes to make EDI seem worthwhile, view electronic trading as an added and unnecessary cost rather than a way of simplifying and extending the trading cycle.

Internet EDI

With all these problems, it is not surprising that EDI has failed to move beyond a powerful group of technology-heavy corporations. But can the Internet make EDI technology more palatable? Federmen certainly thinks so: There is a tremendous pull, he says. Internet-based EDI offers common standards, proven interoperability, anytime/anywhere network coverage, and an abundance of public domain software. There is also the small matter of cost. Internet service providers (ISPs) offer access to millions of potential trading partners for more or less the price of a local phone call, along with the cost of a flat-rate subscription. This contrasts dramatically with conventional EDI costs based on the use of established private EDI services (which store, forward and sometimes process messages), such as Advantis, GEIS or a Sterling Software VAN service. Suppliers levy a subscription, bill for connect time and transmissions, and charge for controlling, logging and archiving messages. Internet-EDI software suppliers hope that ease of use and the potential for dramatically reduced costs will push sluggish traders to move towards EDI. But there are some important hurdles to overcome. First, there is the problem of availability. Time and again, senior managers at large and small businesses have asked how the Internet can be viewed as a valid route for commercial messaging when no guarantee of availability exists. Internet users, they point out, rely for network resources on parties that have no commercial interests in providing them. Incidents like the five- day lock-out of 8,000 Internet users in Hong Kong, for example, have rattled business confidence. Second, there is a concern that the Internet will not always be low-cost. Today, the economics of the Internet has an irrational structure, in that it appears to be largely free to many users, and few people know who is paying for what.

But this is certain to change with charging by connection, volume of traffic, and by connection time, all becoming increasingly common. It is not unrealistic to expect the pricing structure of the Internet and of private EDI services to undergo significant change during the next three to four years. In particular, Internet flat fee access could give way to differentiated, usage-based pricing as government and academic subsidies dry up and ISPs (many of whom are in financial difficulties) start to look for ways to produce a profit. Meanwhile, EDI VAN prices will come down, driven by economies of scale, argues Howard Stern, director of EDI Services at Sterling Software. As on-line ordering increases incrementally, so too should the volumes of back-end EDI sales processing transactions. The upshot, he says, is that the apparently wide gulf between Internet and commercial VAN pricing will narrow or even disappear. But pricing, although always important, is not key. The cost of sending an EDI message is only a very small part, perhaps just 10% to 15%, of the total EDI expense. For the moment, the burning issue is not cost, but transaction security. EDI documents are usually financial, which means that security is absolutely paramount. Where messages are passing over a wide area, there are two main methods for ensuring security: first, secure the channels along which transactions pass; and second, shield the message itself in a secure digital envelope. Only one approach is fully proven and tested, says Stern – secure the channels by making the network inaccessible to non-members. For the past 18 months, established VAN players and experienced EDI users have been openly haranguing Premenos over its Templar software product, a relatively low-cost, flat-rate solution for Internet-based EDI. They argue that Templar is simply not secure enough. Federman, of Premenos refutes this claim. The fidelity of the EDI message is everything. We have had to spend a lot of time going over the details of our plans for public key and private key encryption, message authentication, confidentiality, non-repudiation … But the fact that Premenos has to spend so much time explaining and setting up its security features sets the alarm bells sounding, particularly among existing EDI users.

Concern

The concern is this: might not Internet-EDI evolve to a stage where one closed network of trading partners is merely substituted for another? After all, if messages are exchanged using proprietary security routines and application-level protocols, and if proprietary authentication procedures and proprietary encryption agents are used to scramble and unscramble data, then surely all this will add up to a closed network of users, rather than an open, Internet network. Blaine Erwin of Forrester Research thinks so. Use of proprietary implementations will stall the overall market. Internet-EDI solutions from competing vendors are in no way interoperable, he says. Federman contends that this is only temporary and that Templar will rely on self-certification technology only until an acceptable public- key (encryption method) infrastructure is established. Once this is in place, Premenos plans to offer its emerging WebEDI product as the next step for marrying EDI with the World Wide Web. This will enable anyone with an Internet connection and a Web browser to perform EDI transactions. It will minimize the fuss and lower the cost, says Federman. The link between the World Wide Web and EDI is crucial for mass market acceptance – but skeptics are quick to point to other limitations which may hinder EDI growth. One of these is a lack of interactivity: as an electronic commerce tool, EDI is limited by its store-and-forward email style technology, especially when compared with some of the emerging World Wide Web technologies. It isn’t suitable for the carriage of real-time, on-line business transactions. The problem is not new. For years, organizations such as IATA, the airlines ticketing body, have been calling for so-called ‘interacti

ve’ EDI standards, which would pave the way for on-line applications such as electronic ticketing. It has a strong business requirement: worldwide, the annual bill for selling and distributing airline tickets through agents and distributors comes to $23 billion – more than is spent on aviation fuel. Yet it is unlikely that a standard for interactive EDI will see the light of day for two or three years. For the moment, the development of Web ‘forms’ software and security-enhanced mailers looks more promising, especially because it is the focus of attention for two highly influential software companies, Netscape and Microsoft. Another, Open Market, looks to have the edge with its Integrated Commerce Environment, a software product set for on-line commerce built around existing Internet protocols and which will work with Web browsers. It’s here, it’s live, it’s running today. We are real- time, with authentication and settlement in-band [on-line], says Gail Grant of Open Market. The company says the system will support credit card transactions, allowing order taking and payment, and has hooks to EDI which will close the electronic trading loop, making for reliable, digital purchase order messaging.

Real progress

This is where the real progress is being made, says analysts such as Erwin of Forrester. In the context of the Web, Internet-EDI seems to be just scratching the surface. It is no more than a Band-Aid. I’d advise businesses to skip it. Web server technology is about to swamp the EDI domain, says Erwin. Analysts at Frost & Sullivan disagree. They believe that EDI sales are set to quadruple to $3.2 billion worldwide by 2001. In Europe, Ovum projects a compound growth rate of 21% a year up to 2000, to reach just under $1 billion. A shift in revenue streams is expected, with more revenues coming from services and less from EDI systems sales and support. In actual numbers, companies which trade electronically find themselves in a minority – although those that do use it are large and powerful. EDI use is almost mandatory among Fortune 500 companies while penetration among small and medium-size business sectors is negligible. All agree that the key issue in the widespread take-up of EDI is the overall level of business confidence in electronic commerce in general, and the Internet in particular (electronic commerce will drive EDI use, whether or not the Internet is used as the transport mechanism for back-office administration). There are signs of growing acceptance of electronic commerce. Business generated through the World Wide Web was running at $436 million in 1995, says ActivMedia, the specialist Web market analyst company that monitors all Web-based business activity, dramatically up on the previous year. But, for the moment, most Web sites are seen as ‘experimental’. Indeed, sales volumes are conservative. Across all the Web sites it monitored, mean revenues were just $7,100 a month. However, ActivMedia is projecting 100-fold growth by 1998. By the turn of the century, it contends, nearly any product or service that can be delivered electronically will be. The growth expected in on-line business is almost certain to give rise to an increase in the volume of electronic trade messaging. But it remains to be seen just how much of the increased traffic funnels into public access networks and how much the established VAN suppliers can siphon into their private service systems. In a potential market for electronic trade estimated by analysts at Forrester Research to be worth $21.9 billion by the year 2000, even a small influx would make for a massive injection.

Vans on the highway

The figures speak for themselves: global reach to over 106 countries, availability to well over 24 million users, around 15,000 commercial sites, monthly growth rates exceeding 10%. The spread of the Internet is simply too strong for the EDI message carriers to ignore. GEIS (GE Information Services), the $700 million juggernaut of the value added network (VAN) services world, has already been provoked

into announcing a newly formed Internet division with a business brief to develop Internet-based commerce services. It faces stiff competition. Premenos has positioned itself closely to Advantis and the IBM Global Network, while the maneuverings of Harbinger Net Services promotes it into the top three with 16,000 subscribers and 25,000 mailboxes. In addition, the international telecommunications carriers are racing to build up a substantial base of Internet subscriber connections. France Telecom, NTT, Deutsche Telekom and British Telecom are all busy plotting Internet tactics. And they do so from a position of strength, offering just the type of industrial-strength Internet access that corporates are likely to demand. These giants distinguish themselves from more nimble, niche Internet service providers on several fronts. They offer high bandwidth access through leased lines, using technological standards such as frame relay, X.25 and ISDN, and they offer faster, more reliable message transmit. It is estimated that some 80% of all Internet traffic is transported by the backbone networks of Sprint, MCI and Advanced Network and Services Inc. In April last year, this trio agreed to link their Internet efforts to direct traffic exchange and raise service performance. Companies chasing Internet business are in for a wild ride, warns Forrester Research. So what are the prospects for VANs on the information superhighway? In many ways, the standard protocols and wide reach of the Internet offers an entirely new distribution channel to attract small and medium size companies. GEIS wants to offer this to the 40,000 firms that are already hooked into the GEIS EDI network. It says that they can combine the security and accountability of the company’s worldwide private network and VAN applications with the freedom and accessibility of the Internet. The World Wide Web provides direct access to a global market. That’s undeniable. But customers are telling us that once electronic trading links have been established, they will want to conduct their business behind the firewalls and security blanket provided by a VAN, the company says.