By William Fellows
A handful enterprise application integration CEOs suggest that TSI International Software Ltd CEO Connie Galley should be patting herself on the back in light of New Era of Networks Inc’s ‘explosion.’ Neon’s misfortune appears to have debunked the myth of EAI through acquisition. TSI’s approach is organic growth with strategic acquisitions where required. Indeed TSI is using the opportunity to begin doing some of what it hasn’t been much good at thus far – blowing its own trumpet.
Galley told ComputerWire that TSI’s business model has been clearly vindicated and that the long overdue market correction puts the two company’s market capitalization in the same ballpark, where they should be. Until Tuesday Neon’s value had been at least twice that of TSI. Galley explains that Neon’s in a difficult position because it needs to maintain a fast growth path while scooping up slow-growing ISVs, meaning it needs to add them at a faster and faster rate. By contrast Galley says that TSI prefers to under promise and over deliver. TSI is expected to be at a $90m run rate this year up from $45m revenue last time.
TSI has recently been telling its own story to the analyst community rather than allowing its competitors to do the talking for it. It insists it’s not just a data transformation specialist but offers products up and down the classic EAI ‘stack’ including application adapters and business process flow.
Dun & Bradstreet spin-out TSI was founded in 1995 for a mainframe data extraction toolset which it still markets. It developed EDI solutions for PCs – integrating workflows and transactions with front office applications, and nearly collapsed when that market never materialized. CTO Saydean Zeldin – who wrote the re-entry guidance development of the Apollo flight software at the instrumentation Laboratory of MIT – wrote TSI’s core Mercator data transformation engine which was used initially to support the company’s EDI efforts. It reached an agreement with SAP on EDI in 1993 and then released Mercator for SAP R/3 in 1996. It claims 250 R/3 customers, around half acquired through the channel and OEMs.
SAP sales account for a third of total sale revenue; half of Mercator sales. Its average sale price is $175,000 but sell-on business is usually ten or twenty times that, the company claims. It has formal reseller arrangements with the five big systems integrators. Much of its new business is integration over IP. Remember, it says, IBM Corp’s eponymous MQSeries transport can’t go through the firewall. Its acquisition of UK financial EAI specialist Braid gives it an entry into the STP straight through processing world – where Braid is the number two supplier behind Tibco – plus the Gemini generic mapping tool for equity market applications and Nimbus which manages collateral tracking in wholesale banking.
Mercator 2.0 is at beta 6 and will ship this quarter. Key improvements include multi-byte support enabling TSI to better attract global customers with Asian interests. Performance has also been improved. It is already multi-threaded and doesn’t use a database which it says cripples the performance of other EAI solutions. TSI admits that technology has prevailed over marketing in previous releases which were numbered 1.1 through 1.4 accounting for the number of years it has been in the market. Expect the following release to be 3.0.
TSI believes the market could develop in such a way that systems integrators become the main vendors of applications; ISVs will be their ‘component’ program suppliers. Some large ISVs may even achieve this status too, in which case they will probably buy or build their own EAI solutions. It is not viable for SAP and other vendors to offer their own EAI solutions today as customers naturally want solutions that can tie together multiple applications without vendor bias.
