When young companies go in search of funding, the quality of the money they obtain matters just as much as the quantity. Being bankrolled by a group of prestige investors can make a huge difference when trying to convince customers and future shareholders that the start-up’s value proposition is sound and that its long-term prospects are good. In May, when application integration software company Constellar concluded a round of private funding, there were some big-name takers: James D Robinson III, the ex-CEO of American Express; Charles Wang, the CEO of Computer Associates; and international financier George Soros. They join a list of other backers that includes the CEO of PeopleSoft, the founder of Cisco, and venture capital powerhouses Brentwood Associates and Technology Crossover Ventures. Constellar was able to raise $18m in private money because such sources of funds have concluded that the company is sitting on a technology that addresses a fundamental requirement. Large corporations have found it necessary to build data connections between the complex mix of custom and third-party applications and databases that they have amassed over the years. The matrix of application-to-application communications and database-to- database links have become even more complex as the IT infrastructures of companies have become more widely dispersed globally, and as mergers and acquisitions have created a need for previously separate systems to exchange data. A mainframe payroll system, for instance, may need to draw data from an in-house personnel software program while also using data from a finance department application from package vendor. To cope with this demand, companies are building ‘point routines’ to transfer data between manufacturing, financials, forecasting and many other business applications, creating what the Gartner Group calls ‘inter-application spaghetti’.
Staging and transformation
Constellar’s way of addressing this problem is to offer a hub that acts as a staging and transformation post for data from one application to others. It irons out inconsistencies in data formats and application interfaces, and even offers the opportunity for users to add business logic to the process while it is passing through the hub. Says Constellar CEO Brian Donnelly, it provides ‘an information router’ the data equivalent of network products offered by Cisco. However, the company has not always positioned the technology with such a grand design. Springing from the UK, but now headquartered in Redwood Shores, California, Constellar saw its technology as being most applicable to the booming data warehouse field, where the hub could be used to bring consistency to data from various sources before it was passed for analysis. But although 40% of its deals still come out of the warehousing sector, Donnelly has always appreciated that the technology has a much wider scope. It could, he says, become the key layer in a corporation’s IT infrastructure. Certainly the company’s client list is equally as prestigious as its early investors: Telecoms company Sprint is using nine Constellar hubs and has plans for 40; ScottishPower has five hubs; Deutsche Bank has nine; the British Airports Authority has six. This is not experimentation, says Donnelly. A third to a half of our customers are in production with their hubs, he says, citing First USA, Hoffman-LaRoche, Imation, British Telecom, Imation, Nikko Bank, Yorkshire Electricity as production sites. The hubs do not come cheap. The average initial sales price is $250,000, and that has attracted considerable interest from consulting firms such as Ernst & Young, Coopers & Lybrand, IBM and Oracle, which are interested in related implementation fees. From $3m in the year ending September 30, 1996, Constellar’s revenues rose to $8m in 1997 and should hit between $18 and $19m this year. Still in investment mode, the company does not expect profits to start flowing until next fiscal year. Constellar had expected to go for an initial public offering this year, but one of the reasons the company chose to put that off until 1999 or 2000, and seek a further round of private funding, was that it was little known outside of database circles. A large chunk of the $25m will go on overhauling The worst marketing in the world, says Donnelly, we have done zip in the US. Based on its private placement, the company now carries a valuation of around $125m. Now it has to prove that its trophy investors are onto something big. Computer Business Review.