If Ecclesiastes is right, and there really isn’t anything new under the Sun, then a fine example of that phenomenon can perhaps be seen with the April re-emergence of Marlow, UK based data exchange company Software One Ltd as newfangled metadata maven One Meaning, Inc. For this tale is quintessentially one of how good ideas in the software industry never really ever go away – though they may never actually make their inventors any money at any one given time. With Software One, a company founded in 1989, that idea was metadata, that is, data about data; and while the vehicle for commercially realizing the benefits of that notion was Computer Aided Systems Engineering (CASE) first time around in the early part of the decade, now the company really feels its hour of glory has come with the modern trend to Warehouse that Data. When Andrew Mercer, previously of MSA and Comshare, set up Software One with his wife, the market gap he spotted was between data definitions and repository formats of the major CASE tools of the day, and soon his company was offering Exchange, now capable of moving metadata from some 30 different environments into a common data model for translation/export to others (think of it as a hub with 30 spokes).
By Gary Flood
A rake of CASE tool vendors soon saw that linking up with an independent offering such as this was an ideal way of selling their own products while allaying customer insecurity about lock-in. John Spiers, now VP of Strategic Planning for One Meaning, was previously at Oracle’s UK subsidiary, was on hand to ink the deal between Oracle and Software One five years back. He now notes that the problem with Exchange was that while a sensible product, it was never a particularly sexy one: It was a bit like insurance – it did Oracle a lot of good having it in the portfolio but that didn’t translate into a lot of people either really buying it or using it. Though Exchange has, over eight years, built up a loyal customer base of some 400, the fact that the 45-staff strong company claims its sales last year at a very ho-hum $2.5m. Though CASE is no longer the Church of Whats Happening Now, it is worth noting that not only are there still many development shops out there with CASE software which is still in use, and for which Software One/One Meaning still sells product, the basic needs it tried to address have not gone away. To listen to Spiers, they may even have gotten worse. The questions it asked haven’t been answered and corporations have the same problems. The old idea was to use CASE to build what we then called an enterprise data model which would hold a picture of the company’s operational data structure in order to get a more holistic view of the company and its customers. No-one ever actually managed to build an enterprise model in the 1980s and 1990s, so people gave up, but thats because it was just too hard, not because they didnt need one. And now the new idea is to use cleansed data gathered from everyday operational systems to build what we now call a Data Warehouse – in order to get a more holistic view of the company and its customers. Marlow itself grew out of an aborted early attempt by Sybase Inc to get into Warehousing before its financial woes started two years ago, but after Sybase withdrew funding for further development Mercer saw the potential of applying metadata rigor to the emerging Warehouse trend. Common to both projects is the need to rationalize what may be 20 or more operational/legacy system definitions of, say, ‘customer,’ into one central repository.
Big Man on Campus
And this is where metadata comes in, as it is a way of obtaining information to help you understand the data you have, in Spiers’ words, you need to understand the data structures in your operational systems in order to build the Warehouse. Hence One Meaning’s new product, Marlow, based on an object oriented DBMS from Versant Object Technology Corp, which is a combination of a ‘metabase,’ Exchange-style bridges and a set of tools, the combination of which is claimed to allow firms to move metadata in and out of a neutral data garage which can also track versions of different definitions of data, which One Meaning sees a need whose time has definitely come. But One Meaning claims to be in the position of offering an unmatchable eight years’ experience in the field, a lead which it thinks could translate into it being the Big Man on Campus of the market… when it takes off, which is not yet at least. There are three big potential sectors for us – the infrastructure/information management (repository) market, the enterprise information Data Warehousing market and where companies are involved in the pain of implementing a big applications package like SAP. All of these may come together in a large merger or acquisition, for example. But at the moment the market doesn’t know it’s a market – our major competition is a corporation spending $5m of its own money making do or trying to develop its own solution. The company sold one copy of Marlow in December 1996 for $300,000 to the London based element of European financial services giant ING Barings which is using it for the more consistent handling of customer information across its customer information systems. So few real users yet. But after a year of effort – and a lot of air miles – Mercer, still president and CEO of the revamped company, has been able to score the coup of getting an impressive $6.8m worth of VC funding from Menlo Ventures and Matrix Partners and as of Monday the company is up and running from its new Palo Alto HQ.
Gain credibility
Well, nearly; while the shingle is up the company’s actual offices are being refurbished, it still lacks a VP Americas and VP Pacific Rim, and it knows it has to win some large US accounts in order to really gain credibility in this marketplace. With that kind of serious investment in the company from such well-known Vcs, with us being here in the Valley so we can manage our partnerships better, were about to finally put ourselves on the map in the US, declares Spiers. There is also the FUD factor of the Microsoft repository and its recent tossing of a fragmentation grenade into the whole metadata world (CI No 3,269), so estimates of revenue of $6m, by next March do seem a little ambitious, but given the management talent at board and operational level, plus the funding, Mercer may have finally assembled the better metadata mousetrap. Though one does wonder, taking the longer term view, whether or not we will always try to rationalize consistent data structures out of our frantic everyday operations, call them Data Warehouses or Enterprise Models – but as the Good Book also says, there is a season for everything, probably even for assembling metadata.
Content from our partners
