MDM grows up, slowly

The master data management (MDM) industry has now been around for about a decade: it does not have an obvious date to put on its birth certificate but the term started to be bandied about around 2002, though there had been precursor products a few years earlier that specialised in the handling of particular data domains, specifically customer and product data.

The term became solidified when one of the industry majors, SAP, brought out a product with the MDM name in 2003, though after poor take-up they purchased a company called A2i in 2004 and relaunched it as SAP MDM. What is intriguing is seeing how the industry has developed and the degree to which mergers and acquisitions have played a part in shaping the industry.

With MDM, as with almost every enterprise software development, innovation came from small startups rather than lumbering industry giants. A company called DWL built a customer hub as far back as 1996, mainly for the financial services industry. Siperian was founded in 2000, also focused on customer data but aimed at the life sciences industry. Software dealing with mastering product data goes back further, with Hybris founded in 1997, Heiler as far back as 1987, and Stibo even earlier.

Yet these early efforts remained niche affairs until around 2002, when large enterprises began to realise that the ERP silver bullet was not going to solve all their data inconsistency issues at a stroke. In 1998 the first true multi-domain MDM software appeared in the form of Data Foundations, followed by Orchestra Networks in 2000, with Kalido MDM appearing initially in 2002. But even when I was running Kalido in 2001 we were considering bringing out Kalido MDM and I remember being told by two analyst firms that, as far as they were concerned, MDM did not exist as a market and had no potential to do so.

However in the software industry the power of the sales channel is paramount, and there is a chicken and egg situation. Customers will know that they have a problem (say, in managing master data) but if there is no obvious existing market for software then they will be told by analysts to build their own solutions. It needs either some startup to be wildly successful or for a major industry player to enter the fray. Other large vendors will react, usually by buying competing startups. Industry analysts will then proclaim the new shining market, and customers will start buying in numbers. Once SAP brought out its MDM offering, Oracle followed in 2004, and IBM purchased DWL in 2005. MDM was firmly on the map, and the term ‘MDM’ started to supersede the ‘CDI’ and ‘PIM’ acronyms that had described the single data-domain pioneers.

What I find interesting is how long the inevitable consolidation process has taken. Tibco bought Velosel in 2005, Microsoft purchased MDM startup Stratature in 2007, Software AG bought Data Foundations in 2010, with Informatica buying Siperian earlier in 2010. Just last week Informatica purchased Heiler, a PIM pioneer, so the process of large vendors buying MDM startups has been going on for eight years, and still there are a few independent MDM pioneers out there, and some plausible candidate large companies who have yet to enter the space. New entrants continue to appear: Pitney Bowes launched an MDM product in mid-2012, and the market itself continues to grow at a very healthy clip of around 15?20 per cent annually, depending on which analyst you speak to.

Why has this process taken so long? The customer need has been apparent for decades, the first early software solutions were appearing as far back as the late 1980s, yet it took until 2003 for a major vendor to enter the fray, and even by 2012 new entrants are still appearing and large companies still buying up early pioneers. Compare this with the speed in the internet software industry, where Facebook was only founded in 2004, Twitter in 2006, Google in 1998. I think this is due to the inherent conservatism of corporate software buyers, who have huge legacies of existing infrastructure to deal with. A consumer can flip from Facebook to Google+ in moments, but a company with hundreds of existing applications to consider is by nature an inflexible beast: making a change to its software landscape is like turning an oil tanker. The slow but inexorable rise of MDM is a case study in how long it takes for an enterprise software segment to truly emerge.