The master data management (MDM) market is at least a decade old now, depending on your definition. Certainly there were products around before 2004, such as DWL (acquired by IBM), but the term MDM came into common parlance in that year. At that time the technologies were firmly aimed at either customer data or product data, but not both. These data domains certainly have common management issues, but do need specific functionality.

For customers a common issue is name and address management, to avoid deliveries or mailshots going to the wrong address. Hence solutions focused on customer data tend to have elaborate functionality for name and address validation, often using data quality tools embedded with the solution to help avoid duplicates. If you are a B2C customer then volumes are high too, so you need technology that can handle 100 million records or more in real time.

Product data is usually much lower volume, with just a few million product lines typically, but what it lacks in volume it makes up for in complexity. Product data is often classified in complex ways by marketing folks, so sophisticated hierarchy management is a priority: being able to deal with unbalanced or variable depth hierarchies is important for such tools, but is rarely relevant for customer data. Also, product data frequently arrives in spreadsheets or text files with little or no structure, so the ability to parse a record and assign meaning to attributes is of key importance.

As MDM became more popular it became clear that companies struggled to manage all kinds of data beyond these two domains: asset data, materials, location, people, charts of accounts and more were all candidates for MDM. This led to the development of "multi-domain" technologies, pioneered by companies like Kalido and Orchestra Networks, that were neutral to the type of data being managed. After an initial phase of denial, most single-domain vendors converted to the multi-domain mantra, even if in some case more on PowerPoint that in actual code.

A flurry of acquisitions has seen data integration and infrastructure vendors offering broader platforms, combining not just data quality software with an MDM hub but also data integration software. Oracle, IBM and Informatica have all gone down this path, as have Software AG and Tibco. Some global companies have launched ambitious projects to try and manage all their master data in one architecture, though such projects have great political as well as technical challenges, and few have really succeeded in nailing down the management of their master data across all domains and geographies.

At this point you may think that the market is maturing, and that the vendor landscape would become reasonably settled, but in fact that is not the case. Even as the mega vendors digest their acquisitions, the last year or two has seen a flurry of new entrants.

One example is Semarchy, a French vendor with built-in lineage and version control, and a cloud deployment offering. Another is a product from Pitney Bowes, who are better known for postal franking machines but also have a large software division. Their MDM product is based on a graph database, making certain types of data relationships, such as customer relationships, particularly easy to model and display. Another MDM product from a data quality vendor is Ataccama, which specialises in high volume customer data projects.

Agility Multichannel, though multi-domain by design, is aimed mainly at rapid on-boarding of product data for retailers, with links to eCommerce. This product joins more established vendors such as Stibo, Heiler (bought by Informatcia) and hybris (bought by SAP). Another speciality area is dealing with material and spare parts data. This requires specialist domain knowledge, such as the ability o link to product coding standards such as the international coding standard UNSPSC. Verdantis and Sparesfinder play in this particular niche, the former a company launched in late 2012. There are even add-on products appearing, such as Veriscope from InfoTrellis, which aims to help manage MDM hubs via a reporting suite.

All this activity shows that there is plenty of room in the market, despite the presence of mature MDM offerings from industry giants. To some extent MDM can be seen to be going back to its roots, with products aimed at specialist data domains and competing on domain knowledge and specialised functionality against more general MDM offerings. For most companies, it will be tough to impose a giant MDM hub on its entire organisation, and ensure that all master data is mastered and maintained via such a single master hub to rule them all. Internal politics makes such things exceedingly challenging. Hence most MDM projects today are of smaller scope, picking off specific data domains in certain geographies or business units, or even starting with less volatile master data, known as reference data. What is clear is that MDM is now becoming firmly mainstream, is expanding rapidly as a market in 2014, and growing up as an industry.