Master data management as a discipline had its roots in improving an enterprise's handling of customer data, with some early vendors producing customer hub solutions as far back as the late 1990s. The same approach has long since been extended to product data, and beyond to location, supplier and other data domains, but customer data remains the single most popular data domain to demand attention. A 2012 Information Difference survey found just 48% of respondents reckoning that their customer data was "consistent". This was not especially surprising since further surveys by the same company in 2008 and 2013 found that a large company had a median of six different competing systems actively updating customer data – hardly a recipe for clarity.

There are different approaches to tackling this issue. One is a root and branch style where a new master data hub is installed and this system gradually becomes the single master source, actively switching off the ability of other systems to generate new master data. This fairly invasive step is relatively rare in practice due to the amount of change involved. It is more common for a master data hub to co-exist with existing transactional systems, taking feeds of data and producing new "golden" copy data for downstream systems, such as data warehouses. However it is also possible to leave the various sources of master data in place, and construct an "on the fly" version of master data that removes duplicates and addresses data quality, feeding the shiny new master data to systems that need it, but without trying to fix the data at source.

Such an approach was taken at Nationwide Building Society, the UK's largest with 870 retail outlets and 16,000 employees. Previously the organisation had a customer system with limited information such as basic contact details and their accounts, and a somewhat silo-based approach to dealing with customers through the various different customer touch points. They wanted to provide their front-line employees with a comprehensive contact history in which customer interaction across branches, call centres and their corporate website could all be brought together. The approach that they took left the source systems in place, but added a layer of software that could access these and generate a real-time consolidated view of customer data when needed. This was complemented by a data quality project that improved the state of the customer information at source. This dual approach enabled customer-facing employees to get a more complete picture of clients, and to be able to pro-actively suggest products that made sense for them.

Following implementation, Nationwide found a significant improvement in sales effectiveness, with a doubling of incremental sales. The ability to deliver a more personalised service improved customer satisfaction too, and indeed Nationwide has pulled ahead of its banking peer group in terms of customer satisfaction, measured monthly by the independent market research firm GfK NOP. Its lead in this area has been widening in the most recent polls, and it has been rated as having the highest customer satisfaction of any bank in Europe. The number of products held by each customer has on average increased by 50%, and indeed more sales are achieved through their inbound service channels than through outbound marketing channels. The project paid for itself within two years, with an attractive return on the project investment.

This project illustrates how improving master data can have substantial customer benefit, even when it does not involve putting in place a conventional permanent master data hub. Naturally, this project was not just about data improvement, but also involved staff training that enables sales staff to make effective use of the much greater customer information that the new system made available.

There is no single approach to master data that is correct in all cases. In some instances it will be better to bite the bullet and implement a brand new hub, actively switching out older systems, which may bring additional cost savings although it is tougher to implement. In other cases a "co-existence" route may achieve many of the benefits needed. The right approach will vary from organisation to organisation, depending on the political culture of each and the degree of determination to address the underlying problems. A more invasive approach may have greater eventual benefits, but cost more and involve a higher degree of risk of failure.

A common theme, whatever style of approach is taken, is the need to align the project with a genuine business need, such as the customer satisfaction initiative seen at Nationwide. A further recurring theme in such projects is the need to address data quality, which almost inevitably turns out to be in a worse state than senior management realise. What is impressive is the tangible business results that can be achieved when such a project goes well.