Fixing master data without spending tens of millions

Master data issues are some of the most persistent and costly problems in large organisations. I have seen tens of millions quoted to fix them. Sometimes that is justified. Often it is not.

Over several master data management implementations I have settled on a delivery framework that keeps MDM tied to business need, avoids unnecessary platform spend, and leaves the organisation owning its data afterwards.

1. Make the problem visible

Ask about MDM and people describe architectures and golden records. Those are rarely the hard part. The hard part is organisational: MDM needs a long-term commitment of people and process, and the business has to recognise that before it will resource it.

Make it real for leadership. In one programme we showed the executive team more than four hundred variations of a single flagship product in their systems. Moments like that turn master data from a technical concern into a visible business risk.

2. Pick your design

Three patterns cover most needs, and they run from simple to complex:

• Consolidating master data to support reporting.
• Creating a governed master data source shared across operational systems.
• Consolidating master data across multiple organisations.

The pattern you choose drives cost, resourcing and timeline more than any tool decision.

3. Right-size the solution

Many organisations assume MDM needs a platform investment in the tens of millions. Often it does not. A large share of use cases are about reporting, and a consolidation solution in the data and analytics layer is enough. Synchronising operational systems does not always need a standalone MDM product either: an ERP can act as the master for some domains, with data quality tooling around it.

There are cases where an enterprise MDM platform is the right answer, such as a retailer with hundreds of thousands of SKUs, thousands of suppliers and omnichannel sales. The point is to make the investment proportionate to the problem.

4. Hand over to business as usual

The purpose of an MDM programme is not to cleanse all the data. It is to leave three lasting capabilities: a standard for monitoring master data quality, a process for correcting it in line with policy, and a single trusted source referenced by every system. Data stewards in the business then own it.

You do not always need the biggest toolset. Sometimes the smartest move is making better use of what you already have.

Scott Armstrong is a programme director who recovers and delivers ERP, finance, data and CRM transformation for large organisations. Talk to Scott about your programme.