Skip to main content

What is Master Data Management and why does it matter?

people in a meeting looking at data
(Image credit: Getty)

Many businesses were drowning in data (opens in new tab) even before the pandemic moved everything online. It’s been a recurring topic of interest for the entire C-suite how to maximise the value of owned data - and extract the insight and value hidden within. 

Yet even for those who are investing efficiently in the data teams, systems and processes to unlock and make use of often complex datasets, there remains an issue of its ongoing hygiene. For whilst data can exist in a vacuum, the need for properly instituted data management has only grown in recent months - as both the volume of data being managed and the competitive business environment (amidst testing economic conditions for many) have also surged.

Of course, data management is a hugely broad area, but one area that’s receiving greater focus is Master Data and the data management subset of Master Data Management (MDM). So what is it - and why is it important for a business to get on top of?

Related: The power of data as a business input (opens in new tab).

What is Master Data Management (MDM)?

Essentially, master data is the data about the business entities which provides context for the transactions your business conducts. It describes the operational business data that typically makes up the most valuable data a business has. It typically sits in three main categories: people/legal entities (customers, vendors, suppliers etc.), things (products, assets, accounts) and locations.

Each category has its own unique attributes and characteristics. For example, for a customer’s master data, we might include: first name, last name, address, postcode, buyer type, and any contracts or SLAs they may have signed.

Factoring in each sub-category, this can result in a large amount of business-critical data. It’s not rocket science to know that keeping accurate, readily available information on customers, products and location is essential for everyone across a business, from engineers to finance, marketing and sales etc.

Where and why MDM is needed

woman viewing a delivery app on her phone

Customer records may have addresses that don’t match, or only include part of the name – that's where MDM can help (Image credit: Getty)

If Master Data’s importance is clear, it’s efficacy within many businesses isn’t. Many databases we come across, even within a single system/application, can have confusing, duplicated master data.

A customer record may have addresses that don’t match, or only include part of the name. A phone number might be stored in two different formats, with and without international dialling code for example. It becomes possible that systems this master data is being fed into can’t make use of it, if - as is likely - they don’t recognise certain formats of stored information.

One of the main issues affecting MDM is where siloed, unsynchronised systems lead to a particular customer being created, then recreated and possibly modified, leading to multiple entries for the same customer. Any missing data, inconsistent formatting, or incorrect spellings can lead to these siloed systems (CRM vs. Billing for example) creating unique customer IDs for the same customer.

The knock-on effect of this is an inability to be certain if it’s one or many customers, and it completely undermines the data analysis that can lead to a business unlocking valuable insight on that customer.

It’s a classic example of a common data management problem that could even lead to a business losing customers owing to:

  • Poor customer service because of the wrong contact/product information
  • Poor sales reporting - given the duplicated customer record
  • Inefficient marketing spend - if a customer is targeted multiple times
  • No ability to upsell - if buying history is siloed from customer records

So the need for Master Data Management is clear. The fact that it’s stored in different systems, where mistakes and duplicates can easily occur, leads to unreliable and flawed information going to multiple departments. With success increasingly depending on having an accurate picture of a businesses’ state of play, MDM is a vital cog in the data management wheel.

Related: What does it take to be genuinely data-driven? (opens in new tab)

What’s the solution?

data entry on a laptop

MDM involves laying a data model to standardise data input and importing across different enterprise systems (Image credit: Getty)

Whilst it’s possible to automate solutions in individual systems, even reconciling the previously mentioned formatting issues, it’s the siloed approach to data quality that prevents easy consolidation and alignment of any duplicated or misleading data.

So first things first, MDM involves laying a data model to standardise data input and importing across different enterprise systems. This enables precise mapping and interchanging of data between the sources and the MDM solution, governed by structural rules and formats. Then an MDM solution sets about standardising and cleansing the data, reconciling inconsistencies already present in the database to begin a process of matching.

The matching stage, built upon customisable rules, enables a business to determine whether records from systems are the same piece of data or separate. To take our earlier customer example, it might say “if surname and phone number are the same, this the same customer”. These rules can be sequenced and run in turn until a rule is satisfied.

Merging - not always a required stage - can then finalise the creation of a master record; a consolidated view of a single customer or entity across all the (previously) siloed systems. Similarly rule-based, merging can help determine which records should be prioritised for pulling certain information, i.e. “always take the latest phone number from CRM”.

One size does not fit all

Of course, not all businesses are created equal. So whilst MDM goals rarely shift from being a single representation of a cluster of data, there are naturally different ways to implement MDM, depending on the systems in use, data domains and business processes in play. These fall into four broad styles:

  • Consolidation: pulling source data into an MDM hub, providing the best version to consumers - but source data stays as it was
  • Co-existence: as with consolidation, it pulls source data into an MDM hub, providing the best version to consumers, but it also synchronises this best version back into original source systems
  • Centralised: master data is created in the hub itself, all other systems are connected via APIs and do not need to author data locally
  • Mixed/Hybrid: combines consolidation or co-existence with centralised style - some data is consolidated while other records may be centrally authored within the MDM

What else can MDM do?

Beyond merely fixing user record duplicates, MDM’s broader mission is to centrally merge and govern data from any system in an enterprise, to unlock unthought-of, powerful benefits to a business.

Putting MDM in place, with its structures, rules and systems of data governance, ensures accurate financial, sales, or regulatory reporting. Accurate inventory information empowers sales and procurement teams. And a clean customer database means the CMO can precisely segment audiences for more efficient marketing campaigns. Depending on what businesses offer, the confidence in easily accessed data can open up areas of growth that have not previously been considered. It can create a ‘single truth’ from which all businesses can run.

Roman Kucera is CTO at Ataccama (opens in new tab).

Related: Turning data into opportunity post-pandemic

Roman has over 11 years of experience in software development and data management. As CTO of Ataccama, Roman's main focus is developing innovation and emerging technologies - for the past couple of years, this has involved artificial intelligence and big data platforms.