Quantexa

Desilo, Detangle, Deduplicate: Master Data Management

A comprehensive guide to overcoming your company’s trickiest data record challenges—through contextual Master Data Management.

Dan Onions
Dan OnionsGlobal Head of Data Management, Quantexa
Zuletzt aktualisiert: Oct 23rd, 2024
15 min read

The challenge you face

After decades of data collection, your company’s volume of data will have grown exponentially. But most of this data is collected by multiple applications, which often results in duplicate entries in different (siloed) repositories.

That makes it difficult to get an accurate, timely view of the data your organization relies on to operate—whether for customer service, information management, or decision analytics.

And the result?

Internal problems

  • Your records are incomplete and may include outdated, inaccurate information

  • Your data teams are overloaded with mounting manual remediation tasks

  • Revenue opportunities are missed

  • Potential risks are overlooked

Customer problems

  • Services and products are offered to customers who already have them

  • Customers are frustrated by repeating details and information to different customer service agents operating in different data silos

  • Customer experience takes a nosedive due to inaccurate, missing, or misrepresented information

Regulatory problems

  • An increasing number of regulations require reliable, consistent, and well-managed data

  • Regulators receive information submissions that are inconsistent and do not tie together

  • Inspections demand ever-higher standards for data quality and lineage

  • Enforcement actions involve fines and restrictions on business activities, affecting P&L and reputation

The fact is this: most organizations have the data but not the technology needed to manage master data and harness its value.

Without a way to deduplicate records and accurately chart data, individual entities—like customers, vendors, and related parties—can’t be properly understood and analyzed.

To truly make the most of your data, you need to be able to:

Block img

Consolidate and deduplicate it

Block img

Resolve discrepancies

Block img

Have full visibility - in context of related data

Master Data Management allows companies to tackle this issue. But poorly implemented or limited solutions are often ineffective at best and harmful at worst.

Read on to learn more about Master Data Management, its potential shortfalls, how you can overcome them with best practices – and the pay-off when you get it right.

CHAPTER 1

The inherent problems of Master Data Management

The aim to have one master record for each real-world entity has always been hard to achieve—and it’s becoming even harder.

Companies are contending with:

  • Multiple internal applications—many of which will contain different versions of the same master data record

  • Numerous external data sources provide additional—sometimes contradicting—information about companies and individuals

Bringing together all these views of master data is incredibly difficult, because of:

  • The data quality problem

  • The transformation challenge

  • The varying needs of different data consumers

  • The need for governance and control

The data quality problem

Traditional MDM has an inherent data-quality problem. And it affects your decision-making, regulatory compliance, business effectiveness, and efficiency.

Traditional MDM solutions don’t focus on solving data-quality issues. In fact, they tend to fail at the first hurdle of matching data because they struggle to join data from across disparate data sources such as multiple internal applications.

When you add the volume and variety of data from external sources it’s even further beyond their capability. Instead, they use roughly the same matching algorithms they’ve used for the past 20+ years, which rely on record-to-record comparison that is very fragile when key attributes are missing or different.

And that’s a problem. Because crucial information goes unreported when your MDM solution can’t catch essential links between data, and obscure relationships and connections are often overlooked.

It also makes your MDM implementation very high risk.

Bad data quality means:

  • Data remains trapped in silos and is duplicated across channels

  • Master records aren’t accurate and true-to-life

  • Bad decisions are made, due to the fact they’re based off incorrect or delayed data

  • Key information and critical links go unnoticed

  • Opportunities are missed

The transformation challenge

MDM is both a business and a technology transformation challenge.

Here’s how it usually goes.

You start out with multiple users updating the master records in separate applications, all with their own ways of working. Inevitably, it’s messy, it’s haphazard and it results in duplicate records, inconsistencies, and confusion.

So, to combat this, you decide to standardize things. Everyone is to use just one application, with a standardized set of rules on how to input data, how data should be formatted, what records look like, and more.

It’s a good idea—in theory. But the problem is that the real world rarely plays out quite so neatly.

So what you end up with is this:

  • A logistical nightmare, as you try to migrate decades of data and make it conform to your ideal record format.

  • Confused and frustrated users who need to be transitioned to the new service—with all the training and business change support that entails.

  • Backwards-compatibility issues, as information moves to new places and takes on different—unrecognizable—formats, meaning users and business applications can no longer find information.

  • Challenges updating future records. If your records only track A, B, and C, what happens when users later need to add D and E? Is it updated across the entire data store? And what happens to data that is initially discarded for non-conformity but is later needed?

It’s a big job. If you don’t have a plan for data migration, then your initiative is doomed. And even the best, most precisely executed migration plan can’t save you if the tools you rely on aren’t up to the task.

The varying needs of different data consumers

For an MDM initiative to be considered successful it needs to be able to serve data to consumers across the organization. For instance:

  • Analytics teams who need to link data sources for Decision Intelligence

  • Fraud-monitoring applications

  • Customer services or relationship managers who need rich views of their customers

  • Finance teams who need to aggregate risk reporting

Different business units often have different views on the data they require. So, when an MDM initiative attempts to standardize master data attributes across the organization, it may mean dropping attributes that these data consumers rely on—which, naturally, can result in tension and impact business performance.

MDM needs to be able to present a rich and deep view of data to areas of the organization that consume it, while on the journey to standardizing key attributes. But that often runs counter to the way a lot of existing MDM software works, which relies on a fixed view of data that needs to be adhered to from day one.

The need for governance and control

In large organizations there are often many applications that hold customer data—each controlled by different business units. And that means a wide range of stakeholders with different priorities.

Implementing traditional MDM often requires each business unit to give up control of their applications and data to a central initiative—which can result in a great deal of pushback and agitation.

To successfully implement an MDM initiative, you’ll need to be ready to address the political challenges that come with it. A lot of that relies on bringing people together around a vision of a service that will benefit them, within a transformation program that can actually deliver.

CHAPTER 2

Building a single, golden point of truth

The key to resolving the traditional MDM data-quality issue lies in powerful Entity Resolution that retains context and doesn’t force data into a standardized format. We call this "contextual MDM".

How contextual Master Data Management stands apart

Originally built to tackle financial crime, contextual MDM (cMDM) ingests data from both internal and external sources to build an accurate, connected, and enriched single-entity view using Entity Resolution and network generation technology.

This is different from traditional MDM solutions, which rely on record-to-record matching—a method that does not work well on disparate records, as it relies on many attributes matching.

At Quantexa, we use an expanded range of data—including address, phone, email, country, and third-party data—to make further connections and enrich your data. This is why our solution can make connections between records even when data quality is poor.

When people talk about MDM, they could be referring to the discipline or the technology solutions enabling it

Traditional MDM

Traditional matching does not work well on sparsely-populated records—because it relies on many attributes matching.

Records can only be accurately linked if a number of fields match (for example, if two records have the same name, date of birth, and address on file). The lack of additional contextual data makes deduplication difficult and leaves questions unanswered. Traditional Master Data Management also relies on you to set rules. If the rules are too rigid, records will be under-linked (meaning duplication is more likely to go uncaught).

If the rules are too loose, records will be over-linked (meaning different records are more likely to be mistakenly deduplicated, even when the entities in question are different).

Quantexa's MDM solution

With our Entity Resolution software, connections can be made intelligently across records.

Using additional fields and an expanded range of records from any number of internal and/or external sources, our software makes it possible to accurately determine when multiple records exist of a single entity—and to turn duplicated records into a single, enriched entity view.

Resolved entities can also be seen in context with their networks—so you can see how different entities relate to each other.


With MDM, data across different records is iteratively updated to enrich all sources, leading to better match rates and higher quality records data.

CHAPTER 3

The benefits of contextual Master Data Management

With contextual Master Data Management, you gain:

Block img

A single, complete view of connected data
One trusted source of truth across all systems and resources

Block img

Flexible and open architecture
With open APIs and original-format records, joined-up records can be easily consumed and integrated

Block img

The power to make better decisions
Information is accurate and updated in real time, offering more visibility and greater understanding

Block img

A foundation for trusted data
Enrich and improve the quality of your master data records

Block img

Low risk implementation
Without the need to transform data, it’s quick and easy to deploy

Block img

Consumer-oriented views of master data
Different applications consuming data are able to connect data according to their needs

So you can

Block img

Create and update accurate master records, in real time

Block img

Spot hidden risks and identify high-value growth opportunities

Block img

Share essential data among your teams

Block img

Offer frictionless, digital-first experiences for your customers

Block img

Develop good data practices and upkeep organizational data hygiene

Block img

Scale your business easily