Enterprises do a poor job of mastering their data — and that failure makes it difficult to personalize marketing and create a single view of the customer, Forrester analysts conclude in a report on master data management.

The report (fee charged) was co-authored by Forrester analysts Michele Goetz, Gene Leganza, Shaun McGovern and Matthew Izzi.

IBM, one of the leaders in the report, defines Master Data Management (MDM) as systems that provide a "trusted view of critical entities typically stored and potentially duplicated in siloed applications — customers, suppliers, partners, products, materials, accounts, etc."

Michele Goetz headshot

Goetz, a principal analyst serving enterprise architecture professionals, told CMSWire enterprises often fail to anticipate change and revisions of master data over its data lifecycle. They also strip away facts that provide context about customers and products — and vendors cash in on those errors.

The big data technology and services market according to the International Data Corporation will grow at a compound annual growth rate (CAGR) of 23.1 percent between 2014 and 2019, with annual spending reaching $48.6 billion in 2019

Brendan Witcher, Goetz’s colleague at Forrester, told us earlier this month the average organization has customer data in ten to 15 systems and sometimes in as many as 20 or 30

“Businesses are realizing that you aren't going to be able to have a single system for all data,” Goetz said. 

“Federated ecosystems are the new normal. MDM becomes more critical as a way to crosswalk across data silos, inform how to aggregate disparate data in those systems, and orchestrate master data changes across systems. Adoption and replacement of MDM systems is on the rise to manage and orchestrate master data schemas across traditional, big data and cloud.”

Common Roadblocks

But education is in order beyond the technology, Goetz said. Too often, enterprises assume there is only a single master data definition when in fact each search/query, application and decision has its own schema. 

“Systems are built to run and walk away,” Goetz added.

Where else do businesses stumble?

“Constraining master data models and schemas to a tight number of data elements and attributes and stripping away facts that provide more context about a customer, product, site, etc.,” Goetz said. “Enrichment of master data supports a wider range of business scenarios due to master data relevancy in decisions, processes and actions.”

Enterprises also get in trouble when they treat MDM as an integration hub rather than a model and schema management solution.

It’s an “expensive investment when a lighter-weight reference data management solution or metadata management solution could better serve the task,” Goetz said.
“MDM's value is managing different definitions and views of master data. It is not only for data mapping.”

Vendor Landscape

This brings us to the vendors in Forrester’s Wave. According to Forrester’s breakdown:

“At a minimum,” Goetz told CMSWire, “they are all able to scale across complex ecosystems, managing, integrating, orchestrating and syndicating master data and models/schemas.”

It’s the enterprises themselves, Goetz said, that create the roadblocks. Not the vendors.

Forrester's discussions with vendors and customers shows a lack of data governance and stewardship expertise along with a lack of preparation on what the MDM tool needs to manage master data.

“No definitions, models, schemas, data policies and business rules/standards are defined prior to purchase,” Goetz said. “This can add three to six months to a project. Implementation can be straight forward with a system integrator/consultant.”

Almost all implementations are done with a partner, she added. Organizations that put effort into the upfront data governance and subject-matter experts will have a better chance of successful MDM implementation. “Most organizations are immature in this area and lack even an understanding of who the data owners and stewards should be,” Goetz said. 

New Kids in Town

The newcomers to the leaderboard handle context at scale, Goetz said. 

“They are managing the physical, logical and semantic models in unison," she added. "Reltio and Pitney Bowes have graph databases as their repository that allow master data models to include more metadata for rich context. The graphs also provides additional flexibility in managing multiple domains in unison and natively rather than setting up separate domain models.”

Their created environments thrive, she said, on high fidelity and deep data linkage allowing organizations to both set MDM definitions and continue to expand and evolve these definitions as more context is infused. 

“Additionally, machine learning is becoming part of the MDM solution to interpret and infer the domain models from the data sources,” Goetz said. “This gets MDM up and running more quickly and helps evolve the domain models continuously as data changes or new sources are added.”

Goetz said she was surprised in her MDM research by the number of vendors that have introduced knowledge graphs to visualize master data models and linkage.

“I was delighted,” she added, “to see that some of those vendors also took the next steps to allow changes of the master data and the models through the knowledge graph.”

Manish Sood, CEO at Redwood Shores, Calif.-based Reltio, said in a statement that“Forrester’s research "echoes the demands of enterprises today, emphasizing that reliable data is a core foundation to bring any business vision to reality."

“Our highest score in the strategy category, and the second-highest score in the current offering category, reflect our customer’s confidence and commitment to our platform. Our core mission goes beyond MDM. We are focused on simplifying all aspects of data management for IT, while delivering high value data-driven applications into the hands of frontline business users.”