A briefing by Pitney Bowes Software for the Hub Designs MDM Think Tank
It’s astonishing to realize that Pitney Bowes was founded in 1920. Most of us probably consider this company a postage metering, mail handling and printing giant. A good deal of the company’s business still focuses on solutions for postal and print services but Pitney Bowes continues to build out digital solutions as a complementary line of business.
Looking at its multi-decade history, it’s apparent that Pitney Bowes is an innovation company that has been providing data management solutions for quite some time. From the Pitney Bowes website:
Delivering more than 90 years of innovation, Pitney Bowes provides software, hardware and services that integrate physical and digital communications channels. Long known for making customers more productive, Pitney Bowes is increasingly helping companies grow their business through advanced customer communications management.
Pitney Bowes has very deliberately built out offerings for data management, analytics and business insight with a strong customer focus. Multiple acquisitions have been made over the years to broaden – and deepen – solution capabilities.
Navin Sharma, director of global product strategy, and Robert Cruz, vice president of business development, met with the Hub Designs MDM Think Tank to talk about the approach that Pitney Bowes is taking for master data management and data governance with the Pitney Bowes Spectrum platform.
Solving Customer Problems
Most customers are based in North America, but Pitney Bowes is expanding around the globe.
Pitney Bowes has built a strong beachhead in services for multi-channel marketing and customer communications, including customer analytics and segmentation. Other solution areas include risk, fraud and security management for financial institutions and government agencies, and customer-related services for the telecommunications industry.
Companies that have adopted the Spectrum Platform for data management include very large enterprises and government agencies such as Citigroup, AT&T, U.S. Homeland Security, and FedEx. Facebook recently became a customer for location-based intelligence to integrate with its complex social network.
Vision for MDM
To understand the positioning of the Pitney Bowes Spectrum platform for MDM and data governance, it’s useful to take a quick look at how Pitney Bowes has evolved its data and digital communications business units. Many of its services over the years have revolved around customer, product, location and related activities, bolstered by technology acquisitions.
The Pitney Bowes focus on helping their client companies work with accurate customer data means that Pitney Bowes understands the importance of the customer for any enterprise. Pitney Bowes is also in a great position to really understand how their client companies want to operate their businesses and the challenges that they face.
MDM is a logical step for Pitney Bowes data management. To decide on the direction and technologies for the Spectrum platform, Pitney Bowes took their time to study the current vendor landscape for MDM solutions, and to consider what needs and pain points their customers have had over the years.
Pitney Bowes didn’t agree with most of the other vendor approaches, which Navin termed as too big, too complex, too expensive, and too dependent on infrastructure. His other comments on current vendor offerings: limited views of the master data, and lack of movement towards the deft handling of big data volumes especially complex social media data / content.
So Pitney Bowes decided to do something different that could handle the new aspects of data today and where it may be heading in the future. After all, master data management should reflect how businesses operate and evolve.
Customers and Relationships
Navin noted that for many enterprises data-related processes need to align with the real world intricacies that companies face, especially in terms of complex relationships and hierarchies with respect to customers and their extended networks. He went on to express that MDM should help the enterprise see the customer’s network of relationships as well as reveal non-obvious relationships and spheres of influence – these views frequently are not possible with traditional CRM systems.
Pitney Bowes looked at many different infrastructure approaches to data hubs: Hadoop, columnar, document, relational – but chose something else: the graph database. At the heart of Pitney Bowes Spectrum is a NoSQL graph database engine to create hubs that better enable agility and responsiveness to enterprise needs and constant business volatility.
As Navin pointed out, the graph database structure reflects the increasingly connected nature of data and the relationship networks underlying customer behavior. It allows for the rapid modeling of real world entities and their connections across roles, processes and interactions
The choice of a graph database infrastructure reflects RDF principles for organizing data: model the data to the actual business. The high level structure of a graph database consists of nodes, properties and edges. Graph databases function without indexing since the relationships between data elements “point the way”. Nodes pertain to entities such as people or activities; properties provide information regarding nodes; and edges connect nodes to other nodes or properties. It’s through edges that patterns of data relationships and interconnectivity can be discovered.
The Spectrum graph database runs on Linux as well as other flavors of Unix, and Windows. Spectrum enables high scalability and fast performance when crunching through large data volumes. Performance is optimized for queries, especially for complex networks of data. Additionally, Spectrum is agnostic regarding hardware and OS environments, and supports multi-domain data management.
Expose complex hierarchies and relationships through visual analytics
Source: Pitney Bowes
For Spectrum, Pitney Bowes is pursuing this mantra: Simple. The offering is self-contained, with no dependency on third-party app servers. Leveraging many years of providing data integration solutions, Spectrum covers just about any way to connect to sources. Data virtualization capabilities are also available to work with complex data types.
Spectrum is a SOA platform (Service-Oriented Architecture) – every capability is available as a service, with extensive support for APIs and external web services. The platform has very clean UIs for the design tools and easy-to-use, responsive modeling tools. Visual design elements can be added by applying themes to make flows and models easier to understand and manage.
Data quality and matching are well integrated into the offering. Again, Pitney Bowes has logged a lot of miles in the data quality world, with a long history of services for managing names, addresses, and identity through extensive and sophisticated capabilities:
- Rules and algorithm-based matching strategies
- Match-rule analysis
- Best of breed record creation
- Open Parser – define grammar by domain, country, language
- Multi-cultural name recognition
- Identity matching algorithms with global name variation data
- CASS, SERP, AMAS, SNA, UPU standards based global address verification
- Extensive data enrichment; interface with multiple third-party tools
Data governance capabilities include role-appropriate user portals for business stewards (with browser-based access) and IT stewards. Dashboards are provided for monitoring current issues, to-dos and scoring of data quality. Data stewards have access to detailed profiling and monitoring for low-level data quality issues. Since data governance benefits from team effort, Spectrum supports collaborative discovery and issue management
A key capability that has been brought into Spectrum is based on Pitney Bowes location-aware intelligence and global geocoding capabilities. Pitney Bowes asserts that the ability to visualize spatial data and understand relationships between specific locations will allow organizations to make more strategic business decisions.
Other features: design tools include an automated bi-directional metadata flow when new components are added; and analysis capabilities are available for data in the hub. Spectrum supports flexible MDM deployment styles: Consolidated, Centralized or Co-existence.
Customer Insight: Analytics and Big Data
The Pitney Bowes focus on customer comes through their articulation of the value of Spectrum: it’s less about data management per se, and more about the impact on the customer’s business. Pitney Bowes sees actionable insight as the real purpose of data management infrastructures. They appear to be taking care to provide a high performance infrastructure to bring reliable data into processes to derive that insight, as well as the subsequent decisions and actions.
Analytics capabilities include support for predictive analytics, BI reporting and data visualization.
Spectrum also handles big data mining and analytics, leveraging MapReduce-based implementations for large batch processing. In Spectrum, Pitney Bowes combines social and spatial analytics with the more traditional BI and analytics options, making the claim that it is the only MDM vendor that “can add Spatial, Social and Predictive Analytics dimensions to the master record”. Such an approach can provide greater depth for customer intelligence that can aid product, marketing and sales strategies and execution.
Predictive modeling can be used for customer activities such as the propensity to make purchases under certain circumstances or to identify the interest of particular customer segments for certain products that they have not yet purchased. Such analytics are fueled by the Spectrum ability to navigate social networks and spheres of influence. This type of modeling can also work well for processes that handle risk, fraud, and security analytics.
The Future of Data at Pitney Bowes
Pitney Bowes software solutions have been included in a wide range of Gartner & Forrester benchmark reports for several years: from Managed Print Services to Data Quality Tools to ETL and Data Integration. And Pitney Bowes applications and services run the gamut from call center optimization to customer communications and digital marketing.
This puts Pitney Bowes in a unique position to truly understand the data pain points experienced by many kinds of companies. It also means that they understand a lot about how different companies do business, and why a sharp customer focus is a top competitive differentiator, both for themselves and their customers.
Pitney Bowes has set out a new way to offer master data management capabilities. This company has the chops to understand why their approach should work well for many enterprises. It’s likely that Pitney Bowes will continue to innovate their approach to MDM and that the platform will continue to be responsive to customer needs as organizations and business conditions evolve over time. Pitney Bowes already offers cloud services for other solution domains – this could well be the next move for the Spectrum platform.
About the Author
Julie Hunt is the editor of Hub Designs Magazine, and an independent software industry analyst and consultant for solution and customer strategies.