Chuck Kane

Master Data Management: Moving from the Grid to the Graph

In the age of Big Data, businesses are collecting mountains of information about their customers that, when organized effectively, offers the potential to deliver important and actionable buyer insights. To do this, many organizations store information in data warehouses, or more recently, data lakes in big data environment. However, with customer databases collecting hefty streams of data on a daily basis, wading through and determining what information is useful or relevant for a specific business division or decision can become a very daunting task.

To tackle this challenge, businesses will often deploy Master Data Management (MDM) technologies that aggregate the wealth of incoming transactional data per customer. The goal is to offer a complete perspective of each customer’s relationship with the business, from the first transaction to the most recent. In the past, these MDM solutions were built on traditional relational databases that were able to provide easy views of a defined subset of the information through a trusted data hub. 

However, MDM platforms based on relational databases limit what businesses can do – and how quickly they can do it – with the customer data they collect. With more customer information available today than ever before, there are now greater opportunities to leverage this information across the organization. These opportunities require agile MDM platforms that can react to needs across the business, rather than the narrow demands of a single part of the company. 

via www.pitneybowes.com

Great article on the value of the graph.

Single View of Customer for Banks

It is important to look at gains both in the moment and over time as it relates to improved data and customer view. Without a focus on improving data, quality erodes: people move, companies merge, new products and channels arise, personnel changes—the data problems of today that prompt consideration of improvements become bigger when not addressed.

Costs arise when the data maladies throw a proverbial wrench into the application process. In fact, Gartner comments: “While losses of millions of dollars are significant, we believe these estimates understate the true financial impact on most organizations—the actual magnitude of the problem is typically far greater (by orders of magnitude) than is perceived by business and IT leaders.”

Gains include revenue improvements that result from more efficient programs and higher response rates. They also often include indirect benefits with substantial value: resulting process efficiencies and improved compliance, for example, can both translate into significant indirect savings.

It pays to look all across the organization for ways in which bad data can negatively impact business profitability and customer satisfaction/loyalty. Here’s another example: At onboarding, a bank’s customer information is not collected entirely or accurately:

•           Accounts do not get appropriately linked
•           Fees are charged incorrectly
•           Eligibility for individual accounts and pricing is not recognized
•           Preferences for communications channels, formats and frequencies are not honored
•           Compliance requirements are not met

 The Pitney Bowes Single View of Customer addresses these concerns by:

• Revenue and Profit Growth - Deliver a centralized, precise and comprehensive view of clients across multiple products, across multiple lines of business - enabling a banking wide customer-centric growth strategy that enhances customer sales
• Cost Optimization and Efficiencies –. Thes single view of customer solution will allow improved servicing and marketing processes around the client and should allow banks to better understand customer behavior by providing a consistently reliable, accurate customer master data to the bank's analytics and reporting environment.  
• Compliance and Risk Management – Enhanced regulation restricts fees, increase compliance costs, and further commoditize banking products and services.

Revenue and Profit Growth:

Consumers have changed considerably in the last 20-30 years being increasingly connected and expect every organization they interact with, including banks, to be less complicated to do business with, efficient and more proactive than in the past.   The challenge for banks is to not only remain competitive but increase profits and drive revenue generating activities while facing the challenges of regulation.   Even more so, they continue to need to make investments to update their technology migrating from disconnected legacy systems. 

 The single view of customers creates a power advantage for banks by creating a customer-centric growth strategy that expands sales and profits across all channels.   The most competitive banks are weaving into their products and services additional revenue generating opportunities by leveraging the single view both in the physical and digital world, especially mobile technology.  

Given the experience of the web marketplaces, social media, and online services, customers are expecting banks to keep pace with the evolution of offerings.   Growth opportunities in mobile have become increasingly necessary to drive customer satisfaction.   Most clients have some handheld device expecting 24 hours, seven days a week service.   By not having a single holistic view, customers become frustrated with the inconsistent interactions, inability to quickly process transactions, manage their accounts, perform research on products and services, and interact digitally get information quickly to make better decisions.

Pitney Bowes delivers a solution that enables our customers to create and maintain a consistent view of clients across the digital and physical channels to help reduces customer attrition, by providing real-time access to the comprehensive and accurate view across channels, which results in better marketing, enhanced cross-sell, increased market share all of which impacts the profit growth.

The SVC assists in ensuring offers made are relevant to their customer's that lead to expanding relationships, enhancing the customer experience, increase in portfolio, decreases risk, and compete with non-bank organizations competing for traditional banking products and services.  

Banks can capitalize on this enhanced view by offering financial planning and its related products that directly impact the growth strategy across lines of business.

Lastly, having a single customer view enables banks to expand rapidly by an acquisition that allows access to new markets, customers for cross-selling and reduce operational expenses.

Cost Optimization and Efficiencies:

Like any organization, banks are evaluating options for consolidating aging technologies to support a coordinated channel strategy around the single view of the customer. They should always evaluate the market for solutions that reduce the complexity of implementation, lowering costs while delivering business benefits.   The agility of an organization is defined by its ability to deliver business benefits in less time than the competition and by its ability to quickly identify new opportunities, and find sponsors willing to commit to delivering the business benefits. 

The cost of supporting aging technologies slows many banks from growing and competing effectively in the marketplace.  Pitney Bowes helps banks by incorporating our innovative new technologies to deliver value to customers faster, reduce project overruns, and overall reduce operational costs by eliminating redundant support functions, spreading the cost of technology implementations across profit centers, attract a younger workforce, and eradicate the cost of licensing legacy systems.  This leaves money available for investment in additional projects at a fraction of the current expenses helping banks remain competitive in the marketplace.

Pitney Bowes enables our clients to increase performance and scalability while avoiding costly workarounds caused by limited exception handling which increases the speed of straight through processing, but also ensures the systems are agile enough to handle both new market opportunities and regulatory changes. 

Analytics, Segmentation, and Targeting:

Disconnected systems make it difficult, if not impossible, to discover and analyze data to derive meaningful insights.  This inefficiency makes it difficult to know whom to target, what offers to make at the right time, and which programs to make investments.   When investments do not meet expectations, this wastes valuable resources, time and paralyzes organizations to move forward losing out on opportunity costs.   

Once a single view of the customer has been established, not only can a bank understand wallet share, but also identify segments that positively generate revenue.   The deeper segmentation a bank has, the better it can target customers providing relevant offers for its products and services.  

The Pitney Bowes Single View of Customer ensures a holistic perspective to enabling multiple segments to be leveraged at the same time, enhance targeting strategies and enhanced cross-sell across lines of business.   

However, the benefits extend beyond the customers but also to the employees as the holistic view helps with defining accurate employee compensation plans, effective training, and building incentives drive results in predictable growth, retaining top talent, and lower administrative costs.   

Regulatory Compliance

The cost to comply with these regulations, in just the US Financial Services market, is growing at over 10% per year and will soon exceed $10bn/annually. (Aite Group, 2013)

Dodd- Frank reduces profits and will make it more difficult for banks to differentiate themselves. - IBISWorld, “Commercial Banking in the US” (November 2010).

The Credit Card Accountability, Responsibility, and Disclosure Act of 2009 (CARD Act) reduces the amount of interest and fee income that banks can collect from consumers. It also requires many banks to invest in technology to comply with new rules around billing practices. Some have estimated that the CARD Act will reduce the average annual income of large banks by $500 million to $1 billion and mid-tier banks by $50 million to $100 million. - Diamond, “From Surviving to Thriving: Maximizing Revenues and Profits in the New Banking Environment,” 2010.

Recent changes to Regulation E, which governs electronic fund transfers, are expected to reduce the fee income of the largest banks by up to $1 billion by some estimates.  - - Diamond, “From Surviving to Thriving: Maximizing Revenues and Profits in the New Banking Environment,” 2010.

FDIC insurance premium prepayments already have made an impact, reducing banking industry profits by 37.5% in 2009 according to some estimates.

In addition, changes to the FDIC’s insurance premium assessment framework may reduce the profitability of certain banking activities. . - IBISWorld, “Commercial Banking in the US” (November 2010).

Basel III will more than triple banks’ capital requirements, increasing funding costs and leading to slower growth.

The Pitney Bowes solutions enable banks to leverage the same Single View for AML/KYC in 8-12 weeks compared to other solutions that take 9-12 months by:

• Providing banks a means to quickly onboard data sources and develop insight into customer data regardless of format or source
• Our data federation solution leverages the same entity extraction logic for both source systems and transaction streams
• Prospective customers are included in the same repository as customers
• Provide banks with the high-quality data required to build the transactions that feed into the transaction monitoring systems
• Provide a governance process to review data anomalies of interest and manage the quality of every data source
• Data from multiple sources can be appended to the entities held transaction monitoring system. This can improve the match confidence and insight during investigations.
• Graph’s ability to aggregate data and provide temporal entity analysis will provide bank investigators with tools to keep ahead of money laundering schemes.
• Our solution is Not disruptive, enhances the client AML infrastructure, drive investigative efficiency, and deliver value faster than traditional solutions. 

 Fraud Detection:

• The Pitney Bowes solution has assisted our banking clients to be proactive in the identification of fraudulent transactions by defining a “common nexus” of entities and relationships based on transactions. 
• A common nexus is a single person, branch or group of people, common to a set of transactions that generates a potentially suspicious network.
• The Pitney Bowes solution is used to load and organize the data within the graph so that we can identify transaction anomalies which result in alerts. 
• The graph database is essential to both storing the entities, but also the relationships that drive the common nexus. 
• Our solution was able to identify these fraudulent networks 24 months in advance of traditional methods due to the patterns of the connectivity of the data and the relationships that are unique to generating the suspicious networks.

System of Records transforming to Systems of Knowledge

In the context of business, is a Customer Relationship Management (CRM), order and fulfillment system, a billing system, a customer servicing system each system of record?

The CRM system creates, updates and maintains new client information during the initial sales and origination lifecycle. It serves as the system of record and is considered the authoritative source for customer information for a time. During a successful sale, this same client information is then fed into or entered by hand into the order and fulfillment system. Does the order and fulfillment system now become the system of record or is the CRM still the system of record? After fulfillment is complete, the same overlapping customer data, product or service data is then fed into the billing system and customer servicing system, do these become the system of record? If the billing or customer servicing system update customer data, do these systems become the system of record? Imagine, millions of updates happening during any given time of the life-cycle where prospects become clients, and customers again are prospects, how do you even keep track of which system is the system of record?

Master Data Management is intended to collect this overlapping data per domain and act then as the authoritative source of truth, therefore, relegating the source as subordinate to the master. The master will capture all updates, apply consistent data quality and consolidation, and share its results ensuring all systems are a reflection of the same information. So the master becomes the system of record, and the subordinate systems become systems of knowledge unique to a business purpose.

To further this thinking process, we "know" that knowledge is the collection of experiences and associations that build upon the isolated collections of data. Companies will rarely make decisions based on a single system but rather a collection of systems that support the ongoing business purpose. What separates the high-performing companies in their industries is how well they leverage their collective experiences to create an efficient system of knowledge to drive better decisions.

How implementing a modern data management approach can underpin truly agile CRM?

Change is the only constant in business and to stay competitive in this ever changing the environment, organizations need to ensure they build a flexible, contextually relevant, holistic system of knowledge or the Single View of Customer.

The Single View of Customer, at its core, should be driven by graph databases to exploit connected data connect to drive insight to the influencers, leaders, followers, brokers, decision makers, and the customers most likely to buy or leave your organization and take others with them. The entire domain of relationships in the graph is rich with so much knowledge that when used effectively empowers the CRM system to drive focus and results.

The graph solution is much more than a visualization tool for the data rather its a unique method of storing data as normalized entities with deep knowledge of the relationships between the entities. This model can be easily extended and modified to join vast domains of data together that support the true Single View of a Customer.

The graph empowers your organizations to perform deep traversal queries to ask questions of the data without having done any extraction of the data, denormalization of this information and then fed into some analytical engine to get results. Instead, it inherently can ask multiple different types questions of your data which is essential when looking for meaningful ways to develop campaigns that will yield the most results.

For example, when there is sufficient amount of data a graph-based solution can rapidly traverse the data model to identify customers that work for a particular employer, with a specific salary cluster, who are avid hunters, who are parents of children aged 10-12 that play soccer in a fraction of the time of non-graph solution. Of course, you can do things with big data and other no-SQL solutions, but graph databases are the right fit for real-time operational use cases that require you get answers now.

Going further, within the same query we can also identify customers that are most likely to be influencers not only within their community network but to exert influence with other clients in other communities who share similar attributes leveraging graph real-time analytical capabilities.

Based on this example above, let us assume there are multiple hunters who also enjoy fishing and live in the same communities; we can also infer that over some period they may meet and communicate at soccer games. In fact, the reality is they would positively or negatively influence each other, thus altering future shopping patterns.

If you wanted to create new campaigns but don't have the data integrated yet? No problem, just extend your graph model by leveraging our virtual graph capabilities by leveraging a data federation solution that enables these virtual extensions. Once extended, you can query and traverse the data in the same way you do for any other queries to drive value in your campaigns.

Data enrichment can help uncover new hidden insights

Additional insight into gaining insight into successful campaigns can also include enriching the data with third party data sources such and Experian, Neilson Demographics, and Location Intelligence. By simply taking address information and converting this to a set of lat/long you can now drive additional insights such as the total drive time to the nearest store, the closest competitor to the products and services offerings. You can present your customers on a map and identify store locations where most of your customers originate and areas where there are opportunities.

It's critical that in moving towards a more data-driven market that leveraging as many data sources possible and connecting this data together in a way the drives value in the shortest time as possible will accelerate your business objectives

Data Consolidation – Problems we all face

Businesses across the globe in every industry desire to use data to drive insight that drives positive performance in their respective organizations. Effective integration, standardization gain the foundation to this insight, and consolidating the data and relating the data together to achieve knowledge and wisdom. Consolidating data has proven tough due to the limitations of many of the de-facto systems deployed within all organizations such as relational databases. In fact, much of the related technologies such as data transformation tools are so prevalent because of the shortcomings of relational databases, but still, even with all the tools, it's been elusive in using data to achieve meaningful relationships and contextually relevant insight.

Strict Schema Solutions

Relational databases, while having many strengths over their predecessors, have proven time and time again that they are difficult to create, rigid to change, and unable to support traversing n number of domains with ease or without having to replicate the data into various structures to gain insight.

Once the data is locked into a “hardened” model, along with the different strategies to optimize the performance of the applications, it becomes difficult to extend the model over time without impacting existing applications. Also, these structures are not very relationship oriented, in that the relationships of the entities across the tables are hidden within the tables of relational databases themselves making it harder to create or enhance later as new relationships are formed, or existing ones are updated or removed. The applications that depend on these strict schema structures to also become rigid over time as they are reflective of the underlining foundation.

Organizations have tried many methods to resolve these problems, but the most prevalent one is the creation of the operational data stores representing the use cases of the applications they support. The goal is to increase the quality of the data within that domain and provide an easier method to enable applications to gain access to the data. But what ends up happening is these data stores are lacking consistency, the quality of the data is not always at its best, and data is replicated all over the organization further resulting is the lack of trust.

To combat this one off operational data store, organizations adopted programs to master data in a single application where the quality of the data could be governed and trust in the data restored.

In fact, an MDM program is an operational data store that supports a given domain with prescribed models, queries, services, data management, data quality, data governance. Each supports the idea of simplifying the data management problem and creating a single view and the ability to link that domain to all other domains quickly.

The results have been lacking, and most organizations are not satisfied with the results. Why? Because these rigid data structures just do not lend themselves well to linking the multiple domains together, so solution providers tried to encapsulate multi-domain master data management applications, with the core focus being, say customer, and have a light version of other domains attached to the customer domain. These light versions just steer organization right into the train wreck they were looking to avoid by having a single data store that could house multiple domains to drive meaningful relationships.

Graph Databases = New Insights

But there is an answer to this problem and graph databases powered by solutions that have data quality as the focus, will enable businesses to build a comprehensive view of the data domain, to eliminate silos across data domains which ultimately result in getting to the single view right.

The customer is foundational in collecting, standardizing, consolidating them in a graph model where we can then integrate the various domains to identify relationships, uncover opportunities, and engage customers in a more meaningful and consistent way across the different channels.

The graph inherently supports the multi-domain and multi-contextual approach which is critical in making data fit for use across a much broader spectrum of knowledge to gain insight that leads to knowledge and wisdom.

Also, the flexibility of the graph database allows for businesses to grow and expand the data to include new and every changing requirements which improve the chances of success dramatically. This foundational change will certainly result in our ability to achieve the holy grail of the customer 360 view.

Moving Forward

With every new solution, there is a period of risk aversion and initially slow adoption. In working with the early adopters, we have achieved results that were just not possible in previous solutions.

Leveraging the graph, coupled with an agile solution stack, clients have achieved spectacular results in as little as six weeks. This insight leads to a consistent yet dynamic view of the customer, they are faster to market and deliver results at a much lower cost that in the past.

So why continue to struggle and spend millions of dollars on projects that provide lower returns because the integration, transformation, and consolidation of data are severely limited by its foundation? The plethora of data quality solutions was certainly the same used by organizations across the globe, but after seeing the results by making this shift from relational to a graph, customers are again excited about the future of gaining insight leveraging solutions that scale and ultimately meet the needs of our respective customers.

My advice is not to get left behind. Adopting graph databases is the wave of the future for data consolidation efforts.