.excerpt-thumb {display: block !important;}

How to Achieve Reliable Customer Data


Originally published at Information Management 

Reliable customer data is crucial for the success of any business. It lets you communicate with your customers effectively, informing you which products, offers, services, and messages they’d be most interested in at any given time, and which communication channels they’d likely prefer.

You can see information about previous interactions customers have had with your company, and make sure that you have things like correct affiliations and spelling of their names – all of which can lead to a better customer experience and prevent loss of business due to simple mistakes.

But attaining reliable customer data is often an ongoing challenge. Just look at a typical omnichannel customer journey today – customers may visit your website looking for a product. They may use a mobile application to order it, and if there’s an issue or they need help, they may contact a call center.

They may also go to the store to return or exchange the product. Meanwhile, they subscribe to marketing emails for coupons and deals and follow the company on social media. This is how customers do business today -- they pick the channel and the time of interaction most convenient to them.

Capturing all this information and supporting these complex omnichannel journeys requires systems to maintain and share customer interactions, transactions, and profile information. But disconnected systems and channels make it challenging to create a single source of reliable customer information, so data often remains siloed.

Point-to-point system integrations are expensive to build and maintain. In these scenarios, sales, marketing and support teams have little confidence in the quality of their customer data, often questioning if it’s up-to-date and clean. Unreliable data leads to uninformed and inconsistent communications with customers, and, in turn, poor customer experience and loss of business.

What’s needed is for organizations to be able to bring customer data from all internal and external sources together into a single source of truth -- information that includes demographics, multichannel interactions and transactions, social media feeds, and third-party data subscriptions.

Data from various sources needs to be matched and merged using various rule sets and machine learning algorithms to create a single “golden record” for each customer. Modern data management solutions are helping answer this need, and enabling all business users within an organization to access and contribute to a standard customer profile.

But storing all the customer information in order to create reliable data profiles requires big data storage – a company may have millions of customers, and each customer may have hundreds of attributes and thousands of interactions and transactions.

Options such as Cassandra can help -- with modern database technologies, organizations can utilize a wide rows model to implement as many attributes as they need and add new attributes at any time, providing infinite scalability and flexibility. Not only do business users such as sales, marketing, and support need reliable customer profile data -- they also need to know about customer relationships, affiliations, value, and influence.

Graph technology, similar to that used by LinkedIn, helps establish many-to-many relationships between people, products, locations and other real-world entities. It enables business users to learn about the products purchased, channels traversed, and content consumed by each customer. Additionally, you can discover information about members of their household and identify others that may influence their decision. Having reliable customer data is key to unlocking this ability.

With the reliable data foundation and customer graph created, organizations can utilize big data analytics engines to process the reliable data and make informed business decisions. Options such as Apache Spark can enable relevant insights about customers and even provide intelligent recommendations to teams using graph analytics, aggregations and joins and machine learning.

Since maintaining high scalability, performance, and security is crucial for your customer data store, an intermediate storage to provision normalized data structure and format for processing engines is necessary.

Real-time insights and actions is key as well -- customer data is continuously flowing in and needs to be merged, and business users may make changes to the data, and even add new customer attributes. This requires continuously refreshing the changes in the normalized data in intermediate storage for use in subsequent data analysis.

As data sources and volumes continue to increase, it’s becoming more and more difficult to achieve reliable customer data. Modern data management platform with schema-less model to store information at a big data scale, graph technology, and big data analytics engine can help create a reliable data foundation and deliver best-in-class data-driven applications, with relevant insights and recommended actions, to inform business users of the next-best actions for customer engagement, helping organizations achieve success.

(About the author: Maxim Lukichev is a software engineer and researcher with over 10 years of experience, currently serving as the lead data scientist at Reltio. You can contact him at maxim.lukichev@reltio.com.)