The Information Age has ushered in multitudes of cloud-based data solutions ideal for analyzing, organizing, interpreting and measuring your customers’ behavioral patterns, demographic breakdowns, interests and detailed customer-service KPIs across numerous touchpoints. Every business or startup with an online presence has an opportunity to cement lasting success and flesh out a more thorough roadmap by getting to know customers’ personalities, preferences and pain points better and developing more personally designed products, services and marketing initiatives rife with genuine value. The insights you leverage and the real results they yield depend entirely on how you manage the individual, sensitive data with which your customers trust you.
As a general rule, human beings like your customers would rather be known by their names instead of impersonal numbers. Since countless individuals around the world share nearly exact combinations of first and last names, databases often work around this potential confusion by tracking records according to phone numbers, email addresses or individually generated ID numbers. Still, a name is sometimes all there is to work with as a unifying data point. A name matching tool neutralizes nicknames, multiple possible spellings of the same name, translation flubs and other potentially confusing factors to improve tracking through several useful approaches. Here are just a few methods that many name matching tools utilize.
The Common Key Method pairs similar-sounding names based on a key or code derived from known English pronunciations. Although favored for high recall and fast execution, this technique is limited to Latin-based languages and requires considerable precision to transliterate non-Latin monikers.
The List Method compiles each name component’s full array of possible spelling variations and searches for matching names. While easy to maintain, names the system does not recognize, split between two fields or that contain missing or added spaces are difficult to track. Also, expensive hardware may be required to rapidly parse long lists of name due to this method’s computationally intensive nature.
The Edit Distance Method is easy enough to implement, but it may miss subtle linguistic nuances by weighing all swaps evenly and struggle with names not originating from Latin-based languages. Names are sorted based on the number of character changes needed to transition from one name to another. For example, since “Cindy” and “Cyndi” have a single transposed letter, the two names have an edit distance of 1. Meanwhile, “Catherine” and “Katharine” have an edit distance of 2, because “C” becomes “K” and an “a” replaces the first “e.”
The Statistical Similarity Method assesses potentially thousands of matching name pairs to recognize when two names look similar and assign similarity scores to them. The precision across various scripts and languages is incredible, but accuracy comes at the cost of slow, methodical performance and substantial effort to adjust features and train data. This particular method may not be the best tool for high-transaction settings.
The Word Embedding Method creates numerical vector representations of organization names based on a word’s semantic meaning. “Car” and “sedan” are semantically similar enough to share a vector space, but “fish” and “wallet” would be considerably farther apart. Although this method catches semantic similarities guaranteed to slip past tools rooted in spelling, it typically only proves effective when matching organization names.
When comparing the merits of ticket tracking versus email-based systems, the former often pulls ahead with several clear advantages. Unlike email streams that have to be continually updated with each new development, ticket tracking software centralizes a single case’s information into one catch-all file. With built-in tracking and reporting metrics to monitor efficiency, anyone can run a single report chronicling every ticket worked by a single person within a given period. Workflows can be clearly and easily prioritized to ensure the most critical issues are resolved first.
Along similar lines, prioritizing the clients with the most prolific support requests allows for allocating resources as efficiently as possible while supervisors monitor exactly what needs to be addressed, where various issues are in the resolution process and who is working on what. As a result, your recordkeeping actually promotes collaboration and continuity with equal, easy access to case history at all levels.
Named Entity Extraction
A named entity extraction process passes a string of text through an algorithm as input and zeroes in on relevant nouns such as places, people or organizations mentioned in a sentence or paragraph. This is exactly how news and publishing houses manage massive volumes of online content generated on a daily basis.
A name entity extraction program can identify the key organizations, people and places discussed among countless full article and generate relevant tags for categorization in defined hierarchies for quick, easy content discovery. From a customer service perspective, the same principle can be used by a big-box retailer to isolate a single location (“Santa Monica”) and product (“Diet Coke”) in order to ensure a complaint is properly categorized for handling by the appropriate department within the organization. Think of it as your own personal search engine.
Creating an Online Customer Portal
More than anything, customers want control over their own online experiences with the brands they love. Instead of a one-size-fits-all customer portal, aspire to build a community around personalized interaction and self-service options for answering questions and addressing issues. The earliest online customer portals managed transactional data, purchase records, order histories and bill payments alongside demographic data including personal profile information and address changes.
The resulting format killed two birds with one stone: customers could manage their own data and businesses spent considerably less money having customer service representatives manually change information as needed and answer and route support calls. Customers can turn to your company first for answers to their questions at any time and avoid the frustration of being routed to multiple online locations for the information they need. On the company’s end, an online customer portal maintains scalable, efficient customer relationship management and standards for effective service as your business grows and you gain powerful new insights from a dynamic community into how your value propositions are received.
Information is power. That flows both ways when it comes to customer data management. Your company streamlines its customer service efforts and leverages potentially valuable insights into the consumers who choose your products and services. In turn, your customers assume direct control over how they experience the value your brand has to offer by enjoying total independent access to the product information and answers to their questions when they need them most.