For businesses, one of the most tantalizing promises of the digital revolution is the ability to use data to create value—to generate ROI from the raw information streaming in from professional networks as much as from actual “things.”
Beyond buzzwords and analyst-speak, putting data into practice
This post originally appeared in InfoWorld*
Yet businesses have struggled to articulate how to go about extracting value from this new source. As Gartner’s Andrew White recently noted in a review of data and analytics in strategic plans, “The vast majority of such documents do not make any reference to a real measurable business outcome [and] … misconstrue the term strategy for plan, action, decision, risk, principle, goal, direction, effort, tactic, and so on.”
Why does it have to be hard to link data to simple, easy, and meaningful business outcomes? In part, this is because these databases are often large and full of errors, making them cumbersome to use. For example, in a bank, tellers interact with customers every day. While the bank’s system may have duplicate records of a customer, the tellers aren’t confused about whom they are speaking to, because the customer is in front of them.
But no matter the quality of the teller’s service, if the customer later gets multiple touches from the bank, for each of their duplicate details, the customer may think, “Don’t they know me? I was just there last week!” Couldn’t this easily be avoided by prompting the teller to ask a few simple, unobtrusive questions about the customer’s details?
Over time, a bank may have grown multiple systems, powering different applications, across departments and divisions, that may not talk to each other. The teller, however, is only interacting with one system, and the bank would not (and should not) surface all of the customer details to the teller.
If the bank used a system which could service up just the right bits of information to that teller, issues like this could be solved by a simple question like “Is your current email jim@xyz.com or jim@abc.com, or are they both you?” The pieces could fall into place, and the systems in the back end could make the corrections and roll them through all the pertinent systems.
The power of doing so is clear: Better data for the bank strongly equates to better up-sell and cross-sell opportunities, and ultimately higher customer satisfaction.
Application data management
In my example, each of the bank’s applications (checking, savings, loans, investments, etc.) has data that is specialized to that system. For example, while credit rating and employment details are part of the loan information, they’re not included in checking account information. Similarly, recent ATM transactions are not needed for investment account detail, though they are relevant for checking. All this data is managed by individual applications in the bank.
However, some level of data is pertinent to all accounts: name, address, phone, and perhaps current account balance. This shared data that exists in multiple applications is where problems start to occur, and a whole industry of data quality and stewardship strives to be the solutions. The data that is common to, and thus managed across, multiple applications falls under the concept of application data management (ADM).
ADM is a technology-enabled business discipline designed to help users manage and govern the application data required to operate a specific business application or suite (like a CRM, ERP, or SCM). This could include master data management (MDM), and has the same technical capabilities, but its purpose is fundamentally different from MDM.
Bringing it all together: the data hub
A data hub brings together the concept of mastering that shared data that lives across applications such that it can be governed and managed in a centrally understood, nondisruptive way.
The data hub can be logical and/or physical. It is not about transactional storage, but rather for mediation (enforcement) of information governance and sharing policies. It gives a 360-degree view of everything in the organization by reconciling top-down and bottom-up, as well as inside-out and crowd-sourced, data.
The data hub has all the capabilities of an MDM, augmented with important parts that enable it to be a data management system of record, source of truth and system of engagement at the same time. A host of capabilities (such as data governance, quality, enrichment, and catalog) are required to meet this widespread business need—and ideally, should be in a single platform, for ease of use and fast time to value.
Current thinking
I recently had the chance to see how the industry is embracing this way of thinking at Gartner Data & Analytics Summits in the US and UK. As a handful of those analysts start to codify the data hub and ADM, it seems that talk around data hubs and ADM is finally catching up to the solutions in place and should mollify the frustration end-user clients feel with lots of uncoordinated bits of software that are patch together with glue and marketed as a single “platform.”
Recent developments in artificial intelligence and machine learning will help the data hub become a reality. In this way, the AI in an intelligent data hub can be applied under the covers in a platform to help with meaningful business processes. For instance, smart algorithms learn enough to eventually make suggestions that help with automating stewardship, prioritization of workflows, matching, merging, and calling on external microservices to augment the core horsepower in the hub.
This allows the hub to present master data and non-master data objects (system of record, static or dynamically changing values, transactional data, etc.) in easily digestible form to any type of user—like a bank teller.
The days of a bank teller helping customers get better service, and the bank providing a more integrated experience for its customers are upon us in the age of the intelligent data hub.