In 2017, professional cyclist Luka Pibernik was leading the annual bike race Giro d’Italia by over five seconds when he threw his hands up in victory a few meters from the finish line. The problem was, there was still another lap to go. By the time he realized his mistake, and buckled down for the last lap, he was in 148th place.
So, Luka Pibernik did not win the Giro d’Italia … Dewey did not defeat Truman … the Falcons did not beat the Patriots in Super Bowl LI … Jean van de Velde never won the Open Championship … and Leon Lett never scored that touchdown. History is littered with cringe-worthy moments (or endless replays) of premature celebrations and declarations of victory in sports and in life.
We also see similar tendencies every day in companies striving to execute on modern decisioning programs. In our recent research, an overwhelming majority of companies at various stages and maturity levels of modernization say they haven’t yet realized the full value of their data modernization and artificial intelligence (AI) industrialization efforts.
From our observations, this is a clear-cut case of declaring victory before the job is finished — which, in most cases, is caused by focusing on the tactical elements of data modernization or over-valuing a simple AI proof-of-concept and neglecting the bigger picture that brings the entire initiative to completion.
Play through the whistle
It’s easy to see why this happens. For the CIOs and CTOs who champion and execute data modernization strategies, the migration from legacy architectures to the cloud is a victorious moment. And, of course, it’s exciting to see an AI pilot perform well in a lab environment.
But while it’s completely justifiable to take a moment to appreciate these moments, it’s not time for a victory lap. You are likely somewhere squarely between your point of origination and destination. For the people in marketing, HR, finance and other business areas hoping to realize value from data modernization, initiatives like legacy/infrastructure modernization and AI pilots are merely the opening whistle.
Meanwhile, the more junior roles within the transformation journey are incented to declare victory at milestones along the process. “My job is to transition our cloud data warehouse from on-prem to the cloud” or, “My job is to enable the call center to be able to know the customer before they call” or, “My job is to create the best AI model that will work in a specific proof-of-concept.” While important milestones, these are not wins.
The real celebration
The fact is, modern decisioning relies on both a technical and operational construct that delivers the data to any point of decisioning and transcends to the people or platforms responsible for activating the insights-driven decisions. Those AI pilots need to come out of the lab and be scaled and deployed across the organization and embedded in how work gets done.
In short, true value realization comes from achieving agility, flexibility and fidelity in all aspects of the enterprise in both response to and anticipation of the ever-changing landscape in which the business competes. Achieving that requires equal parts data transformation and organizational change management.
The four elements of data transformation success
Data transformation initiatives need to go beyond the architecture-focused step of legacy data modernization to deliver full value. True data modernization programs incorporate four essential elements, including:
- Data fabric: establishing an enterprise-wide architecture that encompasses all components of the data value chain. The data fabric is usually created through the migration of legacy systems, infrastructure and platforms to a cloud-based ecosystem. Crucial components include a hyperscaler system, a cloud data warehouse, master data management and governance capabilities, and a reporting infrastructure.
- Data operations: deploying the agile frameworks and systems needed to manage the day-to-day processes that monetize the data asset. Businesses need to rethink how they use data within the organization and elevate a dormant asset (data) to a monetizable structure. This includes DevOps, smart operations, Agile processes and modern software delivery approaches like continuous integration/continuous delivery.
- Data in action: creating the pipelines and processes to get data to the people and platforms responsible for making decisions or taking action and then reincorporating relevant data back into enterprise systems. The purpose of data modernization isn’t to build cathedrals to worship at the altar of data but to bring it to “the masses” where it can deliver value. As we’ve said, more data doesn’t make you more digital — the value comes from identifying the exact right data that can achieve the insights you need and planning how you’re going to activate that data.
Important components here include the customer data platform, a data hyperloop, real-time data pipelines and participation within one of the many data marketplaces that are emerging.
- Data responsibility: developing tools and processes to ensure the long-term viability of the investment in your data asset. You can spend millions to build and run a new technology infrastructure, reorganize your team and restructure processes to embrace data modernization. But one slip in the area of governance, privacy, security or ethics can quickly jeopardize the entire data asset.
That’s why key elements for any data modernization initiative include an ESG strategy, privacy and protection platforms, and a way to address the burgeoning area of “green data.” By focusing on these sustainability areas, businesses can future-proof their investment and make sure they continue to deliver value over time.
Reason to celebrate
When all four elements are harmonized, the modern decisioning vision can be fulfilled. Consider a manufacturer we worked with that wanted to ensure it could fix problems with its machinery before breakdowns occurred:
- Data fabric: We equipped 11,000 manufacturing machines with cameras and seismographs that continuously monitor for changes in motion or sound. The data is sent to a Microsoft Synapse cloud warehouse for analysis on the Azure cloud.
- Data operations: Data pipelines were set up that could handle the large audio and sound files, and models were built to transform the data into readable formats. Data is captured and analyzed to determine what was worth keeping and move into the system.
- Data in action: Alerts let people know when a machine’s indicators ae out of bounds and feed data back into the system.
- Data responsibility: Quality assurance was a very big concern, whereas privacy and ethics was less so, as the system was focused on machines, not personally identifiable data.
Be Brandi Chastain
Consider the 1999 Women’s World Cup. Brandi Chastain kicks the game-winning goal. She takes a minute to celebrate. The next day, the team gets an incredible parade, where together they all stand in victory. And, the very next day, they all go back out and practice again.
Understand this: The data transformation journey is long, winding and arduous. There will be many successes and bumps along the way. Take the time to celebrate the successes … take the time to learn from the bumps … and when it is the right time, declare victory. Then, two days later, get back to practice.