Check out the on-demand sessions from the Low-Code/No-Code Summit to learn how to successfully innovate and achieve efficiency by upskilling and scaling citizen developers. Watch now.
Business leaders are increasingly prioritizing digital transformation agendas. However, in their rush to change, many neglect to develop fundamental data and analytics strategies to inform those digital transformation agendas and support data-driven technologies. In fact, a recent report by the Capgemini Research Institute found that 84% of the world’s leading enterprises not only lack proper data and analytics strategies but also the foundational processes, systems, and tools to truly become a data-powered company. As a result, these organizations fall behind their more advanced competitors that achieve 22% higher profitability on average.
To fast-track their data and analytics strategies and ultimately compete with leading data-powered enterprises, organizations must modernize their data estates. And they should start doing so with these five steps:
1. Map value streams
One of the first steps along this journey is to use value stream maps — visual charts that detail each step throughout the product delivery processes that ultimately add value to the customer. These value stream maps help you identify and prioritize larger business objectives, such as expanding market share or improving customer engagement, while also providing a high-level overview of the data modernization efforts needed to achieve these priorities. By thoroughly mapping objectives, technical capabilities, and datasets, you can holistically understand which legacy dependencies and abstract applications no longer serve their purpose and identify what capabilities are still needed to accomplish each business objective and modernize your organization’s data landscape.
To successfully create a value stream maps, you’ll need to answer these questions:
Intelligent Security Summit
Learn the critical role of AI & ML in cybersecurity and industry specific case studies on December 8. Register for your free pass today.
- Who will lead this project, not only from IT but also the business?
- What are the larger business objectives I intend to map?
- What datasets do I have at my disposal and how do they align to my objectives?
- What data modernization tactics are needed to achieve these objectives? And how can I group applications and infrastructure required to launch applications?
- What are the detailed steps I must organize to align these data tactics to my overall objectives?
- What capability gaps do I need to fill in my team and how do I fill them?
2. Decommission legacy data in phases
Many of the enterprises struggling to keep up with their data-powered competitors have used outdated, monolithic systems, or legacy data, for years. Moreover, these companies typically lack the processes needed to thoughtfully transition to modern platforms. Although these companies shouldn’t necessarily uproot their systems and replace them with new platforms too quickly, they still should devise processes to ensure an orderly transition over a designated timeframe.
To begin such a transition, you should map relevant transactional and historical data with the new data systems. Then you can transform data for use in the new systems. After this, you can divide the monolith into containers of larger, independent components to ensure all work efforts ultimately map to potential business value. Although these steps will take time, by decommissioning legacy data in distinct phases, you can modernize your data estates in a manageable, cost-efficient manner and attain significant ROI in the process.
Here are the questions you must as you begin decommissioning legacy data:
- What team will manage this project and does that team have enough of a business lens, with business leadership?
- What are the pain points with the current monolithic system?
- What modern platforms can best serve the data estate?
- How can I carefully make this transition – what are the detailed steps to move each system over?
- Based on the results from mapping the organization’s historical and transactional data, how can I divide my monolith into independent components? And how am I ensuring this process ultimately maps back to my business strategy?
3. Integrate a multi-cloud system
Many of the world’s largest companies are prioritizing multi-cloud systems to drive cost-efficient innovation and enhance analytics capabilities at scale. In the future, competitors should expect to adopt these systems as well. This will require an enterprise to allow each business domain to handle its own data sets, while simultaneously planning for interoperability within multi-cloud environments. To achieve this, you should look to cloud vendors to migrate your systems, code, and applications to various clouds. You should also use federated learning to work with distributed datasets as you move to a multi-cloud system to ensure you can use external data while preserving the privacy of your own internal data.
That said, below are questions leaders will need to think through before making this major change:
- Who will lead this process? Specifically, who will manage the overarching multi-cloud engagement and vendor partnerships? And who, within each business domain, will act as cloud point-person(s)?
- How are data sets currently organized within each business area? Is each domain prepared to make this transition to the cloud?
- What vendors are available, and which will serve my specific business domains best, given their management styles, systems, and applications?
- Is my cybersecurity strategy, including tools and platforms, equipped to take on this multi-cloud frontier? What other measures must I consider before making the leap to a multi-cloud environment that will use external data?
4. Customize data-discovery tools
As companies modernize their data estates, it is essential to develop frameworks through which they can understand their data. This can be done through data-discovery tools that collect, evaluate, and recognize patterns in data from various sources. However, before doing so, organizations need to decide if they will build, buy, or customize data-discovery platforms by identifying existing data assets and current issue areas. In many cases, organizations will use a combination of these platforms to meet their various needs. In fact, enterprises must prepare to continuously update their data-discovery tools into the future as needs change and data and analytics initiatives scale.
This customize process can only occur as the overall data incentives scale as well. Thus leaders will have to address the following questions to see this drawn out process through:
- Do I have a team dedicated to data discovery? And do they have the ability, and the institutional knowledge, to customize discovery tools, based on the organization’s needs?
- Does my company have a thorough process of connecting multiple data sources, cleansing and preparing the data, sharing the data throughout the organization, and performing analysis to gain insights into business processes? Given this assessment, what are my gaps or issue areas?
- Do I have an existing data asset that can be used for each given problem? Does this tool need to be customized in any way — how so?
- As I begin to implement new tools, which teams will be affected by the changes?
5. Accelerate innovation with DataOps
To continuously modernize data estates, organizations should begin looking to the future of data-driven innovation — data operations. As 85% of leading data-powered enterprises are already deploying DataOps practices to improve the quality and speed of end-to-end data pipelines and 90% are using it to deliver faster analytical solutions, competitors should move quickly to adapt. By establishing DataOps strategies that focus on a culture of collaboration with cross-functional teams, metadata management, and automated data provisioning, companies can achieve continuous data flows, gain faster access to actionable intelligence, and spur the creation of valuable products and services.
To create DataOps strategies, leaders must ask themselves:
- Who will sit within the DataOps team, and how can I ensure that this new team is tightly integrated and multi-disciplinary?
- What will be the practices and protocols while improving end-to-end data pipelines?
- How do I tie DataOps into existing frameworks and processes for DevOps? Security? Lifecycle management?
- How will the team identify and handle data drift?
- How will we define our metadata, and what platform within the overarching data platform will handle metadata management?
As enterprises rapidly transform their digital infrastructures to keep pace with the modern market, leaders should not neglect their fundamental data estates. Without modern data infrastructures in place that enable business value through prioritized use case deployment, organizations will not only impede their digital transformation agendas, but also fail to gain business value from data-driven solutions.
Jerry Kurtz is EVP of Insights and Data at Capgemini Americas.
VentureBeat's mission is to be a digital town square for technical decision-makers to gain knowledge about transformative enterprise technology and transact. Discover our Briefings.