Accenture x Starburst: Achieving Data Democratization with Data Mesh 1 E - B O O K Achiev i n g D a t a D em o c ra t i z a t i o n w i t h D a t a M e s h An Accenture eBook, Compliments of Starburst Accenture x Starburst: Achieving Data Democratization with Data Mesh 2 Teresa Tung Cloud First Chief Technologist, Accenture “ I’m excited about Data Mesh because it gives companies a vision to rally around regardless of where they are on their data journey. At Accenture, we recognize that digital transformation is powered by the ability to act upon data. Aiming for a standard, integrated data remains a goal for data-driven businesses but made challenging to achieve due to the rapid shift in data needs, business priorities, and emerging technology capabilities. That’s why there’s interest in how businesses recalibrate towards a federated Data Mesh architecture and the associated strategy and practices. For those who are unfamiliar, Data Mesh is a decentralized approach to data management, striving to evolve the data platform from a technology-led project-centric model into a paradigm about federated business-led product- centric data, by design. This opens the door for more scalability and flexibility as your business evolves with resilience under an uncertain economic climate. This ebook by Accenture’s Cloud First Chief Technologist Teresa Tung explores key touchpoints that have endured as Accenture has helped new and mature organizations (i.e. insurance, finance, health and life sciences, oil and gas) make the shift towards Data Mesh and achieve data democratization. Yes, data democratization is no longer a nice-to-have, but a way to meet your digital transformation goals. Accenture x Starburst: Achieving Data Democratization with Data Mesh 3 The Digital Achievement Gap The time to innovate your digital strategy is now . But the reality is, not many organizations are taking advantage of this opportunity. At Accenture, we surveyed¹ 6,000 executives to measure what we call technology leaders and technology laggards . We define leaders as those who invest in digital growth, and laggards as everyone else. Before COVID-19, we found that leaders surpassed the laggards in performance and growth by double Post-COVID, that gap has grown to leaders demonstrating five times higher success metrics than the laggards. These findings were not a huge surprise. Those who made the investments in expanding their digital foundation carved out a clear and visible path towards business growth and success. With insights, technology leaders saw red flags when they were happening, took decisive action, and implemented changes. This told us that the laggards were more than likely in the dark about how aligned their business and data strategies needed to be. Even worse, they weren’t in the position to even see and understand what was happening, or the options that were available, and were unable to act upon it. Our advice? Tackle data at its core. Data management is hard, and a lot of times, digital investments can fall to the periphery. But in actuality, your data foundation is the core of your business and a difficult one to tackle. DIGITAL ACHIEVEMENT GAP Pre-Covid 2x 5x Covid ¹ Scaling Enterprise Digital Transformation Accenture x Starburst: Achieving Data Democratization with Data Mesh 4 Data Democratization Desired Outcomes When organizing your digital strategy with data democratization in mind, a number of desired outcomes should be at the forefront as business and technological strategies are undoubtedly interlocked: • Unlock from “tech debt” to “tech wealth”: Data is only useful if it can be accessed. Unlock the trapped data value that you’re currently missing out on. • Reimagine how you operate, collaborate, and innovate: Reorganize your approach to get the most out of your newly unlocked data. • Empower the “real experts”: Make sure not only your software engineers, data scientists, and developers are given the right tools and resources but your business analysts and end-users, too. They’re the real experts who know the business process, the customer journey, and the regulations that need to be addressed. • Establish your role in your ecosystem: Consider your role as well as the relationship with the data. When data products are available and considered valuable on a balance sheet, it elevates the overall conversation. It’s time to tackle the core DATA IS HARD DATA DEMOCRATIZATION OUTCOMES • Legacy and silo-ed • Manual, time-intensive, costly • Requires Hi-level of Tech Expertise • Unlock from Tech Debt to Tech Wealth • Reimagine how you Operate, Collaborate, Innovate • Empower the Real Expert • Establish Your Role in Your Ecosystem Accenture x Starburst: Achieving Data Democratization with Data Mesh 5 Data Mesh is a Strategy for Data Democratization Data Mesh is an evolution of the data platform as we know it, from something that is very centralized into something that is much more federated by design and all the processes and operating models that support that. The data platform started with the data warehouse, which is incredibly application-centric and specialized. This then led to a move towards the data lake, which while centrally owned and governed, allows for some limited grouping of end-users and owners. The newest generation is Data Mesh, a decentralized, distributed approach to enterprise data management. It is a holistic concept that sees different datasets as distributed products, orientated around domains. The idea is that each domain-specific dataset has its own embedded data engineers and product owners to manage that data and its availability to other teams. Data ownership drives a certain level of responsibility, which is often lacking in the current data platforms that are largely centralized, monolithic, and often built around complex pipelines. EVOLUTION OF THE DATA PLATFORM DATA WAREHOUSE 1st Generation DATA LAKE 2nd Generation DATA MESH 3rd Generation Principles of Data Mesh • Decentralized ownership of data • Data products • Self-serve infastructure • Federated and global governance Accenture x Starburst: Achieving Data Democratization with Data Mesh 6 Domain-Driven Data Ownership Domain-driven ownership is at the heart of the decentralized approach that allows each specific domain owner to have primary control over their data. This can only be successful if those domain owners are equipped with the right technology and process. As part of the our survey pool in our 2021 Technology Vision, 89% of executives² stated they believe that their organization’s ability to generate value will be increasingly based on the limitations and opportunities of their technology architecture In other words, there is an immense amount of “tech debt” circulating the industry, which can hold organizations back from taking on a new approach. However, we recognize that there’s value in legacy applications. So the answer isn’t in the resistance, but working with what’s available. Decentralized domain ownership means that data doesn’t need to be centralized before tapping into its value. Or tap into data in legacy systems, or modernize as fast as possible if that’s what the data needs. Either way you start looking at the value that you have today, to quantify it in a way that then can be shown to others within the organization and plot that path to unlocking more wealth. ² Technology Vision 2021: Leaders Wanted Data Mesh: Domain Driven Data Ownership From Technology Debt to Technology Wealth 89% of executives believe that their organization’s ability to generate value will increasingly be based on the limitations and opportunities of their technology architecture Accenture x Starburst: Achieving Data Democratization with Data Mesh 7 Domain Driven Ownership Case Study: Data in Smart Tractors One example could be a tractor company for farming equipment. The company collects data on all the equipment they sell: • how the equipment is operating, • the warranties that it needs, • who is buying it, etc. These data points could have value towards the line of business that creates tractors. The team that owns the tractor division could use that data to see how well their products are selling or working, so they use that information to improve their product or service. The same data can benefit other parts of the business: • The service operations team could see an opportunity for more automation to develop smart tractors. • The data set could also inform on the improvement of warranties and predictive maintenance. • The data could make a business case to fund additional data sensors into the tractor, maybe funding a real time stream, so people could see live feeds of the tractor. Next, the smart tractor also could provide its own unique source of data, such as: • unique perspective of the soil quality, • a unique view of how the farmer actually works the land, and • what the farmer actually does. The technological possibilities are huge with the domain owner being organization that knows the data the best. The domain owner is charged with being able to evolve that data and incentivized to share among teams with differing goals. Empower the domain owners who can best evolve the data source with the opportunity to unlock that technological wealth Accenture x Starburst: Achieving Data Democratization with Data Mesh 8 Preface to Thinking About Data As a Product As you’re thinking about your role as a domain owner, if you’re thinking about having a data product, think about it just like you think about any other product that exists. Your data users are the customers of your data. Below are some questions that you might want to ask. Customer (desirability) Business (viability) Technology (feasability) Why Would a Customer Desire the Product? • Who is the target customer(s)? • What is the customer need? • Does this solution fill an existing need? • What is the customer value proposition? • Do we need to change customer behaviors? • Does the solution require additional costs for customer to adopt? Is the Experience Differentiated? • Does business model align with customer needs? • Does technology model reinforce the business model? • Do we have the competencies to realize the customer experience? What is the Business Viability of the Product? • What is the potential value / profit pool? • Does product require disrupting current value chain? • Is the goal of the product to experiment or expand market position? • How certain is the value creation? • Does monetization model support s-curve? • Does the profit model align with our business strategy? How is the Product Technically Feasible? • Is technology market ready? • Can technology be built to be cost competitive? • Is technology easily substituted? • How does current technology increase value of current portfolio? Teresa Tung Cloud First Chief Technologist, Accenture I’m excited to be invited by Starburst to work together to shape why you should and how you can pivot your data projects to Data Products based on my experiences working with clients at Accenture. “ Accenture x Starburst: Achieving Data Democratization with Data Mesh 9 Data as a Product The second Data Mesh principle consists of treating your data as a product. Once you start to look at your data through the lens of a product, you must ask yourself what you can truly accomplish. This is where you must think about reframing how you operate, collaborate, and innovate with those products for the sake of the business. Our research³ found that 87% of [healthcare] executives agree that digital twins, a digital representation of physical things, so whether it’s customers, assets, locations, these digital twins are becoming essential to their ability to collaborate in a strategic ecosystem partnership. ³Digital Health: Mirrored World Data Mesh: Data as Products Reimagine how you operate, collaborate, innovate 87% of executives agree digital twins are becoming essential to their organization’s ability to collaborate in strategic ecosystem partnerships Accenture x Starburst: Achieving Data Democratization with Data Mesh 10 Data as Products Case Study: Digital Twin The element of collaborating and seeing the same element begins with the digital twin, which might occur in the form of sharing models, artificial intelligence, or running simulations as a form of data. For example, you may have different data nodes with views across the business. • data from engineering and R&D , • data from supply chain , • data from manufacturing , and • data from service Unlock additional value of these data products as they mature. It starts just by everybody publishing this data allowing the business to see the same thing. What’s even better, access to the data — there’s value in that. For example: So once you have all this data, you have the simulations to use AI to really help predict the new design, and help you reimagine and create new products as well as new lines of business. Engineering and R&D Better designs that anticipate customer desires and minimize risk, reduces costs, and make sustainable choices. Better able to predict variability in demand and supply and factor this information into forecasting and ordering processes. Improved productivity ensuring materials accurately tracked and on time throughout the lifecycle. Supply chain Manufacturing Improved customer satisfaction from being able to dispatch repair technicians with the correct spare parts and skill-set to get back on-air. Service BENEFITS JUST BY SEEING DATA ACROSS THE BUSINESS: Accenture x Starburst: Achieving Data Democratization with Data Mesh 11 Evolution of the Digital Twin Digital twin is an example of how data products can evolve both in the completeness in how it’s connected and in the intelligence that uses it to drive insights. We can take this idea to the next level by mapping how these digital twins can mature and grow. You’ll see in the graph above, the data dimension and impact dimension scale together as the products enter each phase of visualization, predictive analytics, prescriptive analytics, and intelligent automation. The Data Dimension is where all the engineering design data, manufacturing data, financials, and field data would reside, with the objective of connecting them all together to act as a Rosetta Stone and point of reference for the development process. The AI Impact Dimension is that replicated data with the influence of machine learning and the addition of more intelligence. This ultimately accomplishes the task of providing complete visibility and insight towards where your data can take you. AI Impact Dimension: Increase the application and maturity of AI leveraging existing data to make the product development process smarter Data Dimension: Increase the type, amount, and quality of the data in the twin to better inform the product development process Data Visualization and Dashboard Predictive Analytics Prescriptive Analytics (Intelligent Coaching Intelligent Automation (Lights-out Operations) Field Data Financial/Logistics Data Manufacturing Data Design Data INCREASING APPLICATION OF DATA + + + + + + Legend Increased twin maturity Possible client journeys Accenture x Starburst: Achieving Data Democratization with Data Mesh 12 Self Service Infrastructure The third Data Mesh principle, self-service infrastructure, is focused on empowering the real expert on your team. The real expert should be someone who knows the business process, who knows the data and the domain the best. That is what democratizing data is about, enabling the expert to make the decision, model, or label the data in a way that’s easy for them without needing to understand the nuts and bolts of the complicated technologies. According to our findings⁴, 86% of executives agree that their organizations must train their people to think like technologists, to use and customize technology solutions at the individual level, but without the highly technical skills. While Data Mesh is a decentralized approach, empowering the experts is a move towards a spectrum of responsibility, with clear delineation and agreement upon the business domains and central IT organization. Self Service Infrastructure Case Study: Enabling Scientists One example of this approach is from the health and life sciences. This health sciences organization has scientists who are constantly running experiments, clinical trials, and conducting research. With so many data sets available to evaluate or generate experiments, they’ll need a simple way to access only what will apply to their own domain and workstream. When there’s new data, scientists should ideally be alerted, especially when there’s data from a domain the scientist might be interested in. There’s a necessary self-service balance between what the scientists and data product owners need. With self-service, access to insights is at our fingertips. Previously, because data integration was difficult, many domain owners didn’t have such quick access to insight. And oftentimes, the skills are still within that central organization. So the central organization can guide and assist users with: • Facilitating easy access to data for end users including providing visualizations, analytics models, experimental sandboxes, synthetic data • Providing data architecture and tooling to produce and use data products • Giving domain owners the checklist of things needed to publish a data product and acting as design authority • Giving metrics and methods to measure data product value • Providing common federated governance and discovery framework to manage federated data mesh The end result is that business users or scientists will understand it and be able to use it. Data Mesh: Self-Service Infastructure Empower the real expert 86% of executives agree their organization must train their people to think like technologists- to use and customize technology solutions at the individual level, but without highly technical skills ⁴ Technology Trends 2021 Accenture x Starburst: Achieving Data Democratization with Data Mesh 13 Federated Governance The fourth Data Mesh principle of federated governance aligns with the democratization goal of establishing your role in your ecosystem . This entails establishing the identity and value of your data in the greater business ecosystem as a measurable product. What we’ve seen is that this is strengthened by a network, with 90% of executives⁵ stating that multi-party systems will enable their ecosystems to forge more resilience and adaptable foundations to create new value within their organization’s partners Data Mesh really enables multi-party systems by having a programmatic means to track everything that you have. It means you can automatically check things like lineage, things like quality, and usage. And this is actually a rich place that you could add beyond role-based access control because now you have visibility. Data Mesh: Federated and Global Governance Establish your role in your ecosystems 90% of executives state that multiparty systems will enable their ecosystems to forge a more resilient and adaptable foundation to create new value with their organization’s partners ⁵ Technology Trends 2021 Accenture x Starburst: Achieving Data Democratization with Data Mesh 14 Federated Governance Case Study: Anti-Money Laundering Utility Money laundering is a problem that every financial institution must tackle. We helped a group of mid-sized banks tackle this in a cooperative model where they could share learnings and scale. Each member back already had their own data investments: different data platforms contained slightly different data schemas. We implemented a self-service infrastructure that started with methods for each bank to scan its data before securely bringing into the utility and models. Brought together the banks could share models and benchmarks. Moreover, the metadata programmatically tracked in this new network which could render lineage views from system of origin, to system of record, and to all the different models and user access. Adding metadata hooks enforced governance with a programmatic approach so the right information could be accessed at the right time on a global scale among all banks and on the individual level within each organization and team. Accenture x Starburst: Achieving Data Democratization with Data Mesh 15 Data Mesh is the Answer to Data Democratization To sum up, empowering domain owners to use right technology and process for the business need will unlock technical debt to technical wealth. Next, viewing data as products will reimagine how an organization will operate, innovation and collaborate where data is an asset. With a self-service platform , you’ll be able to empower the real experts to make data-driven decisions for the business. And finally, with end-to-end governance, you’ll finally establish your role in your ecosystem. As you’ve seen, the four pillars of Data Mesh contribute directly with an outcome of data democratization. It exemplifies the effectiveness of Data Mesh in innovating your digital transformation strategy with the goal of driving your business forward. Data Mesh Attributes • Decentralized ownership of data • Data products • Self-serve infastructure • Federated and global governance Data Democratization Outcomes • Unlock from Tech Debt to Tech Wealth • Reimagine how you Operate, Innovate, and Collaborate • Empower the Real Expert • Establish Your Role in Your Ecosystem As you’ve seen, the four pillars of Data Mesh contribute directly with an outcome of data democratization. It exemplifies the effectiveness of Data Mesh in innovating your digital transformation strategy with the goal of driving your business forward. Accenture x Starburst: Achieving Data Democratization with Data Mesh 16 Data Mesh from Concept to Execution with Accenture & Starburst In a conversation with SiliconANGLE ⁶ , Justin Borgman, Co-Founder and CEO of Starburst and Teresa Tung, Cloud First Chief Technologist at Accenture elaborated on the new emerging paradigm of Data Mesh. “Companies inherently acknowledge that data is decentralized and Data Mesh is a framework for thinking about that,” notes Borgman. “It not only acknowledges that reality, but also embraces it. That’s what’s driving the interest level in the Data Mesh paradigm.” “Many clients are already doing some aspect of the Data Mesh,” add Tung. “A lot of that is reviewing your existing data projects and looking at it from a data product [lens]. It’s a business strategy to think about the data, and there’s a new architecture that underlines and supports that.” ⁶ SiliconAngle: The Cube interview with Justin Borgman and Teresa Tung Accenture x Starburst: Achieving Data Democratization with Data Mesh 17 How Starburst Can Help Companies adopting a Data Mesh architecture must have an analytics engine capable of federating across these different data sources. Starburst is the analytics engine for the Data Mesh architecture, providing a single point of access to distributed data and empowering self-service analytics for each of the business domains. With Starburst, there’s no need to chase the idea of a single source of truth. Data is maintained by the domain owners but easily accessible in real-time across your organization. Starburst is built on open-source Trino, a distributed engine that can execute SQL queries against data stored in a range of databases and file systems. With Starburst and Trino, teams can lower the total cost of their infrastructure and analytics investments, prevent vendor lock-in, and use the existing tools that work for their business so that they can concentrate on enabling faster time-to-insights. Trino’s open technology means that integration with other open technologies such as data catalogs and data discovery tools is simpler and reduces the total cost of ownership of the self- service data platform. If you are moving towards adopting a Data Mesh architecture, we want to be there to help. To learn more, visit the Data Mesh Resource Center. Contact Us www.starburst.io/contact/ Accenture x Starburst: Achieving Data Democratization with Data Mesh 18 About Accenture Accenture is a global professional services company with leading capabilities in digital, cloud and security. Combining unmatched experience and specialized skills across more than 40 industries, we offer Strategy and Consulting, Interactive, Technology and Operations services — all powered by the world’s largest network of Advanced Technology and Intelligent Operations centers. Our 674,000 people deliver on the promise of technology and human ingenuity every day, serving clients in more than 120 countries. We embrace the power of change to create value and shared success for our clients, people, shareholders, partners and communities. Visit us at accenture.com. Copyright © 2022 Accenture. All rights reserved. Accenture, its logo, and High Performance Delivered are trademarks of Accenture. This document makes descriptive reference to trademarks that may be owned by others. The use of such trademarks herein is not an assertion of ownership of such trademarks by Accenture and is not intended to represent or imply the existence of an association between Accenture and the lawful owners of such trademarks. This document is produced by consultants at Accenture as general guidance. It is not intended to provide specific advice on your circumstances. If you require advice or further details on any matters referred to, please contact your Accenture representative. Accenture x Starburst: Achieving Data Democratization with Data Mesh 19 STARBURST.IO Copyright © 2022 Starburst