Data Driven Design for Enterprise Organizations
- Start with a data design framework to understand users and data
- Define how data will be used, next best actions on that data, and why it matters
- Data products must produce insights and actionable knowledge
McKinsey predicts that By 2025:
Nearly all employees naturally and regularly leverage data to support their work. Rather than defaulting to solving problems by developing lengthy—sometimes multiyear—road maps, they’re empowered to ask how innovative data techniques could resolve challenges in hours, days or weeks. Source: McKinsey
The proliferation of siloed data over the last several years ushered in an era of graph databases, artificial intelligence (AI), and machine learning (ML) as a method for companies to connect data in the hopes of gaining new business intelligence, insights and competitive advantages. Many enterprise organizations invested millions of dollars in pilot projects on the promise that these technologies would deliver returns at-scale once data was connected at-scale. Some organizations who took this leap are now left wondering where that return is and if they can still reap the benefits they expected. This challenge is going to only grow in scope as we see enterprises shifting from pilot projects to real-time data across verticals.
Predictive UX solves this challenge with our Data Design Framework that combines Design Thinking and Data-driven Design in our approach to data applications.
Instead of first amassing data at-scale, our approach asks organizations to first seek to understand what data they have and what business and user needs could be supported by exploring and connecting data assets further.
With Design Thinking and Data-driven Design the needs of the business and users are first defined with data workshops, personas and journey maps. This allows the journey of the people who are using data to become the foundation of graph database, machine learning and AI projects from both a qualitative and quantitative perspective. Leveraging this approach puts companies on track to reap the expected return on their investment in data because they are first thinking about how data will be used and what actions people need to take on data. The result is an integrated data experience that enables knowledge workers to take action to protect and grow revenue. This approach scales well with real-time data as we consider the future of work where teams are asked to make decisions in near real-time. If organizations invest up front in understanding those moments and what workers need to be effective in a fast-paced data intensive environment, their data solutions can be designed to meet those needs.
As a user experience designer who has worked with companies developing data-driven applications, I have seen many who have skipped or shortened this critical step even though it is the one that causes so many projects to fail. Human-centered and Data-driven design changes that by ensuring that the domain owners, data scientists, engineers and designers of distributed data platforms are connected to an understanding of data from two perspectives: the perspective of experts who work most closely with the data and from the perspective of those who need to consume and use the data — the latter are the end users who most often interact with data through various UIs.
When data systems start with the outcomes and end user needs in mind, the domain model takes shape first alongside the users and domain experts which in turn informs data relationships (ontologies), machine learning, taxonomies (controlled vocabularies), human-in-the-loop flows, metadata, data flows, user experiences, UI's, and governance.
Data Design Framework
Predictive UX has created a Data Design Framework that focuses on the following key principles to drive organizational value. Our approach includes:
- Design thinking and data-driven design to uncover user needs, assets and asset value aligned to business and user needs
- Defining knowledge and data architectures that allow businesses greater, direct access to knowledge and data
- Creating systems with micro UIs and data-rich component libraries to reduce design and development time
- Admin UIs and workflows to drive human-in-the-loop knowledge back into data layers
- Decentralized governance through people, process and technology, modeled after content management and API management best practices
From this model, organizations can:
- Define metric-driven ROI use cases that leverage quick access to data and knowledge assets through graphs, taxonomies, ontologies and UIs to drive business value
- Create a knowledge and data microservice architecture with APIs to underpin various bespoke UIs
- Design interchangeable enterprise data components based on a company-wide design system to drive consistency and ease of use
- Establish a distributed governance model that decentralizes data management with accountability for data insights and data accuracy at the source of the data
The outcome is a sophisticated yet straight-forward enterprise data strategy that is self-governed to improve accuracy, a primary source of concern for consumers of data, while also providing a distributed set of building blocks that work together to enable rapid application and data deployments at the functional level for just-in-time insights.
To achieve this reality, companies need to look no further than how they have traditionally approached content management projects where SME’s from each business unit serve as experts to define, author and govern content. The difference now will be that content and data will both be managed at functional levels within an organization. This provides a natural, low barrier approach to enabling data at-scale while also empowering content SME’s to own data alongside data SME’s with new analytical tools from which to manage and develop insights, put human intelligence back into the flow, and to track data pipelines.
Content managers, product managers and marketers are poised to become data power brokers and will need to work with others beyond their own domains to socialize insights and steer the enterprise data domain at-large alongside their data science peers.
Real Use Case
A large SaaS organization with more than 150,000 subscribers across their product suite needed to know customer feedback and sentiment across all apps so that product teams could understand the impact of any changes to shared components across the product line.
With plenty of data but no idea how to put it all together for meaningful, actionable insights, Predictive UX stepped in to conduct research, design a data-insights dashboard and conduct user testing to inform and iterate a high-value data design.
The solution is powered by graph technology, but it didn’t begin with the data. It began by first understanding user needs, business needs and ROI, and creating and testing with real data and real users to determine the right fit between data and the user experience. If this project had started with connecting all of the possible data first and developing insights using AI, the outcome would have been a long, high-cost project that resulted in a data lake or another data silo but that didn’t first consider how users or the business would use the data or if users or systems even needed all of the available data.
The human experience needs to come first.
To achieve this, we created a data-rich customer ecosystem prototype that surfaces customer sentiment and trends while also providing a method for product managers to drill down to the most granular level of feedback. This allowed different product teams to see, interact with, and discuss customer feedback and the impact around proposed changes.
To ensure the experience worked in this fast-paced environment of highly connected teams, we first normalized data from disparate sources and used nomenclature familiar to users to describe information and actions. Next, we designed dashboards and underlying screens to surface actionable insights to connect people to data for real-time decision support, and then we tested over and over again. The outcome is a data-rich, highly usable, use case centered dashboard that engages data users in meaningful discussion.
Data-at-scale isn’t necessarily costly or bad. However, it is both of those things if an organization spends time and money developing a sophisticated data architecture with machine learning, algorithms and edge relationships only to later learn they don’t know how to realize business value after all of that time, cost and effort.
Connecting data for the sake of the data journey is a costly endeavor. Connecting assets and data for the sake of human needs to create or protect revenue is a much smarter and, ultimately, more scalable and valuable approach for the data-driven organization.
The use cases for connecting content and data-at-scale are there and the ROI is also there if you begin with the user experience of data in mind.
Our data experts:
Predictive UX specializes in humanizing data-driven design for startups and enterprise organizations. Talk to our Data Practice Lead to learn more.