Three ways to enable human-centered BI
In a world powered by more and more data, information alone isn’t a differentiator. By making data useful and actionable—creating a data culture that puts the power of data into the hands of “non-data” people—businesses are moving up the analytics maturity model. But many soon discover the democratization of data presents new and different challenges.
Business leaders today are faced with a proliferation of digital assets made possible by self-serve BI platforms. You likely know the drill…dozens of analysts publish their own version of the same or similar content further fragmenting the view into the business. These dashboards, crammed with graphs and charts, can have the opposite effect of their desired intent.
The sheer number of dashboards is only one issue. Even that is exacerbated by departmental data siloes and different interpretations of the data. Organizations quickly find themselves ironically uninformed. Their vision of data democracy and enhanced business insights have turned into a morass of information overload and analysis paralysis.
The days of dashboard slinging are over. With departments data rich but insights poor, an increasing number of organizations are saying stop the insanity. But where does that leave the organization that’s in the middle of digital decisioning transformation?
Enter human-centered BI, an approach we’ve developed to address some of the biggest challenges in a data project: defining and framing the ultimate analytics experience the users will have once the solution is built.
Design data tools, not dashboards
To truly make data actionable, you need more than charts and graphs—you need to create rich, interactive tools that tell a story and result in action. Think through the business process and the question, or the series of questions, you are trying to answer. You need to be able to tell a comprehensive view or story of what’s happening in the business and then enable employees to act on it.
That can mean creating custom apps that allow people to turn the data into action. For sales quota planning, for example, that could require enabling people to not only surface insights to know what to focus on but also giving them the ability within the same user interface to create a work back plan.
In our work with Microsoft, we utilized our human-centered BI approach to develop a partner planning tool that went beyond just displaying the data. While many of our clients assume a dashboard can solve their data challenges, often the underlying pain point demands more than just data visualization. So when we were asked to create a dashboard that would show both distributor sales and current inventory levels of products stocked, we had to dig a little deeper.
By leveraging human-centered design principles to ensure we solved the root problem, we inquired into how they would leverage the information and what decisions they would make with it. It turned out Microsoft wanted to use the dashboards to actively plan when to buy certain SKU’s with distributors throughout a period. Given these findings, we pivoted to developing a low-code application that allowed users to create an entire inventory plan for a distributor at the beginning of the period and manage to that through the quarter based on demand and inventory. By baking the data natively into this workflow, they were able to direct action and decision on the insights we surfaced.
The more you can move up the KPI stream to focus on leading rather than lagging indicators, the closer you are to operationalizing your data tools. Don’t just build a dashboard with the executive KPI du Jour. Understand the interdependencies behind the systems, the business processes and the people that rely on and contribute to the data.
Ask yourself and your customer, “What is the desired outcome, and what are the precursors to that outcome?” Ask how the end-user takes action today, and what other data they rely on to take those actions. As you continue to peel back the layers, you begin to discover an integrated system of leading processes, tool usage, data creation and rich communication.
One of the largest truck manufacturers in the world turned to us for help in optimizing their process fordemand forecasting. While their existing process was able to predict seasonality, it was unable to account for promotions their marketing teams ran, nor the introduction of new product and the resulting cannibalization of the old product. This was a multifaceted and interconnected business process leveraging data, but more importantly, contributing and communicating data. Of particular importance was to not just focus on a final KPI, like order fulfillment rates or inventory levels, but rather multiple stops along the way in a complex value chain where the real action was taking place.
In the end, the data tools that were created gave users direct visibility to the actions they could take in their piece of an integrated value chain. It also served a valuable capability that was completely missing in their legacy process: orchestrated workflow and handoffs. The level of visibility to each step of the process and the level of collaboration required put them in the position of targeting $10M in bottom-line contribution.
Use human-centered design principles
The principles of human-centered design differentiate our approach from simple data visualization. We focus on building rich, interactive experiences so everyone from analysts to executives can quickly find actionable insights. To achieve those rich, interactive experiences, you first have to get stakeholders in a room to talk about the desired outcomes and answer the simple question, “How can we help employees do their jobs better?” Translate your intention and connect at a human level with your employees, partners and customers.
At Union Pacific, we brought this to life by sketching out the dashboard in addition to specific data and functionality, then fostered a robust debate among senior-level stakeholders and the teams that would use the dashboard. This helped us understand users, the purpose of the information and how it would get used. Our iterative process – talk to users, sketch, get feedback and iterate—represent key human-centered design principles that helped ensure we were creating experiences that would be useful and deliver the expected outcomes.
Because modernized architecture is critical to a human-centered BI approach, our long-term partnership with Union Pacific positioned them to be able to act on the discovery we did in the above design sprints. The larger modernization strategy included shifting from on-prem to cloud, creating a microservices architecture and the ability to connect event data to other systems, including data warehouses and search indexes in real-time for analysis. Combined, these capabilities help form a foundation to a human-centered approach to BI.
“We will always do it this way from here on out.”
Matt Casselton, head of application development, Union Pacific
At the heart of the matter, human-centered BI is one part of a broader movement focused on employee experience. If you want your company to excel, employees need tools that help them be successful in their roles and drive the company forward. For organizations on the path to digital transformation, data is one way to get there, but only if teams are empowered to practice decisioning, and that requires the right kinds of data and decisioning experiences. Human-centered BI enables organizations to move from dashboards to tools that bring corporate scorecards to life, enabling people to reinvent how they solve challenges and drive value.
About the authors
Brad Jackson serves as vice president of Intelligence and Analytics at PK, where he designs data and analytics strategies for the world’s most recognized brands. His passion lies at the intersection of design thinking and analytics, where the most successful projects provide unique experiences, not just a database or a dashboard or a prediction. He has provided data and analytics solutions in the areas of sales, marketing, loyalty, finance, operations and supply chain management, for industries including technology and software, retail, manufacturing, media, transportation, food and beverage, gaming, financial services, education and healthcare.
James is a principal at PK and leads the Experience Design CoE. He brings more than two decades of experience helping brands understand and navigate the complexities of the digital and non-digital world to engage with users and achieve business success. James has extensive experience leading user-centered experience design and digital strategy engagements across retail, automotive, technology, and education. His work includes award-winning and business-shifting experience design work for both B2C and B2B brands, including DSW, Nike, Macys, Calphalon, Seventh Generation, Subaru, Toyota, Jeep, Union Pacific, Microsoft, T-Mobile, Jack Links, TopGolf and Vans.
Jared Dodson serves as sales enablement practice lead at PK. He has deep expertise in customer acquisition with a focus on B2B marketing and sales, and his passion is helping companies leverage technology to create personalized and relevant customer engagements at scale. Working with Fortune companies—such as Adobe, T-Mobile, Microsoft, Johnson Controls—Jared has led large-scale initiatives focused on lead generation & management, account-based marketing, B2B & partner marketing, marketing technology, sales process, and sales enablement. Jared’s experience spans industries including high-tech, software, telecommunications, energy, and insurance.Tags: Business Intelligence