I have always loved diagnostic analysis. In business, we’re often brought in to unravel complex questions, following every data string to find rational explanations for “why” something is happening. By the time the data team is consulted, the return on investment (ROI) is typically immense, and the answers reveal fascinating insights, often translating into compelling data stories. For example, in a recent project, we analyzed two groups of users. Group A completed an application, while Group B did not, despite looking demographically identical. We concluded that a user interface issue was preventing Group B from finishing the process. (Click here for the case study.) This insight delivered an annual ROI of $35M. However, while diagnostic projects are successful, they often come with a significant challenge: realizing the full ROI depends on implementing the outcomes, which may require additional investment, like funding to fix a broken user interface.
Why does this matter? And how does it relate to creating a tipping point?
In business, “stickiness” refers to the ability of a product or idea to capture and hold attention. For example, smartphones are highly sticky products: in June 2024, Exploding Topics reported that people spend nearly four hours daily on their phones, with over two hours on social media alone. This “sticky times sticky” dynamic generates highly profitable engagement. In contrast, analytics projects often struggle with stickiness. Their insights, while valuable, don’t tend to linger because they aren’t inherently engaging.
Analytics projects typically deliver outcomes in three forms: Information, Automation, and Visualization.
Three months after briefing a Senior Vice President (SVP) and team on a user interface issue, I had a conversation with one of the SVP’s top aides. She was exploring ways to shorten the application process. Given our findings, I explained that shortening it was unnecessary—the interface, not the length, was the barrier. Data revealed that only 4% of users exited the application due to reasons unrelated to the interface. Yet despite ongoing efforts to address the UI, she had forgotten this crucial insight. This scenario underscores a common challenge: analytics often clash with preconceived notions, leading stakeholders to overlook or forget new information.
In parallel to the UI project, we also developed a model to auto-adjudicate coverage applications, aiming to streamline decision-making. While 80% of applications were typically approved, the model auto-approved 70%, freeing staff to focus on the remaining 30% of complex cases. While theoretically useful, the model didn’t feel like true automation to the team, as it simply reallocated rather than eliminated work. Even automation projects with clear value tend to transform tasks rather than replace them outright, meaning realized ROI can be underwhelming.
Dashboards are the classic visualization tool, but they often lack engagement. Typically, dashboards either display metrics that rarely change or show high-variance data. Most are effective for anomaly detection—alerting users when something has gone wrong—but rarely provide ongoing, sticky engagement. When stakeholders view dashboards, they either see familiar data or data that signals a problem. In both cases, this interaction is passive, reinforcing the “anti-sticky” nature of analytics.
Analytics struggles to become sticky because it is inherently non-interactive. The insights generated often aren’t easily seen, understood, or valued in a way that fosters retention. If stakeholders can’t feel the value of analytics intuitively, they require strong incentives to drive engagement and adoption. Ensuring that analytics achieves a tipping point—where it becomes a core part of strategic decision-making—requires more than delivering insights; it requires actively aligning analytics projects with business actions and making insights feel tangible and immediately valuable to those who use them.
Explore the implications of integrating generative AI into data strategy, including its impact on business objectives, accessibility of new use cases, and ethical considerations. Learn how to align data strategy with business goals, unlock innovative use cases, and balance innovation with responsibility in the generative AI era.
Explore the parallels between the rise of generative AI and the earlier era of data democratization in the workplace. Learn from past challenges in data management, governance, and ethics and apply these lessons to effectively integrate generative AI.
Explore the advantages and disadvantages of empowering employees during change management. Analyze the five steps of the Engagement vs. Empowerment heuristic to help you make informed decisions about when and how to empower employees in change initiatives. Find the right balance for effective leadership.
