Podcast Summary: The Analytics Power Hour, Episode #280

Title: Dashboards Must Die! Long Live Dashboards!
Guest: Andy Cotgreave
Air Date: September 16, 2025

Overview

This episode dives into the ever-controversial world of dashboards: are they dying, merely evolving, or as essential as ever? Joined by data visualization luminary Andy Cotgreave—co-author of The Big Book of Dashboards and Dashboards That Deliver—the hosts challenge the definition, purpose, and future of dashboards in business analytics, especially in the era of AI. Throughout, the group explores practical frameworks, actionable advice for dashboard creators, the importance of deprecating outdated dashboards, and the limits and promise of automation and AI.

Key Discussion Points & Insights

1. What Is a Dashboard? (03:14 – 08:15)

• Loosening Definitions:
  Andy emphasizes how pervasive the dashboard debate is:

  "It's such a vague term, we just reduced it to virtually nothing." (04:08, Andy)
  The working definition: A visual display of data that is used to monitor and facilitate understanding.

• Different Types depending on Audience & Function:
  There are broad definitions and uses—from high-level executive monitoring to granular, operational tools for frontline workers.

  "So we have a really loose definition of what a dashboard is... Very controversially, we can make an argument that even a single text table is a dashboard." (04:36, Andy)

• Taxonomy Choices:
  Though others (e.g., Nick Dobrat) have categorized up to 13 types of dashboards, Andy’s book opts for a flexible, use-case driven approach.

• Dashboards are not Always the Answer:
  The team notes that dashboards are often confused with other analytics artifacts—reports, applications—arguing that the semantics are less important than the purpose and outcome.

2. Performance Measurement vs. Operational Use (09:26 – 14:58)

• Performance Dashboards are "Easily Defensible":

  "Doing that well is to me easily, easily defensible. When people head off into other areas and start making broad proclamations, they're kind of setting up things that they can then say these are, it's bad for this." (11:29, Tim)

• Operational and Disaggregated Dashboards are Critical:
  Andy shares a Tableau sales dashboard ("Who's Hot Dashboard") designed to help sales reps act on individual lead activity—a fundamentally different but equally important use case from executive monitoring:

  "That is a dashboard that allows people to monitor disaggregated data... That's useless for the CRO. So you have that spectrum of dashboards that need to be included inside any data strategy." (13:30, Andy)

• The Pitfall of One-Size-Fits-All:
  Executives often assume building an aggregated dashboard solves all analytic needs, neglecting diverse users and their divergent requirements.

3. Interpretation & Actionability: How Much Should Dashboards Do? (16:21 – 23:14)

• Embedding Interpretation:
  Julie highlights the gap between surfacing data vs. surfacing meaning:

  "...they feel like if I just surface the data for them that all the interpretation and actual... synthesis of what those... mean will just happen and they'll know what to do. And obviously we know that's false." (16:21, Julie)

• The Challenge of Data Literacy:
  Andy affirms the importance of user-focused dashboard design:

  "Are all your users at the same level of data literacy? Or data fluency? Because that will change the way you design..." (18:45, Andy)

  His advice is to frame the dashboard's goals as "user stories," e.g.:

  "As an account executive, I need to see who's active on my website in order to make a high-chance call..." (18:53, Andy)

• Dashboards as One Tool Among Many:
  Not every business question justifies an always-on dashboard—some problems need one-off visualizations or narrative-driven reports.

4. Dashboards, Storytelling, & Narrative (23:14 – 24:49)

• Story-Finding, Not Storytelling:

  "Steve wrote a chapter in Dashboards That Deliver about... dashboards as story finding, not storytelling. And I think that's a really important distinction." (23:20, Andy)

• Dashboards can answer a few recurring questions, but truly persuasive narrative—especially for strategic decisions—often requires more tailored presentation outside of the dashboard medium.

5. The Dashboard Development Process (24:49 – 32:04)

Andy shares the framework from his new book:

• Spark: Identify the need/use case.
• Discovery & Prototyping: Talk to users, build wireframes (with or without real data).
  • On prototyping:

    "Wireframing is amazing... but the data drives the way your dashboard will look as well." (27:04, Andy)

• Development (Iterative): Build, test, refine with user involvement.
• User Acceptance & Training: Roll out, train users for adoption.
• Review/Deprecation: Plan for regular review and deprecation of dashboards.

  "Before you loop back... Do you still need it? Can you deprecate it? Because we don't deprecate enough dashboards, and many dashboards should die." (26:22, Andy)

6. Governance & Deprecation: Letting Dashboards Die (32:04 – 37:00)

• Planned Sunset:
  Start with an end date or review date for every new dashboard. If unused, delete or revisit.

  "At Tableau... they all get deprecated after two years if nobody's looking at it." (33:32, Andy)

• Usage Tracking > Self-Reporting:
  Monitoring actual views is more reliable than asking stakeholders; Andy and Tim share stories on discovering "zombie" dashboards that nobody needs.

• Announce Deprecation:
  Use visible warnings before deleting a dashboard, prompting actual users to come forward if it’s still needed.

7. Sharing Insights & Collaborative Learning (37:12 – 40:26)

• Arguing about Data: A Feature, not a Bug:
  Andy flips a complaint into a positive, emphasizing that robust, trusted data enables constructive dialog:

  "...what if that was actually a positive thing? What if you had a robust data reporting system... they could spend half an hour arguing about the data because then they're having a data-informed conversation rather than a data argument." (38:12, Andy)

• Leadership Techniques:
  One success: leaders who use dashboards to generate new, focused insights for their teams, rather than cycling through the same canned reports.

8. Dashboards and AI: Automation, Augmentation, or Hype? (41:47 – 54:17)

• Current Limitations:

  • AI-driven dashboards/tools (e.g., ThoughtSpot) offer infinite Q&A potential, but users want recurring quick answers, not to type the same queries every week.
  • AI still struggles with context, messy data plumbing, business logic, and the human nuance needed for trust and interpretation.

  "...I am, I'm lazy. If I go ask the same question every week or every month, I don't want to type that question every week...I just want to open my phone, go to Tableau Pulse, for example..." (42:37, Andy) "If we can't solve this [ambiguity] as humans, then generative AI isn't going to do that." (41:36, Andy)

• Historical Perspective:

  • The data industry's core challenges are nothing new:

    "Millions of dollars are spent every year collecting data with the assumption that having the data solves the problems being studied." (48:11, Andy, quoting a 1914 book)

• AI as Augmentation—not Automation:

  "Augmentation, not automation. I think that's a brilliant three word summary..." (50:50, Andy, referencing Donal Phipps)

  • Best use cases: accelerating prototyping, helping with coding, or data normalization. Full automation still can’t replace analyst context or business acumen.

• Opportunities and Risks:

  • Tools providing auto-generated narrative (e.g., Tableau Pulse) are promising but can become too generic. Generative AI explanations can be unreliable or hallucinate.
  • Trust, painstakingly built between analysts and business, can be lost instantly with misleading AI insights.

9. Purpose Before Dashboards: The Ultimate Takeaway (54:17 – 56:45)

• Dashboards, AI, and narrative tools all have a place—success depends on clearly defining the business problem and aligning the solution to user needs.
• The friction in data exploration—whether human or AI—remains critical for clarity, understanding, and impact.

Notable Quotes & Moments

• On Definitions:

  "I'm just not interested in... semantics of what we're doing. It's the end goal that is." (18:02, Andy)

• On Dashboard Projects:

  "As a data analyst... we have access and use the systems to visually explore data... But... a dashboard is a window on a predefined, finite set of questions." (06:21, Andy)

• On Storytelling:

  "Dashboards are story-finding, not storytelling." (23:20, Andy, referencing Steve Wexler)

• On AI Limits:

  "...in all my experience of trying to use AI to do data projects, I've been left wholly underwhelmed and really frustrated and just thought I could have just done this quicker myself." (44:41, Andy)

• Historic Perspective:

  "It's from... Graphic Methods for Presenting Facts by Willard Cope Brinton, printed in 1914, 110 years ago... we haven't solved the problem in 110 years." (48:31, Andy)

• On Trust:

  "It takes minutes to lose that trust. If we give them tools that say sales went up when they went down, that's a terrible... indictment on what we do." (53:03, Andy, quoting Tableau Tin)

• On Dashboard Governance:

  "Many dashboards should die. I'm not here to say they should all live." (26:28, Andy)

Timestamps for Major Topics

• Definition Debate: 03:14 – 08:15
• Performance vs. Operational Dashboards: 09:26 – 14:58
• Interpretation & User Literacy: 16:21 – 23:14
• Storytelling vs. Story-Finding: 23:14 – 24:49
• Dashboard Development Framework: 24:49 – 32:04
• Deprecating Dashboards: 32:04 – 37:00
• Sharing Learnings & Team Use: 37:12 – 40:26
• AI & the Future of Dashboards: 41:47 – 54:17

Memorable Moments

• Andy’s Etymology Lesson: The origin of "dashboard"—from leather shields on stagecoaches, not technology. (05:11)
• Fabled Internal Tableau Dashboard (Who's Hot): How operational dashboards can make or break sales workflows. (13:30)
• Wireframing with Real Data: The pitfalls of prototyping beautiful designs that fall apart when populated with messier reality. (27:04)
• AI Hype Check: Generative AI tools still can’t replace human judgment or messy data reconciliation (42:45, 44:41).
• Historic Quote Drop: A 1914 book reveals the data industry's century-long struggle. (48:31)

Tone and Style

The episode blends sharp industry insights, practical wisdom, and witty banter. The hosts and guest freely challenge each other’s assumptions, share war stories, and keep the tone conversational—even as they dive deep into theory and practice. Explicit language, self-deprecating humor, and anecdotes keep the content lively and accessible.

Final Takeaways

• Dashboards are evolving, not dying—their value is contingent on clear purpose, user-centered design, and context-aware implementation.
• AI and automation augment dashboard creation and analysis but can't yet replace (and may never fully replicate) human judgment, especially in interpreting data nuance and business need.
• The need to deprecate unused dashboards and the challenge of aligning artifacts to real decision-making is perennial—and still unsolved after a century of data work.
• Keep the focus on business questions, not on technology or definitions.

"Dashboards must die"—and be reborn, better aligned to what people actually need.

For Further Exploration

• Dashboards That Deliver by Andy Cotgreave (2025)
• The Big Book of Dashboards (Andy Cotgreave, Steve Wexler, Jeffrey Shaffer)
• "Graphic Methods for Presenting Facts" by Willard Cope Brinton (1914)
• Articles, workshops, and taxonomy resources from Nick Dobrat, Steve Wexler, Amanda McCulloch

Contact the hosts:
LinkedIn, Measure Slack group, or contactanalyticshour IO

Listen for: Further episodes on data storytelling, AI for analytics, and practical governance.