The Analytics Power Hour: Episode #279

Title: The Process(es) of Analytics (We Have Thoughts)
Date: September 2, 2025
Hosts: Michael Helbling, Moe Kiss, Tim Wilson, Val Kroll, Julie Hoyer

Brief Overview

This episode explores the multifaceted world of process in analytics: when it helps, when it hinders, and how to find the elusive “just enough” process for effective work. The hosts draw from consulting, in-house, and leadership experiences to discuss process extremes (too much/bureaucratic vs. too little/chaotic), real-world frameworks for prioritizing analytics tasks, the psychological side of communicating and scaling process across teams, and the vital (but tricky) connection between process, expectations, and outcomes. Along the way, the hosts spotlight the pitfalls of rigid ticketing, the necessity of adaptability, and the virtue of transparent assumptions—with plenty of sharp, funny, and self-deprecating banter.

Key Discussion Points & Insights

1. The Spectrum of Process: Too Much, Too Little, and Just Right

• Opening Framing: Process can both empower and bog down analytics teams – the challenge is finding balance.
• Julie’s Perspective (02:57): Every client has their own process, and extremes both cause problems (overly automated slows things down and reduces relationship-building; no process means chaos and confusion).
• Concrete Example (04:32): Julie describes clients with no clear communication or stakeholder alignment, resulting in “chaotic random emails” and confusion over priorities.
• Ticketing Systems Critique (07:02):
  • Julie: Deep dislike for replacing conversation with asynchronous ticket queues (“If they're like, ‘We want to work with you and Jira, we'll just send you tickets,’ I'm like, get me off this flipping project, I'm done.”)
  • Moe: Draws a line between using tickets as project management versus as a stakeholder engagement model (“I violently hate this whole like service desk mentality. I believe in data folks being business partners…not submitting tickets.” 09:35)
• Key Maxim:
  • Michael: “Just enough process and no more.” (07:30)
  • But, as Moe points out: “How do you know you’ve hit that?” (07:40)

2. Why Process Bloats & Why It Breaks (06:00—11:00)

• Tim: Process often swells to “plug every leak” after a one-off issue, leading to bureaucratic burdens: “Now you’ve removed judgment because you said, ‘we had one issue one time, let's now inject more structure into the process.’” (06:05)
• Personality Factors: Mo notes some thrive in “non-process-y” environments; Michael observes this is highly personality driven—what feels chaotic to one may feel like freedom to another.

3. Task Management, Process, and Stakeholder Engagement (11:12—14:00)

• Val’s Journey: Admits to once imposing tickets on her analytics team, realizing later she craved not tickets but clarity (the “analysis planning document”) that prompts stakeholders to seriously consider and communicate what they want.
• Critical Distinction: Tools (like JIRA) aren’t inherently bad—but confusing ticketing with stakeholder engagement is:

  “You can have a ticketing system, but don’t use that as the means of communication ... there should be layers to this.” – Michael (14:00)

4. Making Process Work for Teams: Frameworks, Prioritization, and Communication

Moe’s In-Depth Example: Planning at Scale (18:10—20:53)

• At Canva, Moe’s team planned all analytics work for H2: organized by themes, ruthlessly prioritized, and negotiated with stakeholders (“Anytime you come to us in H2...we're going to ask you to cut something else from it.” 18:53).
• Key Principle: Data as partnership, not a service desk: prioritization with, not for, stakeholders.

Scoring and Frameworks (24:32—26:03)

• Moe’s team developed a scoring framework to evaluate incoming work:
  • Dimensions:
    • Strategic impact (e.g., CEO-requested = 5)
    • Multiplying effects (serves one or many)
    • DS uniqueness (can only a data scientist do this?)
  • “If you feel like this task is like sitting at the ones, you shouldn't be doing it.” – Moe (25:43)

Adapting Process to Scale and Context

• Michael (38:44): “Adaptability is the key … if I brought this heavy handed, multi-step process to a smaller, more nimble team, they would just be like, ‘Why are you wasting our time?’”
• Val (39:13): In small orgs, “so much more can just be an unwritten contract.”
• In large orgs, process is needed to manage complexity, reporting, and compliance.

5. Scoping, Assumptions, and Facing Surprises (27:24—54:09)

• Scoping Dilemmas: Even with strong processes, scope will often blow up mid-project.
• “Communicate, communicate early and often and say I’m sorry we missed on this. What should we do?” – Tim (45:44)
• The Power of Documenting Assumptions:
  • Julie describes adding explicit assumptions sections (“a soft contract with the client” 47:49), making change management smoother when surprises arise.
• Moe: In-house teams stress-test assumptions, sometimes informally, often consulting with finance to validate business contexts and calculations (50:47, 53:21).
• Key point: Assumptions should be externalized, checked, and shared.

6. Process, Communication, and Knowledge Transfer

• Scaling Process Beyond the Individual (31:52—34:56):
  • Michael: “A lot of times people like me…don’t externalize my process. My process is very intuitive….”
  • Moe: Having the process in your head isn’t enough when leading a team—need to “connect the dots” for others, not just yourself.
  • “You can't have all that stuff floating around in your head… even if that process works for you, it's not working for my team.” – Moe (32:43)

7. Over-generalized vs. Tailored Processes (35:22—38:57)

• Julie: The risk of trying to generalize process for everyone is overburden (“the tipping point of where is it too burdensome because we've over generalized…” 35:22)
• The best processes provide predictability and expectations without losing the ability to adapt and problem-solve.

8. Process and the Coming of AI (55:40—57:16)

• Michael: “Everything we've just talked about matters even more now because of AI.” (55:40)
• Moe: “AI is really good at process. That is one thing it's very good at.” (55:58)
• Tim: “No, it’s always fucking mattered. And now if…people want to say, well, now with AI, you’ve got to do it more. It has always mattered.” (56:10)
• Key Takeaway: Whether for humans or machines, clear, thoughtful process is foundational.

Notable Quotes & Memorable Moments

• Michael: “Just enough process and no more.”
  • [07:30]
• Julie: “You have to have a lot of technology and tools and logins and interfaces … and it slows us way down. … We lose the ability to have a personal relationship with them.”
  • [03:20]
• Moe: “I violently hate this whole like service desk mentality. I believe in like data folks being business partners and I do not think submitting tickets actually is the right way of working…”
  • [09:35]
• Tim: “Process for the sake of process in many organizations, especially larger enterprises, wind up in that world … She didn’t care if she delivered business value ever, as long as she followed the umpteen gazillion steps.”
  • [30:08]
• Michael: “Anytime you need to transfer something from one person to another...that’s when you need a process.”
  • [33:57]
• Julie: “Listing [assumptions] out ... it’s a soft contract almost with the client ... And if they come with last minute requests … if you really want to change the scope, then we have the right to change the deadlines.”
  • [47:49]
• Moe: “Apparently doing process stuff is good.”
  • [20:53]
• Tim: “Communicate, communicate early and often and say I’m sorry we missed on this. What should we do? ... as opposed to trying to hide it.”
  • [45:44]
• Michael (on AI): “It forces you to extrapolate your process out more cleanly if you actually expect the AI to perform the way you want it to.”
  • [56:54]

Timestamps for Key Segments

• 02:57 – Julie’s opening observations about process extremes in client work
• 04:32 – What “no process” looks like in reality
• 06:05–11:00 – How and why process balloons (including ticketing systems)
• 18:10–20:53 – Moe’s example of team-wide, stakeholder-focused planning and brutal prioritization
• 24:32–26:03 – Moe’s scoring framework for analytics work
• 31:52–34:56 – Why process must be externalized to scale beyond oneself
• 38:40–41:15 – Adaptability and scaling process for org size and context
• 45:44 – What to do when scoping or assumptions go wrong
• 47:49–53:21 – The power and pitfalls of documenting assumptions
• 55:40–57:16 – The importance of process in the era of AI

Episode Takeaways

• The “right” process is always context-dependent—and almost always less (and more flexible) than corporations think.
• Communication, clarity, and expectation management are just as essential as steps, systems, or documentation.
• Process for its own sake is a trap; process that reflects—and serves—real needs and complexity is invaluable.
• Assumptions may seem obvious but must be externalized; otherwise, scope blows up and relationships strain.
• As analytics teams face increasing complexity (and AI disruption), the intentional, well-communicated, and transparent process is a competitive advantage.

Interested in implementing or refining process in your analytics work? Check the show notes for the frameworks discussed, like Moe’s prioritization rubric and Val’s analysis planning document.