Mind Your P’s and Queues - Daniel Vacanti (Talk Outline)

Daniel Vacanti (from Miami; agile ambassador, three books on agile/Kanban, “Craft’s Stanley Tucci”) — golden nugget: be curious and do your homework. The English idiom “mind your P’s and Q’s” means watch what you’re doing — the recurring theme: agile obsesses over Prioritization while neglecting the thing that actually governs outcomes, your Queues (work in progress).

1. Framing — English, and the P-words

• Opening bit on English being a hard language (he attempts five words of Hungarian). “Mind your P’s and Q’s” = behave / watch what you’re doing (origins: P/Q confusion in movable type; or pints and quarts in pubs).
• Agile is full of P-words — portfolios, projects/products, prioritization, planning, predictability: “you prioritize so you can plan, so you can be predictable.” In the interest of limiting WIP, he picks one: prioritization.

2. What Problem Does Prioritization Solve?

• Frameworks mandate it: Scrum requires an ordered product backlog; SAFe uses WSJF (“the less said about SAFe’s weighted-shortest-job-first, the better”).
• The real purpose (per the Scrum Guide): maximize value (maximize, not optimize). You order items because you believe some are more valuable (e.g., “bank setup” #1 over “model my pay”).
• The catch: value is only realized when an item is delivered to the customer → what we truly care about is the order in which items finish. But priority determines the order in which items start. That’s backward.
• Two unverified assumptions: items that start sooner finish sooner, and items finish roughly in start order. “Has anyone ever gone and matched finish order to start order?” (In every org he’s seen: no.)

3. Queues Are Everywhere

• What actually influences when items finish is how many things you’re working on at once — queue size / WIP — which no agile framework talks about.
• Everything is a queue: board columns (up-next, analyzing, creating, validating) and every level of a work-breakdown structure (initiative → feature → team boards) — “queues coming out your R’s.”
• The WIP explosion: with features A–E each broken into stories, a team working stories in priority order pulls story 1 of A, then B, then C, D, E — quietly pulling all five features into progress. It looks disciplined at the story level, but multiply by an initiative board and 10/20/50 teams → a massive explosion of WIP across the org.

4. Monte Carlo & How WIP Scrambles Priority

• Customers only ask: “When will it be done?” — or, for a fixed date, “How many of these features will make it?”
• With a known number of stories, a Monte Carlo simulation can assign each feature a probability of finishing by a date — but it makes two hidden assumptions: the team works on one thing at a time and in strict priority order (neither is ever true).
• Re-modeling with more WIP (18 features):
  • WIP = 2: already funky — priority #9 now has a lower chance than #10.
  • WIP = 9: priority #1 has a lower chance of finishing than #2 through #15.
  • Pull everything into progress: ”stories not assigned to a feature” beats #1 “bank setup”; nine features have a higher chance than #1. A product owner’s carefully-built prioritization is “out the window.”
• As WIP grows (especially org-wide), finish order is dictated by WIP, not priority — everything is at risk, so ”prioritization is waste.” (Extra jab: ”WSJF is a real thing; SAFe’s implementation of WSJF is not — happy to argue it in Q&A.“)

5. The Snowball & Takeaway

• If WIP wrecks priority, why plan? If you can’t plan, you can’t be predictable; if you can’t be predictable, you can’t decide which products to invest in across the portfolio. Every agile P-word comes back to controlling queue size / WIP.
• (Joke: 2025 law requires an AI talk — his next: “You can’t spell Daniel/Vacanti — or Stanley Tucci — without AI.“)

6. Q&A

• Old agile vs. new agile? The Manifesto (2001) was very software-centric and “got a lot of stuff wrong” — no language for managing WIP or flow; worst perversion: Scrum/SAFe became synonymous with agile. (Plug: his talk ”The New Agile Economics”.)
• Should we only work on one thing at a time? No — he’s not saying that. Find your optimal organizational capacity (maybe 1, maybe 2 — “I guarantee it’s not 18”); if you ignore capacity and blow out WIP, finish order is dictated by WIP, not priority.
• Ideal dev-team size to limit parallel stories? There isn’t one — “seven ± two”/two-pizza teams “don’t exist.” He ran a 60-person standup (finished under 15 min); a colleague had a 40-person team. The limit is how many things you work on, not headcount.
• Team picked random stories ignoring priority — a team-level problem? Clarification: the team did work in story priority (story 1 of A, story 1 of B…) and thought they were doing right — but starting one story starts the whole feature.
• Book on queuing theory? No single great one for agile; read John Little (Little’s Law) for queuing, Shewhart/Deming/Wheeler for variation.
• Which of your books first? When Will It Be Done? (forecasting) → Actionable Agile Metrics for Predictability (why forecasts work) → the third to go further.
• Switch to Lean (where limiting WIP is core)? Not mutually exclusive — do both (“conferences used to be called lean-agile”; “there is no agile without lean”).

People & References Cited

• Daniel Vacanti — speaker; books Actionable Agile Metrics for Predictability, When Will It Be Done?, and a third; talk The New Agile Economics.
• Scrum Guide (ordered backlog), SAFe / WSJF (criticized), Monte Carlo simulation, Little’s Law / John Little, Shewhart / Deming / Wheeler (variation).
• Concepts: prioritization vs. WIP/queue size, start order vs. finish order, WIP explosion, flow, lean-agile.

Video: https://www.youtube.com/watch?v=Q5D2odLVEAY — Transcript via yt-transcript.sh; outline generated from the transcript.