ESSENTIALS - Retrospectives: Shifting Towards Value - Tricia Broderick (Talk Outline)

Tricia Broderick (co-author of Lead Without Blame; founder of Ignite Insight + Innovation; 25+ years) on retrospectives that aren’t “communicate more → communicate more.” Framing: she was asked to do an “essentials” talk and thought “doesn’t everybody already know this?” — then attended a retro that was pure going-through-the-motions (”What did we do? Communicate. What are we going to do? More communicate. It’s painful. It’s not helpful.”). She loves retros — but hates the ritualized, valueless version.

1. Why Retros — Not Postmortems

• Poll: who runs the retrospectives, and who’s in the room because they’re tired of retrospectives? (You could answer both.) ”I don’t want to see this anymore” — walking in to “we just did a retrospective… it’s a sailboat activity.” (After co-writing with Diana Larsen, “all of her pet peeves have become my pet peeves.“)
• She loves retros but disliked “lessons learned” / “postmortems” — “the after-death meeting… who can we blame today?” You point the “magical finger” after the project’s over, don’t actually have the discussions, find something to work on — then ”project’s done, move along.” Not valuable.
• Larsen & Derby’s insight: why wait until the death of the project to reflect? As in the keynote, complexity is all about learning — constantly reflecting to understand how to go forward.
• Source: everything references Larsen & Derby’s Agile Retrospectives — now a v2 with David Horowitz (released ~last year; the previous edition was old but still relevant).

2. The Loop — A Retro Serves Value Delivery

• The v2 diagram: the retro (half the loop) exists to reinforce the product of value you’re delivering — not “it’s the end of the month, we do a retro.” “Could you do a retro just for the sake of a retro? Sure. But why?”
• The five stages are the right half of that loop — always in the context of the product/value, ”not because we’re a team and should do trust falls today.”
• Format note: she runs it as a slightly simulated mock retro (didn’t know the room/setup), pulling volunteers and shout-outs — a mini retro about the Craft conference as a whole (not individual sessions — no knocking anyone).

3. Stage 1 — Set the Stage

3.1 Get a vibe / honor where people are

• Goal: set the goal and motivate people. (“How many intentionally come 5 minutes late because meetings are so painful? …how many have a friend who does?“) If you don’t hook people up front, you spend the rest of the retro getting them off their phones and stopping the eye-rolls (“I had one developer continuously roll his eyes — we had words”).
• Quick opener: “Which duck are you today?” (a copyrighted image of duck moods). Seems stupid, but “I need to know where we are on day two at 9am.” There’s power in people just admitting it — everyone says their number at once (“nobody’s listening to you”). ”I don’t go to work happy every day. Nobody does.”
• Her own example: not “hangry” but angry + exhausted after a wonderful dinner that ran past bedtime (“get your bill, let’s go”). Knowing how people show up changes how you motivate them — if everyone’s a low number, skip the shout-out activities and do more thinking/quiet activities.
• This step is just a vibe — it does not set the goal or tone.

3.2 Focus the topic

• Ask the room: what happens if you don’t focus the topic? → people go all over the map (one on technical process, one on marketing, one dislikes lunch) → no chance of converging on anything.
• She hates “the entire sprint” as the focus (“How stupid — ‘what went well last sprint?’ → generic communication”). Instead: “let’s talk specifically — communication with our stakeholders, how’s that going?” Better energy, better motivation.
• Focusing also diminishes “that one person” with the same rant/soapbox every retro.

3.3 Be explicit about goal & decision

• State the goal: e.g. “1–2 highlights and 1–2 improvements.” Without this, people arrive with different expectations — she once saw a retro produce 20 action items for 4 weeks (everyone’s pet project got selected → nobody did anything). Narrow it, start slow.
• Be clear how the decision will be made — consensus / majority / input-then-leader-decides. If people think they’re helping decide and then learn they weren’t, they won’t participate next time (“frustrating”).
• Tip: always start with majority, not consensus — consensus never finishes / creates bad power dynamics. Get the ball moving first.

3.4 Motivate (make it their problem)

• The bullet list isn’t motivation. Ask the room why it’s good to give Craft feedback → “I’d like to come back and have it be even better”; “appreciation” — as a former conference organizer, “no one walks up and says thank you — they tell you how cold it is, how windy the tents are… did anyone like it?” Organizers “don’t know what they don’t know.”
• In a real retro on stakeholder communication, she’d start: “share examples where communication with stakeholders caused massive breakdowns” → people get fired up about the problems they felt → “wouldn’t it be nice to diminish that?” “I made it your problem — not me as the leader coming in to ‘improve you.’ I want your words driving the focus.”

4. Stage 2 — Gather Data (collaborative brainstorm)

• A variation of the Liberating Structure “1-2-4-All” (not exact, so unsourced): each person thinks alone of one highlight/appreciation and one big improvement → find a partner, discuss, come up with one together (could be one of yours or a brand-new idea) → merge with another group for another.
• Facilitation gem: “when I raise my hand, you raise your hand” — lets the speaker finish while quieting the room. “So effective that at a loud hotel-suite conference party, when security showed up, the guy answered the door and just raised his hand — and the security guard walked away.” (“Creepily raise your hand when security shows up; they’ll leave you alone.“)
• Why individual-first? gives time to think and safety (“if I thought of something dumb, I don’t want to say it to everybody; if I blanked, I don’t want to put up nothing”).
• Why pair, not shout to the group? it pre-empts collaboration over competition — “not your idea vs. my idea; I’m making you come up with one together.” By the time the hard part arrives you’ve already been collaborating, not competing.
• Why narrow to ones/twos rather than “collect all ideas”? 10 people × 3 ideas = too many to get through; narrowing shows where the energy is, collaboratively. “Nuanced, but it makes such a difference in whether you get through all five stages.” (Bad version: “throw up all your ideas and fight it out.“)

5. Stage 3 — Generate Insights (the step most teams skip)

5.1 Promote healthy conflict on an impact axis

• Target: promote conflict to get the ”groan moment” and explore ideas. Someone shouts a highlight (e.g. “the sessions”) → it goes into a bucket on a flip chart → “does anyone disagree that’s high impact?” — she wants the card to move.
• Why move it? you want someone to say “well, I actually don’t go to many sessions — I come to network, I spend my time with sponsors, that’s my genuine value.” Without that conversation you just assume “clearly sessions.” Note she uses “impact” not “did you like it” — “we can both like sessions but they have a different effect.”
• Expect more disagreement on improvements than on highlights (highlights may be compatible).

5.2 Dig into vague items; never name people

• Improvement example (risky, on Craft): “registration” left vague could mean many things. Her anecdote: she didn’t realize she got a speaker bag not a tote, the opening said “drink tickets are in the bag,” she had no bag, went back — ”you don’t get them until you get your speaker bag” — a real problem for her but not an all-attendee-impact problem. You must dig in to generate the insight, not just state a thing and walk away.
• Ground rule: the insight is never a person’s name. It can be a behavior or an issue — “the system always contributes.” Not “the negative is so-and-so at the registration desk” — that’s a private 1:1 conversation outside the retro, not a healthy collaborative one.

5.3 The elephant, and topic size

• When people won’t “groan,” she does something “obnoxious” learned from colleague Jake: put a cute little baby elephant on a chair — “the elephant’s here; anybody going to talk about it?” (You don’t want people throwing the elephant — Slido joke: sure, if mutually agreed as a speaking stick.)
• Don’t pick too-big a first topic — “let’s talk about psychological safety” as a first retro is inappropriate; the elephant’s too big to engage. Build up. She will run a whole retro on psychological safety once the team is collaborating, not just cooperating (per the keynote’s stages). (This is the step people skip, then “wonder why you’re going through the motions.“)

6. Stage 4 — Make the Decision

• Easy if you did stage three (options narrow from ~7 → 2–3 as discussion moves stickies). If you skipped stage three, “whoever’s most senior says ‘I think we should do this’ and everyone goes ‘okay’” — and this stage is very hard.
• Dot-voting annoys her (the common majority tool):
  • Anchoring bias — whichever gets the first vote tends to win; don’t let a powerful leader vote first.
  • Bandwagon — walking up with two choices, you’ll unconsciously pick the one with more dots even if you’d decided otherwise (“we like to be part of the winning group; it feels weird to be the only one voting down here”).
• Anonymous voting is worse (given the keynote): “the only way we can decide is by not telling each other how we feel” → not a high-performing, resilient learning team. (Sometimes you still need anonymity — take it for what it is, but it doesn’t promote teamwork.)
• Better alternative: number the options (1, 2) and have everyone write their vote before walking up — reduces the bias. With heavy power dynamics, avoid dot-voting/majority (leads to “teaming up”) — again, focus the topic to limit that.
• If the first three stages are done well, this stage takes 5–10 minutes.

7. Stage 5 — Close

• Not a summary — “nothing worse than an offsite where the last 30 minutes repeats the day; if you have to repeat it, it stunk.” Focus the close on next steps and celebration/appreciation.
• Next steps: “who’s going to take our action items back to Craft? When?” The facilitator need not own the follow-up, but make sure someone does.
• Celebrate the behavior, not the facilitator (“not ‘tell me I did a good job facilitating’”). Positive reinforcement of collaborating, engaging in conflict, and deciding together is huge — it’s what lets you take them to the next, deeper retro (bigger elephant): reinforce that “you’re still here, it’s still fine, it’s working.”
• Appreciation variation — “hopes and wishes” (from Diana Larsen): instead of general appreciations (which single people out — ”it’s harder for me to take a compliment than a criticism; I start panicking”), share hopes/wishes: “I hope we all come back to Craft next year”; “I wish the organizers understand the good intent behind this feedback.” Opens the spectrum without singling anyone out. Sometimes you’ll get ”I hope I never have to come to a retro again” — that’s okay, they needed to voice it.

8. Bonus Tips (whole-retro)

1. Focus on topics within the team’s control. A dev team couldn’t write marketing copy (a separate team did) yet complained about marketing every retro. ”What’s in our control?” → they made “marketing copy already present” a definition-of-ready → stopped pulling in not-ready work → work stopped stalling in the sprint (and ”guess who got us copy the next sprint — because we didn’t pull it in”).
2. Retros take real design/research/facilitation effort — there are tons of activities per stage (sometimes multiple).
3. Switch up activities — “if you did the sailboat four retros in a row, I’ll start mocking your drawing ability.” Different activities engage brains differently.
4. Duration fits the topic, not an arbitrary rule. Scrum framework says up to 3 hours; her first-retro rule: ≤30 minutes — you can do all five stages in 30 minutes. (She now runs multi-day retros because she knows how to facilitate them.)
5. Give the right to pass. A client insisted ”John will hate this and won’t participate.” She announced ”you have the right to pass, you can observe” — and because activities weren’t all large-group shout-outs, John participated all day, never actually passing. At the end: “I don’t know what you did today, but I liked it. I don’t know how I feel about you.” — “the best compliment.” The option to pass gave him back a feeling of control.

• Overall: go for the majority tipping point — enough people seeing the value shifts others toward it; if some still don’t come on board, “you probably have other problems with them that need addressing.”

9. Q&A

• Doesn’t focusing the topic let a biased scrum master define it instead of the team? She never said the scrum master focuses it — “the focus of the topic.” She likes putting up a board: “what should our retro focus on?” and letting people add topics.
• Teams that just whine — how to change the mindset? Come at their biggest problem — whatever they’re whining about, “let’s see what we can do about it,” from their circle of control. Even if it’s ”what’s served for lunch,” let them talk about it — they often don’t feel heard, and hyper-focusing on one thing blocks participation in anything else. (Hence the self-listed-problems board.)
• A lot of this needs in-person — remote tips? Every single thing can be done on an online whiteboard — collaboration, breakout rooms, all of it. (Susan has a whiteboarding session later; she won’t “pimp a tool.“)
• Isn’t focusing too narrowly problematic / risks missing real problems? Depends on how often you retro. She dislikes the generic “previous sprint” focus, but if you retro more than once a sprint (not everyone’s on sprints), sometimes ”the sprint as a whole” legitimately is the topic. Depends on frequency and length.
• Does “don’t single out people” apply to kudos sessions? No — kudos can single people out. But as a leader, be cognizant if the same person always gets kudos or someone never does — that may signal something in the team.
• Manager/team barely support monthly retros; time is a convenient cue — what other cue? End-of-sprint is a lovely cue (she just dislikes it tied to the sprint generically). On her Kanban/flow teams, once the board has a certain number of topics, that triggers a retro.
• How long should this format be? Hard in an hour. Depends — for ~80 people doing that specific simulation, ”probably a half-day” (more people = longer to get through the groan zone).
• Retros used as a vent — how to form concrete action points? If the groan zone and decision are set up well, it’s easy — a narrow topic yields action points straight out of the decision. It’s hard only when you don’t set it up.
• One bad-mood member destroys the retro — handle separately? If it’s a pattern (not a one-off), handle it individually from curiosity — what’s going on for them, what are their goals, are they a leader who isn’t leading, what’s their impact (“you don’t know what you don’t know about how you influence others”). (Offer: ask that and get one of her books.)
• How to accept things we truly can’t change? She challenges “we just have to accept it” — often there’s a slice in your control (marketing example; or performance reviews — she hates individual ones, found a way to do them within org constraints that aligned to her values). If it really isn’t in your control, get radically transparent about the impact and honor it.
• Nuances for small teams (~5)? Watch who talks the most (she’s “an extreme extrovert”). Groupthink can be worse in small groups (especially long-together teams) than large ones — plan activities to balance the power among the five.
• Hybrid retro — tool/method? A whiteboard tool (won’t name one “unless they pay me”); treat a hybrid retro like a fully-remote one — put the in-person people on the remote whiteboard too, even in the room.
• How to get someone active who’s never active? After the right to pass and adjusting activities (could be neurodiversity / cultural / safety dynamics), if still not participating, have a conversation (not “please participate” but “what’s going on — I know you’re amazing, what’s here for you?”). One person had a prior team lead who weaponized retro content to managers — “I am never speaking in a retro again”; that’s legitimate, and safety had to be rebuilt. (“Can I buy you a drink, that’s horrific.“)
• Closing appreciation (turned on her): “I wish we had more time to listen and learn from you.” — “I love it.”

People & References Cited

• Tricia Broderick — speaker; co-author of Lead Without Blame; founder, Ignite Insight + Innovation; self-described “extreme extrovert.”
• Diana Larsen & Esther Derby — Agile Retrospectives (the five stages; source of all the material); hopes-and-wishes close (Larsen).
• David Horowitz — co-author of Agile Retrospectives v2.
• Jake — colleague; the “elephant in the room” prop.
• Liberating Structures — “1-2-4-All” (the gather-data flow, adapted).
• Concepts: retro serves value delivery (the loop), five stages (set the stage / gather data / generate insights / decide / close), which-duck-are-you, focus the topic, majority-before-consensus, make-it-their-problem, impact axis, groan zone, “insight is never a person’s name,” elephant/topic-size, dot-voting anchoring & bandwagon bias, number-and-write-first, right to pass, definition-of-ready fix, majority tipping point.

Video: https://www.youtube.com/watch?v=LI8gMsriy3Q — Transcript via yt-transcript.sh; outline generated from the transcript.