Ship It Conversations: Running Retros as Tabletop RPGs, With Jinger Meilani of Mountain on Chapters and Characters, Technical Archetypes, Reading the Room, Following Through, and Why Silliness Gets You Honest Answers

The thing that stuck with me most from this conversation with Jinger Meilani is that a Tabletop Retro is still a retro.

That sounds like an odd takeaway from an episode with orcs, character sheets and a campaign that opens with everyone imprisoned in the Hall of the Mountain King.

But Jinger was clear about it. She doesn't see it as a replacement for the standard retro. In her words, "I would call it retro plus plus plus."

Everything a normal retro is supposed to produce is still the foundation. What did we do? What do we want to do better? What's the posture of the team? Where are we aligned, and where aren't we?

The story is the delivery mechanism. She compares it to hiding vegetables in a fruit smoothie. The silliness disarms people, and the honest answers go down with it.

I appreciated how honest she was about why she built it in the first place: "it's because I was bored of them." Even the themed retros got rote. On a remote call with cameras off, she'd tune out. Sometimes she couldn't remember what she did yesterday, much less what she did two weeks ago.

Her critique of the standard format is that it asks people to introspect without helping them do it. As she put it, the standard retro basically says, "let's attempt to introspect, but I will not set the scene for you."

That one landed for me, because it's the same reason a good incident review starts with the timeline. Nobody remembers the order things happened in at 2 AM. You rebuild the scene first, and then you ask what went wrong.

Jinger builds the scene with a storybook. Pictures, music, her own drawings, pages turning on a shared screen. Then, at the plot point that maps to something the team actually went through, she hits them with an activity. Sometimes it's a typical retro activity, what went well and what we want to change, the kind of thing that would feel dry anywhere else. People are so tuned in to how weird the whole thing is that they don't notice.

The orc line is the part people will remember, and I think it's worth slowing down on.

Some people get a character and feel emboldened. They'll say the thing they'd never put on a retro board, because "It's not really how I feel. I'm just an orc."

That's a layer of indirection. It's the same move a blameless postmortem makes when it talks about the system instead of the person. The character gives someone a way to say something true without owning it as a complaint. It's the same thing she said about the story's mappings: sometimes everyone thinks "we're talking about one thing when secretly we're talking about another thing."

But she was just as clear that it cuts the other way. For some people, role-play is the intimidating part.

So she hands out parts the way a lead hands out work. A teammate who isn't shy got an improv scene where he had to talk the party past a guard. "The persuasion didn't work." Other people got a line to read off the screen. And for some people, the right call is no speaking lines at all.

Her rule is "you have to be attuned to your team." Don't make it your first retro at a new job. And "Every time you're silly, every time you put in a game, it should be for a purpose."

That's the same skill as facilitating any review. Some people will walk through their part of an outage in front of twenty people. Some will only tell you the real story one-on-one afterward. You have to know who's who.

There's one part she's less flexible about, and I liked it a lot. She uses the retro for short knowledge-sharing sessions. If someone is the best person to walk the team through a new cluster architecture, that part isn't in character. "Let's not do metaphors. We're not gonna do mappings." The technical content comes first, and "we'll map it to something silly after the fact."

The game never gets to water down the engineering.

Timing matters too. Jinger won't run one in the middle of a fire. "Let's not play games right now." People have to be comfortable before they'll speak their truth, and a team that's stretched thin isn't comfortable.

What's interesting is that her team does joke during fires. She said the jokes help them de-stress, and "They make us better troubleshooters."

So the line isn't humor versus no humor. A well-timed joke on an incident bridge can reset the room. Asking a team that's been paged all week to open up about how things are really going is a different request entirely.

The character sheets were another place where the silly part was doing real work. Each in-game race maps to a technical archetype. A Goliath is a high-level systems thinker. A gnome is a tinkering automation engineer. She compared it to a horoscope, and people love it.

It isn't a secret mapping she keeps to herself. The team talks through what their picks say about their strengths and weaknesses. Her framing was "tell me something about yourself that maybe we've never asked directly before."

Most of us never get asked that at work. We get asked to go do the stuff that needs to be done. She joked that you can tell her you never want to look at IaC again, and the answer might still be "okay, doesn't matter." But she also said the whole exercise helps a manager see where the team is lacking and where to put resources.

She tied it back to a skill engineers already need, too. When we explain technical things to non-technical stakeholders, we reach for metaphors. That's an exercise in building a shared language, and mapping things correctly is, as she put it, "a fundamental engineering skill."

Anyone who has written an outage summary for executives knows she's right.

When I asked how someone should start, her answer wasn't a world. It was figuring out what the retro should produce. The right question isn't "how do we fix this?" It's "in what specific ways are we unaligned?" and "when specifically is our communication breaking down?" Where are the things engineers keep getting beat down on and not telling anybody about?

I said in the episode that this matters for normal retros too, and her follow-up is the part I keep thinking about. When a retro runs on a cadence long enough, people forget why it exists: "wait, why are we doing this? It's just something we do." A bespoke experience has to justify itself up front.

That's the recurring meeting nobody remembers the purpose of. It's the alert everyone acks without reading. Every retro should have to earn its slot the way hers does.

And she holds it to a higher bar, not a lower one. For a regular retro, she said, sometimes the bar is "we've talked about the past two weeks." A failed Tabletop Retro is "nobody had fun, nobody laughed even once."

But the failure she cares about most is the one where everybody has fun, tells her something honest, and nothing changes. She framed a good retro as a kind of contract: "give me more of yourselves and in return I'm going to find a way to actually do something about these things."

I think that's exactly right, and not just for game nights. The fastest way to kill honesty in postmortems is to collect action items nobody does. People notice. Next time they give you a little less. The same failure modes come back, and the room gets quieter.

So my takeaway from this one is simple.

Silliness can buy you honest answers. Following through is how you keep getting them.

And you don't need a fantasy world to start. Kids trick-or-treating is enough. Start with the questions you need answered, then pick a setting your team already bonds over and put a slideshow on a shared screen. Ask your manager first.

Just don't tell anyone else. Half the fun is that nobody knows what's coming.