I Ran My Staff Meeting Off a Live Dashboard for a Quarter

This is a ship story, not a framework. For one quarter I changed how my product leadership team met, measured what happened, and kept some of it and threw some of it out. Here is the whole thing, including the part that did not work.

The short version

I banned status updates from my weekly product staff meeting for a quarter and replaced them with a live dashboard that a small set of agents refreshed before each meeting: what shipped, what moved, eval scores and trends, cost per outcome by workflow, and a flagged list of anything off track. Everyone read it beforehand, so the meeting opened at the decisions. It killed thirty minutes of narration nobody acted on, and it surfaced uncomfortable numbers that used to hide inside confident verbal summaries, including a drifting eval score and two underwater costs no one had examined. The mistake was letting the dashboard get too dense until people stopped reading it; the fix was cutting it to the eight numbers that change decisions. Net, the best operating change I made that year.

For the meeting questions that replaced the status, see Five Questions That Make a Product Review Worth Your Time. For the cost line on the dashboard, see The Product Budget Is a Compute Budget Now.

The meeting I was tired of

The weekly product staff meeting started the way most of them do. Round the table, each leader narrates what their team did this week. Thirty minutes, sometimes more. The people whose updates were not being given half-listened and triaged their inboxes. The updates were performances of competence: confident, smooth, light on the awkward parts. And reliably, nothing was decided in that first half hour. The decisions, if any, happened in the back half when we finally got to the things that were actually contested.

I had run this meeting, in one form or another, for most of my career. I had also sat in the equivalent meeting at Salesforce, at Smartcat, and at a couple of startups, and it was the same everywhere. The format was inherited, not designed. So I decided to stop inheriting it.

What I built

The premise was simple. Status does not require a meeting. Status requires a place to look. Decisions require a meeting. So I moved all the status into a dashboard and reserved the meeting for decisions.

A small set of agents assembled the dashboard before each weekly meeting. It pulled from the systems we already used and put five things on one screen. What shipped since last week. What outcome actually moved, separate from what shipped. Current eval scores for our production AI features, with thirty-day trend lines. Cost per outcome by workflow. And a flagged list: anything off track, drifting, or underwater, surfaced automatically rather than waiting for someone to volunteer it.

The rule was that everyone read it before the meeting. The meeting itself was not allowed to start with status. It started with the flagged list and the decisions those flags demanded.

What it killed

The thirty-minute narration, immediately and permanently. Nobody missed it.

The performance of "we are on track." When the status is verbal, there is enormous room to present a yellow situation as a green one through tone and framing. When the status is a number on a shared screen with a trend line, that room collapses. You cannot narrate a declining eval score into sounding fine when everyone is looking at the red line.

The meeting as a place to look busy. With reporting gone, there was nowhere to hide and nothing to perform. You came with judgment about the flagged items or you did not have much to add.

What it surfaced

This was the real payoff, and it surprised me. When the awkward numbers are always present rather than verbally summarized, the awkward numbers stop hiding.

One of our AI features had a quietly drifting eval score. In the old format, the owning leader would have given a confident verbal update and the drift would have been a footnote, or unmentioned, for weeks. On the dashboard it was a red trend line every single week, impossible to ignore, and we caught and fixed it far earlier than we would have.

Two workflow costs turned out to be underwater. Not dramatically, but persistently, and nobody had ever looked because cost per outcome had never been on a screen anybody saw weekly. The dashboard made the absence of attention visible, which is half the battle.

The general effect was that the team's relationship to the numbers changed. When you know the dashboard is coming and everyone will see it, you start owning your numbers upstream instead of preparing a narrative about them.

What I would not do again

I let the dashboard get too dense. Over a few weeks it accreted metrics, because every metric felt useful to someone and nothing is easier than adding a tile. At some point it crossed a line and people quietly stopped reading it. That is the failure mode that matters, because a dashboard nobody reads is strictly worse than the meeting it replaced. The meeting at least forced attention by putting people in a room. An unread dashboard forces nothing.

The fix was ruthless editing. We cut it back to roughly eight numbers, the ones where seeing the number actually changes a decision. Everything else went into a drill-down nobody had to look at unless a flag pointed them there. A dashboard is a decision tool, not a data archive, and treating it as the latter kills it.

The plain version

Status does not need a meeting. Decisions do. Moving the reporting into an agent-refreshed dashboard killed thirty minutes of theater, surfaced numbers that used to hide in confident summaries, and turned the staff meeting from narration into triage. The one trap is density: keep it to the few numbers that change decisions, or people stop reading and you are worse off than when you started.

If your staff meeting opens with a lap around the table of what everyone shipped, try this for one month: move that into a doc or dashboard people read beforehand, and start the live meeting at the first hard decision. Watch how much of the old meeting turns out to have been theater.

---

If you want to compare notes on operating cadence, or what to actually put on an executive product dashboard, that is a conversation I am always up for. Find me on LinkedIn.

Further reading

• Marty Cagan / SVPG on product operating models
• Lenny Rachitsky on operating cadence and team rituals
• Anthropic on building useful agents