Our daily agency call took nine minutes. Here is all of it.

You could say one word, name one move, or name one blocker. Everything else was already on the screen before the call started. That is the part agencies skip, and it is the part that makes the other eight minutes possible.

Before the call

The scoreboard is filled in before anyone joins. One row per brand. The columns: status, the one move today, the owner, open tasks, days since the client was last touched. The numbers come from the email and SMS platforms, pulled by a script the night before. Nobody types a number live. If a number has to be typed live, the call is already failing.

The row above is invented. Every real row looked like it.

Minute 1: the colour call

The facilitator reads each brand’s name. The brand’s lead answers with one word: green, amber, or red. No sentence. No context. The count of non-greens is today’s agenda. “Six reds” means six things get decided before the call ends. “Two ambers” means a short day.

The first week, people want to explain their amber. Don’t let them. The explanation happens in minute six, and only if it is a blocker.

Minute 2: the touch ledger

Every client gets touched once a day, minimum. Not a deliverable, a touch: a message, a note, a number. The scoreboard shows days since the last touch, who touched, and on what channel. Any brand over one day is named aloud and gets a recovery touch assigned on the spot, with a name.

This rule sounds soft. It is the single biggest predictor I have seen of whether a client renews, because clients do not leave over results as often as they leave over silence.

Minutes 3 to 5: shipped and shipping

Per person, not per brand. Each person says what went live yesterday and the one thing shipping today. One thing. If someone names three, the facilitator makes them pick.

The point is not accountability theatre. The point is that a pod of three people with eight brands cannot hold eight priorities each. It can hold one.

Minutes 6 to 8: blockers, trades, and one proactive play

Real blockers only. Waiting on a client asset, a deliverability flag, an approval. Each gets an owner and a clear-by time. “I’ll follow up” is not a clear-by time. “Before noon” is.

Then capacity trades, live: if two people are on briefs all day, who covers the third brand’s sends? Decide now, not in a thread at four.

Then the proactive play. One per day, rotating across brands: a strategic move the client did not ask for. A back-in-stock flow nobody built. A plain-text note from the founder. This is the slot that proves the agency builds machines instead of answering tickets, and it is the slot that disappears first when a team is tired. Protect it by putting it on the agenda.

Minute 9: the number of the day

The facilitator restates the one metric the whole team moves today, and the one-sentence answer to “how’s it going?” for the brand in focus, so anyone on the team can speak to any client the same way. Hard stop at nine.

Why nine

Nine minutes is short enough that nobody prepares for it and long enough to decide six things. Fifteen becomes twenty. Twenty becomes a status meeting. The stop is the discipline.

The weekly version

Once the daily runs clean for two weeks, add a sixty-minute weekly on the same scoreboard. The red-cell count is the agenda. Every gap is spoken in one format: target, actual, cause, the number, the action. Every action is logged with the exact cell it should turn green and by when, so next week grades this week’s forecasts. The board stops being a snapshot and becomes a ledger.

What this has to do with automation

The scoreboard is the automation. Everything the huddle needs, an agent can pull, format, and post before the call: the platform numbers, the days since last touch, the open task counts, the flows that dropped more than ten percent week over week. The humans keep the nine minutes. The machine does the hour before it.

That is the pattern the book keeps returning to. Automate the preparation. Keep the decision.

What broke

The touch ledger depended on people logging touches, and people did not. The fix was reading the shared channel for the brand’s name instead of asking anyone to log anything.

The proactive play got skipped on heavy days, quietly, until it was moved to the top of minute six, before the blockers. Put the thing you want protected first.

The scoreboard’s task counts were wrong for two brands whose work was filed under product names instead of the brand. The count is only as good as the filing, and nobody notices a wrong count on a green row.