12 May 2026 · Field note

Charts that survive a stand-up

Open notebook beside a laptop in a quiet cafe

Stand-up is where Data Storytelling for App Teams most often collapses, because the ritual rewards completeness. A PM in Leeds last autumn arrived with six tiles “in case engineering asks”. Engineering asked about none of them. They asked why the recommendation was missing.

In the Narrative Metrics room we treat eight minutes as a physical limit, like a column width. You get one claim, one series, one confound spoken in a sentence, and one verb. If the series needs a second axis to be honest, the chart does not belong in stand-up; it belongs in a written pack.

What we cut first is comparison-to-target when the target was set in a different season. British grocery apps learned this the wet way in 2024. A green arrow against a Q2 goal during a rail strike is not bravery; it is a different story pretending to wear the old clothes.

Second, we cut live dashboards shared on a laptop camera. Compression makes small multiples illegible, so the speaker starts narrating cells. Use a single exported frame with the window printed in type, not in a filter chip the camera cannot read.

Third, we cut apologies. “This might be noise” can be true; it is not a stand-up beat. If it might be noise, your recommendation is to wait, and waiting is a verb. Say the number of extra days you want and sit down.

None of this makes a weak metric strong. It only stops a strong metric from dying of courtesy.