Why is a trend more useful than a status?

The same number, two situations

Every point-in-time status format has this flaw, and it's easiest to see with a concrete pair.

Team A hit a hard integration problem six weeks ago. Confidence fell to 3.1. They found a workaround, brought in help, and have climbed back: 3.1 → 3.3 → 3.4 → 3.6. The team is solving its problems and knows it.

Team B started at 4.5 and has drifted: 4.5 → 4.2 → 4.0 → 3.6. Nothing dramatic happened. Scope grew slightly, an estimate is looking thin, a dependency keeps almost arriving. Nobody has raised anything because no individual week was worth raising.

Both teams report 3.6 today. Both would show amber. In a status meeting both would be discussed for roughly the same amount of time, and Team A — the one that's fine — would probably get more attention, because they have a story to tell and Team B has nothing specific to report.

Team B is the one about to miss a date.

Why people can't see their own drift

The deeper argument for a trend is that it does something humans reliably fail at.

A team whose confidence erodes over six weeks doesn't experience six weeks of erosion. It experiences six ordinary weeks, each marginally worse than the one before, none different enough from its predecessor to be worth mentioning. There was never a Tuesday on which things noticeably changed — which is precisely why nobody escalated.

Nobody carries a six-week baseline in their head. Ask a team lead in week six whether things feel worse than in week one and you'll get a considered guess, coloured by how this morning went. The comparison has to be arithmetic to be reliable, and arithmetic requires a record.

This is what the niko-niko calendar got right decades ago, on a whiteboard: the value was never any single day's mark. It was the pattern in the grid, which nobody could have reconstructed from memory.

What the shapes tell you

A trend is legible in ways a number isn't. Four patterns, each meaning something different:

• Flat and high. The team believes the plan. Nothing to do.
• Gradual decline. The most valuable pattern and the easiest to miss, because no individual week looks alarming. Usually something going wrong slowly that nobody has named yet.
• Sharp drop. A discrete event. Someone knows exactly what it was; the fastest route is to ask.
• Volatile. Swinging between 4 and 2 week to week suggests the team doesn't share a view of the goal, or that the goal keeps changing. Worth investigating as a goal problem rather than a delivery one.

Covered in more depth in what to do when the trend goes down.

One thing a trend also establishes: what normal looks like. Confidence moves — problems get found and solved, and a healthy project is not a flat line at 5. Without history you can't tell an ordinary dip from a meaningful one, so you either over-react to everything or ignore everything. With a few months of history, a team's own baseline tells you which this is.

What a trend requires

Three things, and most reporting formats fail at least one.

The same question every time. If the prompt changes, the series isn't comparable. This is why free-text status updates can't produce a trend — "what happened this week" has no consistent answer shape.

The same scale every time. RAG technically qualifies, but three buckets are too coarse to show drift. A project can lose a third of its confidence without changing colour, which is exactly the case you needed to see.

A regular cadence. Irregular observations make direction ambiguous. Weekly is enough for most teams; daily suits short cycles.

Add a fourth, practical one: it has to be cheap. Anything costing real effort won't be done often enough to produce a series, which is why most attempts at longitudinal project data die within two months.

How Genchi does this

Every check-in is recorded. Each initiative displays the current aggregated score alongside the trend across recent check-ins, so the direction is visible without anyone reconstructing it.

The individual vote values are shown too — the values, not the voters — so you can see whether a moving average reflects the whole team shifting together or one person dropping sharply while the rest hold steady. Those are different situations, and the average alone would hide it.