MetrqSign in

Read the cycle-time scatterplot

The cycle-time scatterplot is the single chart that tells you how long work actually takes on your team. Every dot is one finished item. This guide gets you reading it in a couple of minutes; for the underlying idea of why Metriq measures instead of estimates, see How forecasting replaces estimation.

This is a how-to. It assumes you have completed at least a handful of items so there are dots to read.

Find it

Open Analytics and go to the Flow group (use the chip row at the top of the page to jump there). The scatterplot is the cycle-time chart — dots plotted against time.

What the axes mean

  • X axis — Completion Date. Where a dot sits left-to-right is when the item finished.
  • Y axis — Cycle Time (days). How high a dot sits is how long it took, from the moment it started to the moment it finished.

So a dot in the upper-right is an item that finished recently and took a long time. A dot near the bottom finished fast. See cycle time and lead time for the precise definitions and how they differ.

Read one dot

Hover any dot for a tooltip: the item's title, its cycle time in days, and its completion date (with the actor-kind lens on, it also shows a Person/Agent chip). Click a dot — or focus it and press Enter / Space — to open that item's card, where you can see everything about it, including when it started.

Opening the card is how you investigate an outlier. See a dot way above the rest? Click through, read the title, go ask what happened to that item.

Read the percentile lines

The horizontal dashed lines are percentiles of your cycle time. By default P50, P85, and P95 are shown; each line is labeled on the right with its value in days.

  • P50 — the median. Half your items finished at or below this line.
  • P85 — 85% of items finished at or below this. This is the line to plan around: it is fast enough to be realistic and conservative enough to usually hold.
  • P95 — 95% finished at or below this. The tail; the rare slow item lives above it.

Click any percentile in the legend below the chart to toggle its line on or off. P70 is available too. The lines are pooled across all contributors — turning on a lens (below) recolors dots but never changes the percentiles.

Read the spread

The shape of the cloud is the message:

  • A tight band of dots means consistent flow — items take roughly the same time, so your P85 is trustworthy and your forecasts will be tight.
  • A wide scatter means high variability. The gap between P50 and P95 is large, work is unpredictable, and the fix is usually to reduce that variability (smaller items, fewer blockers, less work in progress) rather than to work faster.

You are not trying to push every dot to zero. You are trying to pull the cloud into a tight band low on the chart.

Compare people and agents

Click Lens: actor kind above the chart to color dots by who completed them — People vs. Agents. The dots recolor; the percentiles and totals stay exactly the same (the caption says so). Use it to spot whether human and agent work have different cycle-time profiles, not to change the numbers.

See your SLE on the chart

If you have set a Service Level Expectation, it shows as a bold target line, labeled like SLE: 85% ≤ 7d. Dots below that line met the expectation; dots above it breached it. If you have not set one yet, see Set an SLE.

Related