v6.6.1 is the biggest flow-analytics release we've shipped since the original 5.x. Four additions: shaded percentile bands on the Cycle Time scatterplot, Vacanti zones (healthy / watch / stale) on the Aging trend chart, a brand-new Process Behavior view that runs Wheeler's XmR / SPC method against your sprint metrics, and a brand-new Cycle Time Heat Map that shows item-count density across (completion week × cycle-time band). The aim was simple: bring the extension to feature parity with what flow coaches teach in Vacanti and Wheeler workshops, without forcing anyone to relearn the navigation. Existing settings are preserved byte-for-byte and no reactivation is required. The "What this means for your existing settings" section below covers exactly what the upgrade will and will not touch.
What's new
- Cycle Time scatterplot — percentile bands. The chart now shades the regions between P50 / P70 / P85 / P95 so the typical-vs-outlier distribution reads at a glance. The dashed percentile lines stay; the bands sit behind them and the dots. Find it at Plan → Cycle Time.
- Aging & Blocked — Vacanti zones. The trend chart at the bottom of Aging & Blocked now shows three shaded zones — healthy (below P50), watch (P50–P85), and stale (above P85) — drawn from your team's own historical cycle time. The single P85 dashed line stays as the precise escalate threshold.
- Process Behavior — new view (under Measure). Wheeler's XmR (X-chart + moving Range) statistical method, as documented in Donald Wheeler's Understanding Variation and Daniel Vacanti's When Will It Be Done?. Separates routine variation (noise) from genuine process signals using the natural process limits (UCL / LCL = X̄ ± 2.66 × avg moving range). Out-of-limit points, runs of 8 on one side of the centerline, and 3-of-4 beyond ±2σ are automatically flagged. Switch the metric between throughput, cycle time (P50), and flow efficiency.
- Cycle Time Heat Map — new view (under Measure). A 2D density grid where each cell is the count of items completed in a given (week, cycle-time-band) cell. Warm cells show where the bulk of delivery sits; visible drift of warm cells downward over time is an early indicator of cycle-time creep — often a leading signal of WIP overload before the period-aggregate percentiles shift.
What this means for your existing settings
Every 6.6.1 change is additive. If you've been running the extension across 6.5.x, here's exactly what the upgrade will and will not change.
- Workflow Mapping, Item Types, WIP rules, Sprint Alerts, Notifications, SLEs, Access Control, AI, Privacy, and License are all left exactly as you have them. The two new views land under Measure as new sidebar entries — they do not overwrite anything.
- The percentile bands on Cycle Time and the Vacanti zones on Aging are pure visual overlays. The underlying data, percentile calculation, and configuration are all unchanged.
- If you previously edited any setting — added a state, picked a custom item type, set a WIP limit, configured an alert — your edits are preserved byte-for-byte across this upgrade.
- Both new views are gated to the starter tier (same as Cycle Time and Cumulative Flow), so any licensed or trial user can open them.
Why Process Behavior, not a control chart
Wheeler made a specific point in Understanding Variation: "control chart" implies engineering tolerance limits that don't apply to knowledge work. The natural process limits (UCL / LCL) describe what the process IS doing, not what it SHOULD do. Same math, different framing — and the framing matters when you're presenting to a team that's already nervous about "control" language. Throughput points that fall outside the natural limits aren't "out of spec" — they're a signal that something genuinely changed (new team member, process change, holiday week). That's what the chart flags in red.
Why a heat map, not just another scatterplot
Cycle-time scatterplots show every dot, but they get visually noisy when you have a few hundred completed items. They also don't answer the question "is our distribution shifting over time?" cleanly — that's a 2D question (week × bucket), and a 1D dot cloud can hide drift. The heat map collapses each (week, bucket) cell to a single intensity. Warm cells drifting downward across weeks = cycle time creeping up. That visual lead-indicator is usually visible a sprint or two before period-aggregate percentiles (P50 / P85) move enough to register.
Updating
The extension auto-updates from the Marketplace, so existing customers don't need to do anything. Existing trial and licensed orgs upgrade in place; no reactivation required. The first time you open the hub after the upgrade, the in-product changelog popup will show the v6.6.1 highlights once and then dismiss permanently. New installs can grab v6.6.1 from the Marketplace listing at https://marketplace.visualstudio.com/items?itemName=Baytek.agile-analytics.