Reference
Widgets
Each Compass dashboard has two tabs: Overview (business-level) and Segments (client-centric). This page walks through every widget on each — what it shows, how it is computed, and when it is most useful.
Overview tab
Business-level widgets — cash, revenue mix, concentration, cohorts. Drag-and-drop to reorder; hover the kebab menu for Export CSV or Save PNG.
Revenue by sector
What it shows. Groups your clients into industry sectors and reports revenue per sector plus a growth arrow versus the equivalent-length window immediately before this one.
How it is computed. Sectors are inferred deterministically from contact names using a keyword taxonomy (Digital, Retail, Consulting, and so on). Growth is (current − prior) / prior on the revenue side; client-count moves are reported alongside so a sector growing on one client is visible.
When to use it. Diagnostic on where the book is trending sector-by-sector. Pair with sector-momentum for the growth × share quadrant view.
Revenue per month, by client
What it shows. Stacked bar per month with your top-5 clients named and the rest folded into "Other". Shows both total revenue trend and the concentration story in one chart.
How it is computed. Sales invoices grouped by month; top 5 by total revenue in the window kept as named series, everyone else summed into "Other clients". Uses invoice.total (gross of receipt).
When to use it. A single-glance answer to "how has my monthly revenue shifted and who is driving it?" Useful for spotting a growing dependence on one client without needing the concentration-risk widget.
Overdue invoices
What it shows. Donut of your open accounts-receivable ageing — outstanding sales invoices split by how overdue they are.
How it is computed. Every unpaid ACCREC invoice is bucketed by days-overdue against the scan's as-of date. Uses invoice.amountDue where set, else (total − amountPaid).
When to use it. A quick visual for the age of your unpaid book — 0-30 healthy, 30-60 needs a chase, 60+ becomes a collections priority. Sits alongside smart-collections-triage in the Actions worklist.
Top clients by revenue
What it shows. Horizontal bar chart of your biggest revenue contributors in the selected window.
How it is computed. Sum of net invoice revenue (invoices minus credit notes) per client in-window, ranked descending.
When to use it. Always-on context for every decision. If a chart lower down says "your top client concentration is 35%", the top-clients bar tells you which client that is.
Revenue sources
What it shows. Donut splitting revenue into four buckets: retained recurring, new recurring, repeat purchases, single purchases.
How it is computed. Client-by-client classification based on invoice cadence and first-invoice date within the window. "Recurring" is 3+ invoices in the window; "retained" is a recurring client whose relationship pre-dates the window.
When to use it. The single best predictor of next-quarter stability. A book that is 60%+ retained recurring is much more forecastable than a 60% single-purchase book of the same total revenue.
90-day cash flow forecast
What it shows. Projects your bank balance forward 90 days based on the ageing of outstanding invoices, expected bill payments, and historic burn.
How it is computed. Combines outstanding AR (weighted by each client's historic pay-lag), outstanding AP (weighted by due dates), and average monthly burn to build a per-day cash position. Deterministic — no AI in this projection.
When to use it. Answer to "am I about to run into a cash pinch?". Pair with runway metric for the runway-months framing.
Concentration risk
What it shows. Top-1 client share of trailing revenue, plus a stress test — how many months of runway you lose if the top client goes.
How it is computed. Top-1 share = revenue(top client) / revenue(all). Runway-without-top-client re-runs the runway metric with the top client's revenue zeroed out.
When to use it. The single most concrete way to communicate concentration risk to a co-owner or investor. Above 30% top-share is a diversification conversation.
Sector momentum
What it shows. Growth × share quadrant of sectors — which sectors are big and growing (right), big and slowing (fixing), small and growing (invest), small and slowing (kill / accept).
How it is computed. Each sector plotted as x=share of revenue, y=growth vs prior window. Bubble size proportional to client count.
When to use it. The single best portfolio-shape widget for a services business. Complements the invest-harvest-fix-kill signal at the service-line level.
Cohort retention
What it shows. Client survival curves per acquisition month — each line is a cohort, x-axis is months since acquisition, y-axis is % of cohort still billing.
How it is computed. Groups clients by month of first-ever invoice. For each cohort, at each subsequent month, computes the share of the cohort with at least one invoice in the last 90 days.
When to use it. Diagnosis for month-N drop-off patterns. If every cohort loses 30% at month 4, something specific is happening at month 4 (onboarding hand-over, first-invoice friction).
Concentration stress test
What it shows. Animated simulation of "what happens to runway if we lose the top client tomorrow?"
How it is computed. Same maths as the concentration-risk widget above, presented as a before/after animation instead of two numbers side by side.
When to use it. A one-visual way to make the concentration case in a board or co-owner conversation. The difference between "35% of revenue" and "6 months of runway" often clicks harder in animation.
Sector momentum over time
What it shows. Animated 12-month scrubber of sector positions on the growth × share plane.
How it is computed. Recomputes the sector-momentum widget on a rolling 12-month window and scrubs through the months, so sectors trace their trajectory over time.
When to use it. Shows whether a sector is heading up or down — a snapshot lies about direction; the scrubber tells the truth.
Cash flow — animated
What it shows. Line-draw animation of the 90-day cash flow forecast with a break-even pulse.
How it is computed. Same data as the static cash-flow-forecast widget, presented as an unfolding line with a highlight where the projected balance crosses the break-even line.
When to use it. Useful for demos and screenshots. The static version is faster to read for daily use.
Cohort retention — animated
What it shows. Staggered curve reveal of cohort retention, oldest cohort first.
How it is computed. Same maths as the static cohort-retention widget. Reveals cohorts in chronological order so the evolution across cohorts is easier to follow.
When to use it. Same use as the static version — animation is preference, not new data.
Clients per sector
What it shows. Donut of client count per sector — a different lens on the same sector taxonomy as revenue-by-sector.
How it is computed. Same sector taxonomy as sector-revenue, but summed by client count rather than revenue.
When to use it. Reveals the "many small vs few big" shape of a sector. A sector with high revenue and low client count is a concentration risk within that sector; a sector with many clients and modest revenue is a healthier base.
Segments tab
Client-centric widgets — how the book is shaped, who sits in which band, and which clients this week\'s insights routed into which situational segment.
Keep / Grow / Renegotiate / Review
What it shows. Situational segmentation of clients derived from the insights the engine fired this week. Each client sits in its most-severe segment.
How it is computed. The engine reads the ranked insights on the current scan and maps each affected client into one of the four segments based on which signals fired for them (silent-churn / smart-collections → Review; keep-renegotiate-fire → Keep / Renegotiate; land-vs-expand → Grow).
When to use it. The one segmentation view that changes week-to-week rather than being fixed by revenue. Useful for planning the Monday-morning worklist.
Client pillars
What it shows. Structural bands (Pillar / Growth / Occasional / Fringe) with a diversify-or-consolidate recommendation.
How it is computed. Bands are computed from revenue share × consistency (how regular the invoicing pattern is) × tenure. "Pillar" is high on all three; "Fringe" is low.
When to use it. The structural counterpart to the situational Keep/Grow/Renegotiate/Review chart above. Pillar-heavy books are stable; Fringe-heavy books are fragile.
Top clients by revenue (Segments tab)
What it shows. Same widget as on the Overview tab — leaderboard by revenue in the window.
How it is computed. Sum of net invoice revenue per client in-window, ranked descending.
When to use it. Available on both tabs so you can stay in the Segments dashboard without needing to switch to see the leaderboard.
Ideal Customer Profile
What it shows. Attributes of your top-quartile clients — sector, size band, tenure, cadence — surfaced as a card you can use in prospecting.
How it is computed. Deterministic: takes your top 25% of clients by revenue, computes the most-common values for each attribute (sector, tenure band, invoicing cadence) and displays them.
When to use it. Turns your best clients into a shape you can go and find more of. Feeds directly into the reactivation-shortlist and attach-rate-gaps prospecting flows.
Client pillar revenue mix
What it shows. Donut of revenue by band (Pillar / Growth / Occasional / Fringe).
How it is computed. Sums each client's revenue by their pillar-classification band from client-pillars.
When to use it. Complements client-pillars — the bands tell you how many clients sit where; this donut tells you how much £ sits where.
Segment revenue split (KGRR)
What it shows. Donut of Keep / Grow / Renegotiate / Review by revenue.
How it is computed. Same client-in-most-severe-segment classification as the Keep/Grow/Renegotiate/Review chart above; sums revenue per segment.
When to use it. Shows revenue weight per situational segment — how much money is in the Review bucket vs the Grow bucket. Often more revealing than the client count.
Client tenure mix
What it shows. Donut of revenue by how long clients have been with you — New (<3mo), Growing (3-12mo), Established (1-3yr), Veteran (3+yr).
How it is computed. Each client's tenure = first-invoice to most-recent-invoice in months. Revenue in the current window is summed into their band.
When to use it. A veteran-heavy donut is a seasoned book with sticky relationships. A new-heavy donut is fragile — this quarter's revenue depends on this quarter's wins.
Client size distribution
What it shows. Donut of client count grouped by revenue band — Small (<£5k), Medium (£5-20k), Large (£20-100k), Whale (£100k+).
How it is computed. Client revenue in-window bucketed by absolute £ into four bands. Counts, not revenue, drive the slices.
When to use it. Complements client-tenure-mix. A whale-and-minnows distribution (a few big clients, many tiny ones) is concentration risk; a balanced middle is a resilient book.
Related
- Metrics — the 22 headline numbers behind the widgets.
- Concentration risk (signal) — the maths behind the concentration widget.
- Cohort retention (signal) — the maths behind the cohort widget.