Churn Graph
Open menu

Find churn where your data disagrees.

Churn Graph links your CRM, email, product usage, support tickets, contracts and decision log into one graph, then shows the places where they contradict each other.

Open the risk radar
CRM dashboardHealthy

Kopi Lintas Nusantara

Critical, renewal in 75 days

  • contact history

    Champion left the company

    Aug 15, now at a prospect

  • decision log

    Promised feature slipped

    Jul 20, no new release date

  • meeting notes

    Competitor mentioned

    Sep 18, by the new CFO

Account C01 on October 1, 2026. Fictional data from the KasirNusa case study.

What it does and why

Your health score reads one system. Churn risk hides between six. We connect them, show where they disagree, and point to what worked the last time.

Six separate systems flowing into one context graphCRMEmail & meetingsProduct usageSupport ticketsContracts & billingDecision log6 sources, one graph
The CRM still lists a champion who left the companycrm_accounts.csvChampion: Rina HapsariC01, snapshot Oct 1, 2026contact_employment_history.csvLeft on Aug 15, 2026Now GM Operations at prospect P01

Evidence, at its core.

A green dashboard can hide a champion who left, a promised feature that slipped, or a competitor named in a meeting. Each fact sits in a different system, so no single report shows it.

Churn Graph keeps the file and row behind every fact, so each warning can be traced back to the records that caused it.

Built for the weekly retention meeting.

Start from the riskiest account, open the path that explains it, and leave with an action your team has approved before.

Open the risk radar
Sources connected
6
Accounts ranked
40
Days of early warning on C01
148

Every warning arrives with the records behind it, so your team acts on facts instead of a colour.

  • Risk radar

    All 40 accounts ranked by risk, with the dashboard colour beside our finding, so a green account that is really critical stands out.

  • Evidence path

    Open an account to see the path through contacts, contracts, decisions and tickets that explains its rank.

  • Source A vs source B

    Two records that contradict each other, side by side, each with its file, ID and date.

  • Retention actions

    Suggested actions grounded in past decisions, with who approved them and the cost against the revenue at risk.

  • Signal timeline

    When each signal first appeared, counted in days before the renewal date.

  • Ask the graph

    Ask a new question in plain language and get an answer that cites the node IDs behind it.

Evidence path for account C01 across contacts, decisions and meetingsAccount C01Champion K017Prospect P01Decision D-2025-11Feature FEAT-07Meeting I0331
The path for account C01: a champion who moved to a prospect, a discount tied to a feature that slipped, and a meeting where a competitor came up.
Usage drops only on outlets running version 4.12, while control outlets stay flatRelease 4.12, Jun 29OctNovDecJanFebMarAprMayJunJulAugSepTransactions per outlet, illustrative6 outlets on 4.1219 control outlets

A drop in usage is not always churn.

Account C03 looked like it was leaving: transactions fell by about a third. The graph shows the drop only on outlets running version 4.12, where offline sales failed to sync because of a known bug.

The right response is a bug fix and compensation, not a retention discount.