# 02 — Findings Analysis

**Dataset:** 316 customers · $5.05M ARR · real Suger-provided export
**Analysis script:** `../../analyze.py` · **Outputs:** `summary.json`, `top_priority.csv`, `churn_candidates.csv`, `expansion_candidates.csv`

## Headline

- **Total ARR:** $5,053,865
- **ARR at risk:** $2,610,572 (**52%** of the book) — this is the number every finding is measured against
- **Active:** 206 / **Inactive:** 110 (**35%** of accounts sit dormant)
- **Customers past contract-end:** 142 · **Within 90d:** 55
- **Health score avg:** 38.99 · **Below 40:** 185 · **Above 70:** 13
- **Disbursed through Suger:** $4.09B TTM / $2.70B last 6mo — the platform is *moving money*, but CS coverage is patchy

## Portfolio mix

| Maturity | # | $ ARR |
|---|---|---|
| Emerging | 180 (57%) | $1.60M |
| Scaling | 95 (30%) | $1.91M |
| Advanced | 41 (13%) | $1.54M |

**Read:** the book is bottom-heavy. Most customers never cross the Scaling threshold. Revenue is concentrated in the top 13% (Advanced), but growth motion lives in the middle 30% (Scaling) — that's where the expansion plays go.

| Sentiment | # |
|---|---|
| unknown | 238 (75%) |
| excellent | 40 |
| good | 22 |
| poor | 16 |

**Read:** 75% capture gap. Not a CSM-effort problem — a data-capture-at-source problem. See Finding #3.

## The seven findings (ranked by $ ARR exposure)

![7 findings ranked by $ ARR exposure — Week 1 closes 3 of 7 covering $3.80M](diagrams/rendered/findings-exposure.webp)

### 1. Sentiment dark zone — $3.69M exposed · 238 accounts untagged · **HIGH**

75% of accounts have no sentiment captured. CSM health calls are happening; the data isn't landing in Suger. Discovery is greater than data-cleaning here — the gap is *capture*, not analysis.

**Samples (top-5 ARR):** see dashboard · **Fix:** Week 2 — AI-enabled sentiment capture. Gong/Fathom call summaries auto-post to Suger account timeline with sentiment tag. Close the capture gap at the source, not the spreadsheet.

### 2. Bug debt on 108 customers — $2.68M exposed · **MED**

108 customers carry open bugs against their integration. Quality tax compounds into renewal risk — every week a bug ages, the quality pillar of health erodes.

**Fix:** Day 31–60 — weekly triage with Eng; tie bug-age to health penalty explicitly; top-5 $ARR accounts get a named bug-owner on the Eng side.

### 3. Renewal book uncovered — $2.39M exposed · 142 overdue · **HIGH**

142 customers are past contract-end with no visible renewal motion. $2.39M ARR sitting in limbo. This is the single largest preventable-churn exposure.

**Fix:** Week 1 — stand up a renewal stage in Suger (90/60/30/0) wired to `contract_end_date`; auto-task the CSM at T-90, escalate to ops at T-30. Every overdue account gets a named owner by Day 5.

### 4. 173 accounts with no CRM integration — $1.82M exposed · **MED**

55% of the book isn't wired to a CRM. Co-sell visibility, quota credit, and field attribution all broken. RevOps cannot measure what they cannot see.

**Fix:** Day 61–90 — CRM-connect quick-win offer. Free white-glove setup for top-50 $ARR accounts; packaged onboarding flow for the rest. Integration = visibility = expansion path.

### 5. 180 accounts stuck Emerging — $1.60M exposed · **MED**

57% of the book never crossed the Scaling threshold. Revenue is there, compounding motion isn't. Customers paid, shipped one marketplace, stopped.

**Fix:** Day 31–60 — explicit maturity playbook. Emerging → Scaling requires (a) 2+ marketplaces integrated, (b) cosell motion live, (c) first disbursement. CSM success metric = maturity promotion, not ticket close.

### 6. Contract dates missing on 95 accounts — $1.05M exposed · **HIGH**

30% of the book has no `contract_end_date`. Pipeline forecasting, renewal timing, and GRR all guessing. You cannot run a CS org on a book whose renewal dates are unknown.

**Fix:** Week 1 — required-field rule on Suger Contracts; backfill sprint driven by AR/billing export. Block renewal motion until the field is dated.

### 7. 41 customers with zero marketplaces live — $357K exposed · **LOW**

Paid for Suger, never integrated a single marketplace. Either mis-sold or blocked on implementation. Either way, $357K of ARR that customer isn't getting value from.

**Fix:** Week 1 audit — 41 accounts split into (a) in-flight implementation (accelerate), (b) mis-sold (refund/reprice), (c) abandoned (churn-now, recover the license capacity). Stop the bleed.

## Top-5 dollar exposures in one table

| Rank | Finding | $ ARR | Count | Severity |
|---|---|---|---|---|
| 1 | Sentiment dark zone | $3.69M | 238 | high |
| 2 | Bug debt | $2.68M | 108 | med |
| 3 | Renewal overdue | $2.39M | 142 | high |
| 4 | No CRM integration | $1.82M | 173 | med |
| 5 | Stuck Emerging | $1.60M | 180 | med |

## Bottom risks (where I'd focus Week 1)

1. **142 overdue contracts ($2.39M)** — renewal-stage rule closes the trigger gap.
2. **95 missing contract dates ($1.05M)** — required-field rule + 1-sprint backfill.
3. **41 zero-marketplace accounts ($357K)** — triage into accelerate / reprice / churn-now.

Week-1 ships all three. No code, no migration, no training — declarative rules plus one audit.

## Where AI-enablement fits

Three of the seven findings (sentiment capture, bug triage, stuck Emerging) are *mechanical* work — writing summaries, routing tickets, running checklist follow-ups. Once the capture gate is live, MCP-based agents reduce the human CSM load on those three by 40-60%, which buys the CSM the time to *actually call* the 20 priority accounts from the churn-risk tab.

The AI doesn't replace the CSM. The AI replaces the CSM's spreadsheet.