# Abandon opportunity sizing + recovery t-shirts

**Date:** 2026-07-29  
**Scope:** Start→Submit abandon only (lead bet · intent already expressed)  
**All numbers [SIM]**, not Kanmon disclosed. Hold credit box / accept / fund rates fixed.

**Companions:** `kanmon-funnel-volume-simulation-20260729.md` · summary · F14 in workbench

---

## 0. Two paths (prevent → recover)

Abandon has **many reasons**, like preventable declines, no silver bullet.

| Path | Question | Shape | Priority |
|------|----------|-------|----------|
| **Prevent** | What’s blocking completion *in session*? | Shorter/clearer apply, progress, prefill, handoff, bank/KYB help, meet them where they are | **Primary / lead bet**, easier to keep than to win back; most starters never finish (~40–42%) |
| **Recover** | They left, how do we act on intent? | Email + OTP → resume last step | **Secondary**, worthwhile residual |

**Logic SoT:** [`kanmon-drive-down-submission-experience-20260730.md`](kanmon-drive-down-submission-experience-20260730.md)

**Start→Submit ~40–42% [SIM]** → majority of starters don’t finish. Same pot (288) and t-shirts apply to **either** path: both move Start→Submit → funded $. Anthony’s call: lead with **submission experience** (prevent), not recover-first.

---

## 1. Why this pot

They **started** → intent detected. Job = finish submit (prevent dropout first; recover second). Not cold Seen traffic. Preventable *declines* (post-submit) stay a separate, later bucket.

---

## 2. Opportunity size (monthly)

| | N | Notes |
|--|---|--------|
| Started apply | **480** | Intent pool |
| Submitted | **192** | 40% Start→Submit |
| **Abandoned mid-apply** | **288** | **Pot** |
| Downstream Submit→Funded | **26.6%** | 51 / 192 |
| Blended ticket | **~$18k** | Mix of invoice/AP/term/LOC |
| Base book | **51 funded · ~$912k** | / mo |

### Gross addressable (ceiling: not the bet)

If **100%** of 288 completed submit and faced today’s downstream CVR:

| | |
|--|--|
| Extra funded | **~77** / mo |
| Extra financed $ | **~$1.4M** / mo (~$16M / yr) |
| vs book | ~**1.5×** current funded volume |

That’s the **the hiring manager theoretical** if every abandoner finished and converted like today’s submitters. Reality is far below.

---

## 3. What’s inside the 288 (speculative mix)

| Cause | Share | N | Recovery fit |
|-------|-------|---|--------------|
| Bounce at handoff / login | 22% | 63 | Medium, resume helps if session salvageable |
| Abandon mid-form | 28% | **81** | **Strong**, classic pick-up-where-left-off |
| Bank link friction | 20% | 58 | Weak–medium, need bank UX, not email alone |
| KYB / entity stuck | 18% | 52 | Weak–medium, need validation/entity, not email alone |
| “Save later” never returns | 12% | **34** | **Strong**, email + OTP is the job |

**~115 journeys (mid-form + never return)** are the **core recovery-addressable** slice; the rest need UX/KYB/bank work too. T-shirts below still use the full 288 as the pot (simpler talk), say “core slice ~40% of pot” if pressed.

---

## 4. T-shirt recovery (value if it works)

Define recovery success = **% of abandon journeys that become extra complete submits**, then apply Submit→Funded **26.6%** × **~$18k**.

| T-shirt | Recover → submit | Extra submits | Extra funded / mo | Extra financed $ / mo | Start→Submit becomes | vs book |
|---------|------------------|---------------|-------------------|------------------------|----------------------|---------|
| **S** | **10%** of 288 | ~29 | **~8** | **~$140k** | ~46% | +~15% |
| **M** ★ | **~17%** (~+10 pts CVR) | ~48 | **~13–14** | **~$230–250k** | **50%** | +~27% |
| **L** | **30%** of 288 | ~86 | **~23** | **~$410k** | ~58% | +~45% |
| Ceiling | 100% | 288 | ~77 | ~$1.4M | 100% | fantasy |

★ **M = default case bet**, matches prior sensitivity (+10 pts Start→Submit). Credible for “email + OTP resume works well” under low N.

### Rough mechanism read (how a t-shirt might happen)

| Size | Mechanism story (illustrative) |
|------|--------------------------------|
| **S** | Ship email + OTP; weak copy / single nudge; mostly hits “never return”; open→resume→finish is soft |
| **M** | Persist draft · clear “continue” email · OTP · land on last step · 1–2 nudges · mid-form + save-later respond |
| **L** | M + partner re-entry deep links + bank/KYB friction cuts in parallel, more than pure recovery |

---

## 5. Annualized (M t-shirt: say this)

| | / mo | / yr |
|--|------|------|
| Extra funded N | ~13–14 | ~160 |
| Extra financed $ | ~$230–250k | **~$2.8–3.0M** |
| Gross pot (ceiling) | ~$1.4M | ~$16M |

North star metric = **partner-platform originations** → talk **extra funded count + financed $**, not take-rate (unknown).

---

## 6. Kill / prove

| Signal | Read |
|--------|------|
| Email open < ~15% or OTP→resume ≈ 0 | Kill or rethink channel |
| Resume→submit strong but Start→Submit flat | Timing/segment wrong |
| Start→Submit already >60% | Pot too small, don’t lead with recovery |
| Diagnosis: most “abandon” is bot / fraud / ineligible | Re-size pot down |

Low-N: no A/B required for case, use before/after cohort + return-to-complete rate.

---

## 7. Say on the call

> Illustrative: ~50 funded/mo · Start→Submit ~40%. ~288 started with intent but didn’t finish, abandon has many reasons, so I wouldn’t ship one silver bullet. I’d explore two paths: **prevent** (make complete easier in-flow, usually higher leverage) and **recover** (email + OTP for those who leave). Gross ceiling ~$1.4M/mo; medium lift ~+10 pts ≈ **~+14 originations / ~$230–250k**/mo. Diagnose top drop reasons first, then wedge.