# Payment Cascades — Turkey & China

This is the core document in the pack. A cascade defines the order channels are tried in for a given GEO/vertical: **L1** lines take traffic first, fallback tiers pick up what L1 doesn't clear. This document shows each cascade as actually operated in August, then the confirmed or proposed rebuild for September–October — cost-ordered, with a conversion-rate check wherever that data exists, and with limits/traffic-fit called out explicitly where cost alone would be misleading (Turkey Betting).

## Methodology for the proposed cascades

1. **Sort by PayIn cost, ascending**, within each vertical's pool of Live/AFU-approved channels.
2. **Apply an 80% conversion floor only where conversion data is on file.** A channel below 80% is flagged for review regardless of how cheap it is — cost alone doesn't justify first position if it's failing to convert. A channel with *no* conversion data isn't penalized (there's nothing to measure yet) — it's flagged "unmeasured, monitor once live."
3. **Never reorder a channel that isn't Live/AFU-approved.** In-integration and not-yet-enabled lines stay out of the active cascade until they clear that bar — they appear here as pipeline, not as ranked options.
4. Every proposed change is justified against a real number from `docs/turkey_channel_review.md` / `docs/china_channel_review.md` — nothing here is reordered on judgment alone.

---

## Turkey — Betting

### Current cascade (August, as actually operated)

This supersedes the tracker's raw L1/L2 tags — confirmed by the payments team as how the cascade is actually run:

```mermaid
flowchart TD
    A[Betting traffic] --> L1["L1 — Payinextra — 4.5% — discontinued early Aug — carried most volume while active"]
    A --> L1B["L1 — BigIdea — 4.75% — idle"]
    L1 --> L2A["L2 — Corytech — 5% — idle"]
    L1 --> L2B["L2 — Prime — 5.2% — 30,000 USDT/mo, absorbed volume after Payinextra stopped"]
    L1B --> L2A
    L1B --> L2B
    L2A --> L3["L3 — the rest: Favori, E4A, Astrum"]
    L2B --> L3
```

**Why Prime ended up carrying the volume despite sitting at L2, not L1:** Payinextra (4.5%, the cheapest channel on file) carried effectively all of Turkey's volume while it worked, on low fee and low, stable limits. It stopped working in early August. Prime then absorbed nearly all of that volume instead — not because it's the cheapest (5.2%, above BigIdea's 4.75%), but for the same reason Payinextra worked: low, stable settlement limits fit to small-ticket traffic. BigIdea, still Live and AFU-approved throughout, never picked up this volume despite being cheaper than Prime — limits, not rate alone, determine where volume actually lands. See `docs/cost_analysis.md`.

**Still worth flagging:** E4A sits in L3 at a higher cost *and* the only measured conversion rate in Turkey below an acceptable floor (50–70%) — kept as a review item below regardless of the limits point above.

### Proposed cascade — September–October (desired)

Payinextra does not exist for this period — it's discontinued. Two new channels (Continental, BnPay) and a renegotiated Astrum rate (5.5% → 4.5%) combine into a strong, low-cost L1 tier that effectively replaces the capacity Payinextra carried in August, at the same 4.5% rate:

```mermaid
flowchart TD
    A[Betting traffic] --> L1{"L1 — Sept–Oct"}
    L1 --> Continental["New channel — Continental — 4.5% — limits PayIn 500"]
    L1 --> BnPay["New channel — BnPay — 4.5% — limits PayIn 500"]
    L1 --> Astrum["Astrum — 4.5% — renegotiated from 5.5% — limits PayIn 500"]
    A --> L2{"L2"}
    L2 --> BigIdea["BigIdea — 4.75% — idle"]
    L2 --> Corytech["Corytech — 5%"]
    L2 --> Prime["Prime — 5.2% — keeps its limits-fit role"]
    A --> L3["L3 — the rest"]
    A --> Review{"Flagged for review"}
    Review --> E4A["E4A — 5.5% base — conv. 50–70%, below 80% floor"]
```

**Rationale:**
- Continental and BnPay enter directly at L1 at 4.5% with a 500 PayIn limit — both confirmed for Sept–Oct, both cheaper than everything in the current L1/L2 tiers.
- Astrum's renegotiated rate (4.5%, also limits PayIn 500) moves it from L3 into L1, alongside the two new channels — three low-cost, stable-limit channels now lead the cascade, ahead of BigIdea.
- BigIdea moves from its old solo L1 spot to L2, ahead of Corytech and Prime on cost (4.75% vs. 5% and 5.2%) — still idle, still worth investigating, but no longer the cascade's sole first line now that three cheaper channels exist.
- Corytech and Prime stay in L2 behind BigIdea: Prime keeps a strong position given its limits fit the traffic, but no longer needs to be the primary volume carrier once L1 is genuinely cheap and stable.
- E4A stays flagged for review, unchanged from August — cost and conversion concerns are independent of the L1 rebuild.

**Finding:** a strong L1 for September–October — low rates paired with stable, low limits — ahead of everything currently in the August cascade.

---

## Turkey — Forex

```mermaid
flowchart TD
    A[Forex traffic] --> L1["Astrum — 5.5% — 0 vol — sole channel"]
```

**Note:** Forex has exactly one channel on file (Astrum) and no traffic launched yet. There is no cascade to optimise here — the finding is structural: **single-channel dependency, no fallback tier exists.** Worth a line item once Forex traffic actually starts.

---

## Turkey — Adult

```mermaid
flowchart TD
    A[Adult traffic] --> Pending["Corytech (Adult) — 6% — AFU: No — not yet launchable"]
```

**Note:** Not a cascade yet — this is a launch-readiness item, not a reorder item. Corytech-Adult needs to clear AFU before it has a cascade position at all.

---

## China — General Payin (Low/mid-risk + Betting + Forex)

### Current cascade

```mermaid
flowchart TD
    A[Payin traffic] --> L1{L1 tier}
    L1 --> TCL["TCL — 4.5% — conv. 90% — 0 vol"]
    L1 --> FSP["Fastsecurepay — rate unconfirmed — 0 vol"]
```

### Proposed cascade

```mermaid
flowchart TD
    A[Payin traffic] --> L1{"L1 — priority order"}
    L1 -->|1st, confirmed| TCL["TCL — 4.5% — conv. 90% — cheapest AND best-tested"]
    L1 -->|2nd, conditional| FSP["Fastsecurepay — do not rank on cost until rate direction is confirmed — see data_gaps.md"]
```

**Rationale:** TCL is the clear priority the moment payin traffic launches — it's both cheaper and the only one with conversion data. Fastsecurepay keeps its slot but shouldn't be treated as cost-competitive with TCL until its actual payin rate is confirmed (see `docs/data_gaps.md`) — ranking it today would be ranking on a number that may not even be the right metric.

---

## China — Adult Payin

```mermaid
flowchart TD
    A[Adult payin traffic] --> Sole["Moneystery — 9.5% — sole Adult-vertical channel — 0 vol"]
```

**Note:** Single-channel, no alternative on file for this vertical — the 9.5% rate is the cost of accessing Adult in China at all, not a routing inefficiency. No reorder possible; worth sourcing a second Adult-capable channel if volume here grows.

---

## China — Payout

### Current cascade (August)

```mermaid
flowchart TD
    A[Payout traffic] --> L1["TCL — 2.9% all-in (2.7% payout + 0.2% agent) — conv. 85/85% — 1,500,000 USDT/mo — sole confirmed channel"]
    A --> Pipeline["TCL corporate payout — in integration, no rate yet"]
```

### Proposed cascade — September–October (confirmed)

```mermaid
flowchart TD
    A[Payout traffic] --> L1{"L1 — Sept–Oct"}
    L1 --> HTX["New channel — HTX P2P +3% nominal — ~4pp cheaper all-in — confirmed, starting integration"]
    L1 --> Second["New channel — second candidate — rate not yet confirmed"]
    L1 --> TCL["TCL — 2.9% all-in — proven, kept for stability"]
```

**Rationale:** TCL's all-in cost (2.7% payout + 0.2% agent = 2.9%) and the HTX channel's nominal margin (~3%) look close on the surface — that's expected, since the HTX +3% is a **single fee blended directly against the live exchange rate**, not a separate markup stacked on top of a reference rate the way TCL's structure is. That's exactly why the comparison isn't a simple "3% > 2.9%, so it's pricier": per the payments team, once the live HTX P2P rate is accounted for, the true all-in cost comes out roughly **4 percentage points lower**. A second new payout channel is also planned for Sept–Oct with its rate not yet confirmed — kept as a pipeline candidate in the cascade, not cost-ranked. All three sit at or alongside L1: the confirmed HTX channel absorbs the cost-sensitive share of volume, TCL stays for its proven track record and stability while both new channels start integration. This is a confirmed change for the HTX channel, not a scenario to verify — see `docs/data_gaps.md` for the provenance of the 4pp figure and the second channel's status.

**Tied growth targets (by end of October, as provided):** payout volume 1,500,000 → 3,000,000 USDT/month; blended margin 0.1% → 4–5%. See `docs/china_channel_review.md`.

---

## Summary of changes proposed in this document

| GEO | Vertical | Change | Why |
|---|---|---|---|
| Turkey | Betting | Added Payinextra (4.5%, discontinued early August) to L1; corrected current cascade to L1: Payinextra, BigIdea / L2: Corytech, Prime / L3: the rest | Reflects how the cascade was actually operated, per the payments team, not the tracker's raw tags |
| Turkey | Betting | Sept–Oct: add Continental + BnPay (4.5%, limits PayIn 500) to L1, renegotiate Astrum to 4.5% (limits PayIn 500) and move it to L1 ahead of BigIdea | Builds a genuinely low-cost, stable-limit L1 tier — effectively replacing the capacity lost when Payinextra was discontinued |
| Turkey | Betting | Keep E4A flagged for review | Only channel with a known conversion problem (50–70%), unchanged by the L1 rebuild |
| Turkey | Forex / Adult | No reorder — structural flags only | Single-channel dependency (Forex); not yet launchable (Adult) |
| China | Payin | Rank TCL first once traffic launches | Cheapest and only channel with conversion data |
| China | Payin | Don't cost-rank the second payin channel yet | Rate direction unconfirmed — see `docs/data_gaps.md` |
| China | Payout | Sept–Oct: add HTX channel (confirmed) and a second new channel (rate TBC) at L1 alongside TCL | HTX channel confirmed ~4pp cheaper all-in once live FX is accounted for; second channel pending rate confirmation |