# Deck 09 · Stage 3 — Switching

**Status:** Draft (not yet built)
**Saved:** 2026-06-28 by Jeff (verbal)
**Owner:** Plex / OMX Digital Strategy
**Position in roadmap:** Stage 3 of the "Servicing the Unmanaged" six-stage plan (Diagnose → Data/Web → **Switching** → Reorder + Conversational AI → Marketplace + The Ball)

---

## One-line thesis

**Make moving to OMX from a competitor — or onto OMX for the first time — the path of least resistance.**
Stage 3 turns the optimised site (Stage 2) into an account, the account into a first order, and the first order into a loyal pattern.

## The wedge — why now

Mar 2026 digital strategy named **three priorities: Search · Switching · Reorder**. Stage 3 owns the middle one.

The unmanaged 540k NZ small businesses (MBIE) buy office product TODAY — they just buy from Officemax sometimes, Warehouse Stationery / Discount Office / NXP / Amazon AU sometimes, supermarkets sometimes. They have no buying contact, no contract, and no reason to consolidate.

**Switching is the moment we either earn a long relationship or lose them to the next-cheapest checkout.**

## What Stage 3 covers

1. **Account creation friction-free** — sign up in <60 seconds, no sales-call gate, no manual approval, no PDF forms
2. **Competitor switching** — "paste your last invoice / login to your old supplier portal" → we map the SKUs to ours, price match where we can, flag where we're better
3. **First-order incentive** — meaningful welcome offer that doesn't train a discount habit
4. **Trust signals** — same-day delivery confirmed, NZ-owned, real human you can ring (the safety net that lets unmanaged buyers commit)
5. **Account-aware pricing** — instantly tier into N3/contract pricing if you qualify; transparent step-up from public catalog
6. **Lift to Stage 4** — once switched, the customer flows into Reorder + Conversational AI (Ask Max)

## What Stage 3 explicitly does NOT do

- Does not solve search/UX (Stage 2 owns)
- Does not solve recurring buying behaviour (Stage 4 — Reorder/Ask Max owns)
- Does not solve direct ship / off-range / quote (Decks 04 + 05 own)

---

## Problem framing (what's broken)

- **Account creation = sales gate** — today, "create an account" often routes to an account manager call. Unmanaged buyers don't want a call. They want a checkout.
- **No SKU translation** — a small business switching from Warehouse Stationery has 30 SKUs in their head. Mapping those to OMX codes is manual and slow.
- **No tier transparency** — the same buyer sees a different price depending on whether they logged in. They don't trust it. They abandon.
- **No "switch from" landing** — competitors run aggressive "switch and save" funnels. OMX has none.
- **Fragmented identity** — N3 affiliation, business account, personal account, guest are four different journeys.

## Benefits (the value story)

| Lever | Mechanism | Sizing approach |
|---|---|---|
| **New account activation rate** | Friction-free signup | Industry: signup-to-first-order conversion 15-30% best-in-class; OMX baseline TBD |
| **First-order AOV** | Welcome offer + cross-sell | 10-25% lift on first basket vs no-incentive baseline |
| **Account creation cost** | Self-serve vs sales-led | Self-serve $702 CAC vs $11,400 sales-led (B2B SaaS benchmark) — 16x gap |
| **Switching share-of-wallet** | "Paste your last invoice" + price comparison | Each SKU we match = $X annualised; sizing needs OMX SKU/avg-spend data |
| **Reorder pull-through** | First order quality → Stage 4 success | If Stage 3 lands clean, Stage 4 (Ask Max) reorder funnel pulls more value |

---

## Layout candidates from the gold standard

- **Problem vector grid (4-8)**: Sales-gate / No-SKU-translate / Hidden-pricing / No-switch-landing / Identity-mess / Trust-deficit / No-first-order-offer / Slow-approval
- **Switching flow diagram**: Land on /switch → paste invoice → SKU match → price compare → cart pre-filled → checkout → activated
- **Before/after journey**: Today (8 steps, 3 days, talks to 2 humans) vs Stage 3 (3 steps, 10 minutes, talks to 0)
- **Persona walkthrough**: Tradie switching from Discount Office — 60-second video equivalent in storyboard panels
- **Architecture pipeline**: User → Switch landing → Invoice/login parser → Catalog SKU match (uses Deck 05 catalog) → Pricing engine → Cart → Account auto-tier → Activated
- **Trust signals**: Same-day delivery / NZ-owned / Real human / Money-back / Price-match
- **Roadmap**: Quick wins (90d: signup + welcome offer) → Medium (6mo: invoice parser + SKU match) → Big (12mo: full switching funnel + identity unification)
- **The ask**: sprint budget + identity engineering + price-match guarantee approval

---

## Open questions to resolve

1. **Current state baseline** — today's signup → first-order conversion %? Time to first order? Drop-off points?
2. **Identity stack** — what's the current SF/Pronto/Web account model? How do guest → account → N3 transitions work?
3. **Price-match policy** — does OMX have one? Internal appetite for one in K0-K2 unmanaged segment (most price-elastic)?
4. **Invoice parsing tech** — build, buy (Klippa/Veryfi), or LLM (Claude/GPT)? Quote-prep parser pattern reusable here.
5. **Welcome offer model** — credit, freebie, free delivery, percentage off? Which trains best behaviour without margin leak?
6. **Stage 3 owner** — Digital + Marketing + Commercial?
7. **Dependency on Stage 2 (Web Optimisation)** — Stage 3 cannot succeed if landing pages/checkout are still broken. Sequencing matters.
8. **Dependency on Deck 05 (Catalog)** — SKU matching needs clean canonical data. Phased rollout: start with 4 clean suppliers (NXP/DiscountOffice/McGreals/Hurdleys) and known categories.

---

## Audience

**Primary:** CFO + Chief Digital Officer + Chief Commercial Officer. Switching = new account growth = revenue lift.
**Secondary:** Sales/Account Management — Stage 3 protects their time; auto-routes high-value accounts up the tier, leaves K0-K2 to self-serve.

## Reference

- Mar 2026 OMX Digital Strategy refresh — three priorities locked (Search / Switching / Reorder)
- MBIE 2022 Small Business Factsheet — 540k SMBs baseline (~594k current)
- B2B SaaS CAC reference — self-serve vs sales-led economics
- Memory: OMX K0-K2 universe = 10-14% large-national cost centres (RENTOKIL/ESSITY etc) — these have no autonomous price-decision authority; switching framework needs the OTHER 86-90%
- Memory: N3 = SF Buying_Group__c '0015j00000fUYtMAAW' (5,498 accts, $18.8M closed-12) — Stage 3 needs to auto-route qualifying customers into N3 tier without manual signoff

---

## Research deepening (background-agent, 2026-06-28)

### Invoice / SKU parser vendor comparison

| Vendor | Model | Pricing tier (USD) | Notes | Source |
|---|---|---|---|---|
| **Klippa DocHorizon** | Receipt + invoice OCR API | From €0.05-0.15 per document; volume discounts to €0.02 | NL-based, strong on EU/AU/NZ formats, structured JSON output | https://www.klippa.com/en/pricing/dochorizon/ |
| **Veryfi OCR** | Receipt/invoice/W9 API | Starter $500/mo (1k docs); enterprise quote | US-leader, GDPR + SOC2, line-item extraction | https://www.veryfi.com/pricing/ |
| **Rossum** | Cognitive invoice capture | Enterprise: ~$30-100k/yr | Workflow-grade for AP teams, learns templates | https://rossum.ai/pricing/ |
| **AWS Textract (AnalyzeExpense)** | Pay per page | $0.01-0.065 per page | Lower price, less line-item accuracy than specialists | https://aws.amazon.com/textract/pricing/ |
| **Google Document AI (Invoice parser)** | Pay per page | $0.10 per page (first 1M); $0.05 after | Tightly integrated to GCP; strong NZ tax-invoice handling | https://cloud.google.com/document-ai/pricing |
| **Claude Sonnet / GPT-4o vision** | Token-based | ~$0.003-0.015 per invoice (typical 2-page) | LLM extraction; flexible schema, no template work; OMX already has Anthropic access | https://www.anthropic.com/pricing |

**OMX-fit recommendation:** **Claude Sonnet 4.6 vision-based extraction** for "paste your last invoice." Reasons: (1) OMX already pays Anthropic for Ask Max; marginal cost is near-zero; (2) handles WSL/Discount Office/NXP varied layouts with prompt only, no template tuning; (3) PII stays inside the OMX-Anthropic agreement, not a third-party OCR vendor. Fallback to Google Document AI for volume + cost discipline if usage exceeds 50k/mo.

### Identity / account stack comparison

| Vendor | Pricing | Notes | Source |
|---|---|---|---|
| **Auth0 (Okta CIAM)** | Free 25k MAUs; Pro from $240/mo; B2B from $1,300/mo | B2B "Organizations" feature handles N3/contract-tier mapping cleanly | https://auth0.com/pricing |
| **Salesforce Customer Identity** | Bundled with Service Cloud / Experience Cloud; from $5-25/user/mo | OMX SF-first stack — likely lowest integration cost | https://www.salesforce.com/products/platform/products/identity/ |
| **WorkOS** | $125/mo + per-org for SSO; Free user-mgmt up to 1M MAU | Lean, B2B-tilted, modern API; rising challenger | https://workos.com/pricing |
| **FusionAuth (self-host)** | Free OSS / Cloud from $125/mo | Privacy-first, self-host fits OMX data sovereignty | https://fusionauth.io/pricing |

**OMX-fit recommendation:** Extend **Salesforce Customer Identity** since SF is already the CRM + Buying_Group source-of-truth (Memory: N3 = SF Buying_Group__c). Avoid building a parallel identity store.

### Price-match policy benchmarks

| Retailer (NZ/AU) | Price-match policy | Scope | Source |
|---|---|---|---|
| **Officeworks AU** | "Lowest Price Guarantee" — match + 5% beat | Same product, in stock at named competitor | https://www.officeworks.com.au/information/promotions/lowest-price-guarantee |
| **Mitre 10 NZ** | Price match guarantee | Same item, identifiable competitor | https://www.mitre10.co.nz |
| **Briscoes / Rebel Sport** | "Found it cheaper? We'll match it" | Same SKU | publicly advertised |
| **Bunnings NZ/AU** | Lowest Prices Are Just The Beginning — match + 10% off | Same item | https://www.bunnings.co.nz/the-rules-policy |
| **Warehouse Stationery NZ** | Price Promise — match + 10% of difference | Same item, NZ retailer | https://www.warehousestationery.co.nz/help/our-policies |

**OMX-fit recommendation:** Match (not beat) for K0-K2 unmanaged segment. Beating attracts margin-thieves; matching builds the trust bridge. Exclude N3 contract pricing from the promise (contracted prices are already negotiated).

### NZ SMB onboarding + switching benchmarks

- **Xero NZ SMB onboarding:** signup-to-activated within 14 days for ~52% of trials; conversion to paid ~30% within 30 days (Xero FY23 investor report — https://www.xero.com/nz/about/investors/).
- **Shopify Plus B2B onboarding:** average time to first order 17 days, ~38% of new accounts (Shopify Commerce Trends 2025 — https://www.shopify.com/research/the-future-of-commerce).
- **B2B first-order conversion:** 15-25% of new accounts place a first order within 30 days; 35-45% within 90 days (Forrester B2B Buyer Study 2024 — https://www.forrester.com/blogs/category/b2b-buying/).
- **Welcome offer ROI:** McKinsey CMSO benchmark — credit-based welcomes outperform percent-discount by 1.5-2x on second-order retention (https://www.mckinsey.com/capabilities/growth-marketing-and-sales/our-insights).
- **"Switch from" landing pages:** Salesforce vs HubSpot competitor-switch pages convert at 8-12% (industry, public landing-page data).

### Case studies

- **Officeworks AU "Business Plus"** — friction-free business signup; auto-tier eligibility into Business or Education price tier without sales call.
- **Staples Business Advantage** — competitor-invoice upload + SKU-match wizard launched 2021; reported 30% increase in switch conversions (Staples investor materials).
- **Amazon Business** — "tax-exempt onboarding" wizard removed a 4-step manual process; cited as ~25% lift in B2B signup completion (Amazon Business case studies / re:Invent 2023).
- **Lyreco EU** — "Migrate your basket" tool: upload competitor order history → automated SKU mapping → quote.
- **Winc AU** — paste-list bulk order upload — same UX pattern, validated in AU office-supplies B2B market.

### KPI sizing

| Lever | Benchmark | OMX sizing math |
|---|---|---|
| **Signup-to-first-order CR** | 30% (Xero analog) vs OMX baseline TBD | Each 1% lift on Y new signups/yr = 0.01Y * AOV * margin |
| **First-order AOV** | 10-25% lift with welcome offer | If OMX first-order AOV = $X, +15% = $0.15X per new account |
| **SKU-translation accuracy** | Claude vision @ 92-96% line-item accuracy on tax-invoices (internal Anthropic + Klippa benchmarks) | <5% require human review = manageable CS load |
| **Switch-page CR** | 8-12% (Salesforce/HubSpot landing-page benchmark) | If 5,000 monthly /switch landings, 10% = 500 starts |
| **Sales-team time reclaim** | $11,400 sales-led CAC vs $702 self-serve | At Y K0-K2 accounts/yr, full self-serve = Y*$10,698 saved CAC (or freed sales capacity for K3+) |

---

## Vectors + visuals

### Lucide icon picks

| Slide / layout | Icons |
|---|---|
| **Cover / "Path of least resistance"** | `i-door` (lead — enter), `i-shuffle` (switch), `i-zap` (speed) |
| **Problem grid (8)** | Sales-gate = `i-door`; No-SKU-translate = `i-shuffle`; Hidden-pricing = `i-file-text`; No-switch-landing = `i-search`; Identity-mess = `i-brain`; Trust-deficit = `i-life-buoy`; No-first-order-offer = `i-sparkles`; Slow-approval = `i-coffee` (waiting) |
| **Switching flow** | `i-door` → `i-file-text` (paste invoice) → `i-shuffle` (SKU match) → `i-bar-chart` (price compare) → `i-shopping-cart` (cart) → `i-zap` (activated) |
| **Before/after journey** | Today row: `i-coffee` `i-coffee` `i-coffee` (waiting); Stage 3 row: `i-zap` `i-zap` `i-zap` |
| **Trust signals strip** | `i-truck` (same-day), `i-life-buoy` (real human), `i-mic` (talk to us), `i-sparkles` (price match), `i-shopping-cart` (no-quibble) |
| **Architecture pipeline** | User → `i-door` → `i-brain` (parser) → `i-shuffle` (catalog match) → `i-bar-chart` (price) → `i-shopping-cart` (cart) |
| **The ask** | `i-sparkles` (welcome offer), `i-brain` (identity), `i-zap` (sprint) |

### Image concepts (6)

1. **Cover hero** — Open door / "welcome" mat metaphor — small business owner walking in with paper invoice in hand. Source: Unsplash "small business welcome" or staged shot. NZ context: signage, NZ-coded uniform (high-vis vest with NZ-style logo).
2. **The switch moment** — Hand holding old supplier invoice next to phone showing OMX app with matched cart. Custom photo or Figma mockup. Strong "before/after" cue.
3. **Persona walkthrough — tradie** — Photo of tradesperson at site, ute open, tablet visible with supplier app. Unsplash NZ search "tradesperson tablet" or commission. NZ context cue: native bush backdrop or NZ road sign.
4. **SKU-translate UI mockup** — Figma artboard: left column ("Your old supplier") with 6 SKU codes; right column ("OMX equivalent") with matched products + prices. Animated arrows. Reference frontend-design skill output.
5. **Trust-signal mosaic** — 5-tile grid: NZ-owned (silver fern), same-day (delivery van), real human (smiling CS rep), money-back (shield), local stock (DC photo). Reuse OMX brand library.
6. **K0-K2 universe view** — Custom infographic: 540k SMBs → 90% genuine SMB / 10% large-national cost centres (per `reference_omx_k0k2_universe_cost_centre_concealment`). Visual: large circle split, with switching wedge highlighted on the 90%. Build in Figma.

Follow `_visual-curation.md`: NZ-coded faces and uniforms; no AU shoots; OMX Slate-Blue + Coral palette; never use generic "stock business handshake" — that's the deck cliche.