Scope & Roadmap¶
This page is the authoritative scope contract. Every feature in the Bible carries exactly one phase tag — MVP, Scale, or Vision — and this page is where that split is decided. When a scope debate starts, it ends here: either the feature's row says MVP, or it waits.
- MVP — everything needed to launch in Kuwait and validate the marketplace. Nothing extra.
- Scale — built after traction: retention tools, AI, GCC expansion.
- Vision — the long-term Event OS: escrow, AR, ticketing, white-label, APIs.
The architecture is built for all three phases from day one (multi-country, multi-currency, multi-language, modular monolith along the 12 domains); the feature set is ruthlessly staged.
MVP — launch in Kuwait¶
The MVP proves one loop end to end: a customer creates an event, gets quotes, books, pays a deposit, and reviews — and a vendor pays a subscription because RFQs arrive. Every MVP feature below serves that loop; anything that doesn't is out.
| Feature | Domain | Phase | Rationale |
|---|---|---|---|
| Auth: email, phone OTP, Google, Apple | Identity | MVP | Table stakes; OTP + social cover Kuwait's mobile-first users. |
| Roles & permissions (customer, vendor owner, admin roles) | Identity | MVP | Three user types share one platform from day one. |
| Vendor registration + commercial license upload | Vendor | MVP | Kuwait legal reality: licensed businesses only. Verification is the trust product. |
| Admin vendor approval workflow | Operations | MVP | Human gatekeeping of supply quality; cheaper than remediation. |
| Vendor subscriptions (Free/Basic/Professional/Enterprise) + PSP billing | Vendor + Financial | MVP | The primary revenue stream must exist and be validated at launch. |
| Vendor profile, service catalog, gallery | Vendor | MVP | Vendors need a professional storefront or nobody sends RFQs. |
| Categories + search + filters (city, area, budget, rating) | Marketplace + Platform | MVP | Discovery is step one of the lifecycle. Meilisearch from day one. |
| Customer event creation (type, date, city, guests, budget) | Event | MVP | The core entity; without it Yam3at is just another RFQ site. |
| Multi-vendor RFQ with full status machine (Draft → … → Expired/Cancelled) | Booking | MVP | The aha moment: one request, many quotes. |
| Quote comparison view | Booking | MVP | Comparison is the customer's payoff for structured quotes. |
| In-app messaging (text, images, PDF) | Communication | MVP | Negotiation must stay on-platform for evidence and retention. |
| Booking with deposit payment + status machine | Booking + Financial | MVP | Converts the marketplace from directory to transaction platform. |
| Payments: KNET, Visa/MC, Apple Pay, Google Pay via PSP abstraction | Financial | MVP | KNET-first is non-negotiable in Kuwait; abstraction keeps PSPs swappable. |
| Booking commission (5% of total) + basic vendor settlement | Financial | MVP | Second revenue stream from launch: commission is captured from the platform-collected deposit at confirmation and the vendor's share is settled. See Confirmed Decisions. |
| Refund requests (admin-mediated) | Booking + Operations | MVP | Money-back path is part of the confidence promise. |
| Reviews & ratings with moderation | Experience + Operations | MVP | Reputation is the marketplace's long-term moat; must start accruing at launch. |
| Notifications: email + push (FCM) | Communication | MVP | The waiting-for-quotes anxiety is managed by notifications. |
| Customer, vendor, and admin dashboards | Event/Vendor/Operations | MVP | Each side needs one home screen with its truth. |
| CMS basics (about, terms, privacy, FAQ, banners) | Operations | MVP | Legal pages and SEO landing content. |
| Audit logs | Platform | MVP | Disputes and admin accountability need history from the first day. |
| Arabic/English with full RTL | Platform | MVP | Arabic-first is brand identity, not a locale option. |
| Analytics event instrumentation | Growth | MVP | Can't steer Scale decisions without funnel data from day one. |
NOT in MVP — explicit exclusions¶
Written down so nobody relitigates them mid-build.
| Excluded feature | Why it waits | Arrives in |
|---|---|---|
| AI recommendations / AI quote drafting | AI needs behavioral and transaction data the MVP hasn't generated yet; premature AI = generic suggestions that erode trust. | Scale |
| Event planner suite (budget tracker, checklist, guest list, timeline) | Retention tooling matters after the transactional loop is proven; the MVP event dashboard is enough anchor. | Scale |
| Vendor calendar + Google Calendar sync | Real-time availability is high-effort and low-trust until vendors are active daily; MVP availability is handled in quotes/messaging. | Scale |
| Wallet | No balance-bearing money before volume justifies the regulatory and reconciliation burden. | Scale |
| WhatsApp notifications | Email + push covers MVP; WhatsApp Business API adds cost/approval cycles. | Scale |
| Promo codes & referral program | Growth mechanics amplify a working loop; they can't fix an unproven one. | Scale |
| KSA / UAE expansion | One market at a time; Kuwait proves the model (architecture is multi-country ready regardless). | Scale |
| Instant booking | Requires trustworthy real-time availability (calendar first). | Scale |
| Escrow & milestone payments | Heavy legal/financial machinery; the MVP deposit-plus-payout-schedule policy delivers most of the protection. | Vision |
| AR venue previews / 360° tours | Expensive content pipeline; zero effect on marketplace liquidity. | Vision |
| Seating planner, QR invitations, digital check-in, ticketing | Event-day tooling for a mature Event OS, not for proving the loop. | Vision |
| Vendor CRM | Vendors must first live in the dashboard for leads; CRM deepens an existing habit. | Vision |
| White-label & marketplace APIs | Ecosystem plays that need a dominant platform first. | Vision |
| Voice notes, video calls in chat | Text + attachments negotiate deals fine; heavy media infra later. | Scale/Vision |
| SMS notifications | Cost per message is high in GCC; push + email suffice until proven otherwise. | Scale |
Scale — after traction¶
Goal: turn a working Kuwait loop into a habit-forming platform and expand to KSA and UAE.
| Feature | Domain | Rationale |
|---|---|---|
| AI vendor recommendations | Intelligence | Now trained on real RFQ→booking outcomes; lifts match quality and conversion. |
| AI quote drafting for vendors | Intelligence | Cuts vendor response time — directly attacks the #1 customer drop-off (quote silence). |
| Event planner suite (budget, checklist, guest list, timeline) | Event | Retention: makes Yam3at the workspace between booking and event day. |
| Vendor calendar + Google Calendar sync | Vendor | Enables availability filtering and prepares instant booking. |
| Instant booking for fixed-price packages | Booking | Shortcuts the funnel for commodity services (e.g., photo booths). |
| Wallet (customer & vendor) | Financial | Refund credits, goodwill gestures, faster repeat checkout. |
| WhatsApp notifications | Communication | The GCC's default channel; big lift on quote-received open rates. |
| Promo codes & referral program | Growth + Experience | Growth mechanics on a proven loop. |
| Advanced vendor analytics | Vendor | Renewal driver: show vendors the revenue Yam3at generates. |
| Sponsored & featured listings | Marketplace | Paid placement revenue once search traffic and category density justify it (booking commission already ships in the MVP). |
| KSA launch, then UAE | Platform + all | New country config (currency, PSP routing, categories, Arabic dialect copy) — architecture ready, content and ops work real. |
| Support ticketing & fraud detection rules | Operations | Ops tooling scales with volume. |
| Team accounts & branches for vendors | Vendor + Identity | Larger vendors and franchises join once the platform is proven. |
Vision — the long-term Event OS¶
| Feature | Domain | Rationale |
|---|---|---|
| Escrow & milestone payments | Financial | Full payment protection for high-value weddings/corporate events. |
| AR venue previews & virtual walkthroughs | Marketplace | Preview the hall before visiting; differentiator at maturity. |
| Seating planner | Event | Deep planning tooling for large events. |
| QR invitations & digital guest check-in | Event | Owns the guest experience end to end. |
| Ticketing for public events | Event + Financial | Opens the public-events market (festivals, conferences). |
| Live event-day dashboard | Event | Command center during Execute. |
| Vendor CRM | Vendor | Vendors run their whole business on Yam3at. |
| White-label deployments | Platform | Hotels, government entities, large planners on Yam3at rails. |
| Marketplace & enterprise APIs | Platform | Partners ecosystem: integrations, corporate procurement. |
| Financing & insurance products | Financial | Event financing with partner institutions. |
Phased launch roadmap¶
Durations are indicative effort windows, not calendar promises. Each phase gates on exit criteria, not dates.
flowchart LR
P0["Phase 0<br/>Foundations<br/>~2–3 months"] --> P1["Phase 1<br/>MVP build<br/>~4–6 months"]
P1 --> P1b["Phase 1b<br/>Private beta<br/>~1–2 months"]
P1b --> L(["Kuwait public launch"])
L --> P2["Phase 2<br/>Scale: retention + AI<br/>~6–9 months"]
P2 --> P3["Phase 3<br/>GCC expansion<br/>KSA → UAE<br/>~6–9 months"]
P3 --> P4["Phase 4<br/>Vision<br/>ongoing"]| Phase | Focus | Key outputs | Exit criteria |
|---|---|---|---|
| 0 — Foundations (~2–3 mo) | Repo, CI/CD, modular-monolith skeleton along the 12 domains, auth, i18n/RTL scaffolding, PSP sandbox integration, design system tokens | Deployable walking skeleton; design library; seeded categories & locations for Kuwait | End-to-end deploy pipeline green; test KNET payment succeeds in sandbox; RTL renders correctly on web + Flutter shells |
| 1 — MVP build (~4–6 mo) | The full MVP table above, built loop-first: vendor onboarding → listings → RFQ → quote → booking → payment → review | Feature-complete MVP on web + mobile apps + admin | All MVP acceptance criteria in the SRS pass; QA plan executed |
| 1b — Private beta (~1–2 mo) | 30–50 hand-recruited vendors across priority categories; friendly customers run real events | Supply seeded; funnel instrumentation validated; pricing sanity-checked | Launch criteria checklist below fully satisfied |
| Kuwait public launch | Marketing switch-on per playbook | — | — |
| 2 — Scale (retention + AI) (~6–9 mo) | Planner suite, AI recommendations & quote drafting, calendar, WhatsApp, promos/referrals, wallet, advanced analytics | Habit loop beyond the transaction; second revenue streams piloted | Repeat-event rate and vendor renewal targets hit; unit economics per KPIs |
| 3 — GCC expansion (~6–9 mo) | KSA first (largest market), then UAE; country config, local PSP routing, local supply ops, dialect-aware content | Two new country launches on the same codebase | Kuwait playbook reproduced: liquidity in priority categories within target window per market |
| 4 — Vision (ongoing) | Escrow, AR, ticketing, CRM, white-label, APIs — sequenced by data, not by this document | The Event OS | — |
Sequencing principle
Supply before demand, always. Every market opens vendor-first: recruit and approve supply in priority categories (halls, catering, photography, decoration) before spending on customer acquisition. A customer who finds three quiet categories never returns; a vendor who waits two weeks for the first RFQ never renews. The beta phase exists to break this chicken-and-egg deliberately.
Definition of MVP done — launch criteria¶
MVP is done when every box below is checked. This list — not feature completeness alone — is the go/no-go for the Kuwait public launch.
Product completeness¶
- A customer can complete the entire loop on a real device in Arabic and English: signup → create event → RFQ to ≥3 vendors → receive quotes → compare → negotiate → book → pay deposit via live KNET → receive confirmation → leave a review
- A vendor can complete: register → upload license → get approved → subscribe & pay on a live PSP → publish catalog & gallery → receive RFQ → quote → win booking → see the deposit, the platform commission deducted, and the payout schedule for their settled share
- An admin can: approve/reject vendors with reasons, moderate reviews, resolve a refund request end to end, manage subscription plans and CMS pages
- All state machines enforce legal transitions only (quotation, booking, event) with audit trail
- Notifications fire on every critical moment: RFQ received, quote received, message, booking confirmed, payment result, review request
Quality & trust¶
- Full RTL correctness audited on the 20 highest-traffic screens (web + mobile) — no mirrored-icon or truncation defects
- KWD displayed with 3 decimal places everywhere money appears; totals reconcile with PSP records to the fils
- Payment failure paths tested: declined card, KNET timeout, double-submit, webhook replay — no orphan bookings, no double charges
- Security baseline per Security: rate limiting, OWASP top-10 checks, PII encryption at rest, role-based access verified per the SRS permission matrices
- P95 search and listing page loads < 2s on Kuwait mobile networks; crash-free sessions > 99.5% on both apps
Supply & operations¶
- ≥ 30–50 approved, subscribed vendors live across the priority categories, each with complete profiles (gallery ≥ 5 images, ≥ 1 priced service)
- Vendor approval SLA (≤ 48h) and dispute SLA (≤ 5 business days) staffed and rehearsed — at least one full dispute drill executed
- Terms of service, privacy policy, and cancellation/refund policy published in Arabic and English and reviewed by counsel
- Support channel live with bilingual coverage during launch hours
Instrumentation & readiness¶
- Funnel analytics verified end to end (event created → RFQ → quote → booking → payment) with dashboards for the journey health metrics
- Error monitoring, uptime alerting, and on-call rotation in place; rollback procedure tested
- App Store and Google Play listings approved (Arabic + English); production PSP account approved and reconciliation process documented
- Launch-week runbook written: seeded FAQs, canned support replies, escalation paths
When the last box is checked, Kuwait launches. Until then, no feature outside the MVP table gets built.