FoodFlow Home — Research (P2-1) — Prism main line
Product #2, Stage: RESEARCH. Sources: the 4 verified FoodFlow strategy docs (distilled via agent read) + web research (dual-persona landing patterns, 2026 AI-SEO/structured data). Note: a parallel session's artifacts are preserved untouched in
../_archive-parallel-session/per the CD's directive (2026-06-13); this file is the canonical research of the main line.
1. Positioning (verified, competitive research 2026-06-09)
"FoodFlow is the first multi-persona platform that closes the loop from professional prescription → client execution → grocery cart → compliance reporting, localized in Hebrew and integrated with Israeli retailers."
- The gap is real but conditional: Practice Better / Healthie dominate B2B nutrition but have zero meal-planning of their own, zero grocery integration, zero Hebrew. No Israeli-local competitor. The future moat is grocery integration; today's claimable moat is Hebrew-first.
- Working metaphor: "Shopify for nutritionists."
2. Personas
| Persona | Pain | Pays |
|---|---|---|
| Ravit — nutritionist ~15 clients | PDFs over email, WhatsApp chaos, zero follow-through visibility | ₪99–149/mo |
| Shelly — clinic, 20–50 clients | + intake forms, compliance, multi-practitioner | ₪200–500/mo |
| Ben-El — trainer (later wave) | nutrition beside workouts | ₪150–250/mo |
| Home cook / client | "what to cook today", shopping list, following the plan | mostly via professional |
CD decision (gate, 2026-06-12): the homepage serves nutritionists AND home cooks as equals — two doors. Mitigation for the known split-hero conversion risk: one shared promise line + immediate persona segmentation within the first viewport; each door opens a persona-complete journey. SEO still leans B2B on practitioner keywords via dedicated sections.
3. Landing patterns (2026)
- One persona per message; multi-audience pages win via persona-based pathways and progressive disclosure — not a diluted shared hero.
- Dual CTA pattern: "Try free" (self-serve) + "See how it works" (guided). Real product screenshots outperform illustration.
4. SEO / AI-SEO (2026)
- Structured data = AI citability layer: JSON-LD (Organization, SoftwareApplication, FAQ; Recipe when marketplace content is indexable) determines AI Overview/Perplexity citation.
- Hebrew keyword vacuum is the opportunity: "תוכנה לדיאטניות", "ניהול מטופלים בתזונה", "תוכניות תזונה דיגיטליות", "אפליקציית תכנון ארוחות" — no local SaaS competition; first indexed defines the category. English secondary ("nutrition practice management", "Practice Better alternative" — future comparison page).
- hreflang he⇄en; RTL must be flawless — "RTL bugs destroy trust immediately" (strategy doc).
5. Honesty constraints (must NOT overclaim)
- ❌ "Seamless grocery cart / Rami Levy integration" (unofficial API) → ✅ "smart shopping lists linked to Israeli grocery stores"
- ❌ nutritional analysis/macros (NULL until Phase 2) → ✅ "track your meals"
- ❌ "3,000–5,000 nutritionists" (unverified) → ✅ "a fragmented market with no local solution"
- ❌ precise compliance measurement → ✅ "help your clients stick to their plan"
- ✅ Safe: Hebrew-first RTL-native · digital plan delivery · shopping lists from plans · client portal · AI recipe import (deployed) · household sharing
6. Brand seeds (input to BRAND)
Orange #FF6B35 / teal #4ECDC4, warm kitchen tone, Rubik-class Hebrew-friendly type,
"פודפלו". Professional voice for door 1 ("simple, organized, time-saving"); action voice for
door 2 ("Today's meal is ready"). Wolt = local gold standard for Hebrew-first product design.
Naturally divergent from ben-portfolio (warm/light/food vs dark/indigo/film) ✓.