admin-baked/skills/dispensary/menu-gap-analysis

Menu Gap Analysis

Purpose

Surface the delta between what a dispensary currently stocks and what the local market demands — so operators can make confident procurement decisions and stop losing customers to competitors carrying products they don't have.

When to Use

• Operator asks what competitors carry that the store doesn't
• Weekly competitive intel refresh surfaces product gaps
• Smokey repeatedly fails to match customer requests to in-stock items
• Seasonal reset (spring, fall, 4/20) warrants menu evaluation
• New brand or product category is gaining market share locally

When NOT to Use

• Pricing strategy → Money Mike owns margin/price decisions; gap analysis flags the product, not the price
• Campaign drafting for a new product → hand off to Craig after gap is identified
• Real-time inventory management → POS system handles stock levels; this skill analyzes category gaps
• Compliance review of new products → Deebo checks label claims; gap analysis doesn't evaluate legality
• National trend analysis without local signal → trending nationally ≠ gap locally; require local evidence

Required Inputs

• org_id — required
• current_menu — current store inventory snapshot (pulled from POS or searchMenu())
• location — city/state for competitor scan radius
• categories — optional; default: flower, vapes, gummies, prerolls, beverages, concentrates, topicals
• competitor_data — optional; uses cached CI report or triggers fresh scanCompetitors() if absent

Reasoning Approach

A menu gap is only meaningful if local demand exists for the missing product. Three signals confirm demand:

1. Competitor velocity signal Products appearing on 2+ competitor menus with visible popularity (review count, featured placement, "staff pick") indicate local demand. One competitor carrying something = weak signal. Three = strong.

2. Customer request signal Smokey delegation logs, missed product searches, and "do you carry X?" queries from customers are direct demand evidence. Weight these heavily — customers told you what they want.

3. Category trend signal Category-level shifts (beverage cannabis growing, flower declining nationally) provide context, but only act on these if local evidence supports them. Label as "trend" not "gap" if unconfirmed locally.

Gap scoring:

• Priority 1: Present on 3+ competitor menus + customer requests logged → strong procurement case
• Priority 2: Present on 2 competitor menus OR customer requests with no competitor confirmation
• Priority 3: Category trend only, no local confirmation → monitor, don't procure yet

What to NOT flag:

• Products in legal gray areas for NY
• Products the store has consciously chosen not to carry (check brand guide / operator notes)
• Temporary out-of-stocks on products already in the catalog → that's a restock, not a gap

Output Contract

## Menu Gap Analysis — [Store] — [Date]

DATA SOURCES: [CI report dated X / live scan / customer request logs]
COMPETITORS ANALYZED: N
CATEGORIES REVIEWED: [list]

### Priority 1 Gaps (Strong procurement case)
| Product / Category | Seen At | Customer Requests | Estimated Weekly Demand |
|-------------------|---------|------------------|------------------------|
| ...               | 3 competitors | 5 logged requests | ~$800/wk est. |

### Priority 2 Gaps (Investigate further)
| Product / Category | Signal | Confidence | Suggested Action |
|-------------------|--------|-----------|-----------------|

### Priority 3 Watch (Trend only — no local confirmation)
| Trend | Evidence | Monitor Until |
|-------|---------|--------------|

### Recommended Actions
| # | Action | Owner | Timeline |
|---|--------|-------|----------|
| 1 | Add [product] — strong local demand, 3 competitors carry it | Operator | This week |

### What's Working
[1–2 sentences on categories where the store's menu is strong or differentiated vs. competitors]

Edge Cases

• No competitor data available: State clearly — do not fabricate gaps. Run scanCompetitors() first.
• Product exists but is regularly out of stock: Flag as restock gap, not procurement gap — different action
• Customer requests for a prohibited product: Do not recommend procurement; route to Deebo for guidance
• Competing menus scraped but data is stale (>7 days): Flag confidence as Medium; recommend rescan
• Store deliberately doesn't carry a category (e.g., no concentrates): Respect that constraint; don't flag it as a gap

Escalation Rules

• Gap product is in a legally ambiguous status in NY: Route to Deebo before any procurement recommendation
• Priority 1 gap worth >$2,000/week estimated revenue: Escalate to super_user for visibility alongside operator
• Customer requests logged for same product >10 times: Flag as Immediate — revenue is actively being lost now

Compliance Notes

• Only recommend products legal under NY OCM rules
• Do not recommend products based on competitor data alone if NY licensing status is unclear
• Demand estimates are approximations based on competitor velocity and request logs — label as estimates
• Gap analysis does not constitute procurement approval — operator makes the final sourcing decision