The methodology behind your inventory intelligence
Orderacle transforms raw POS exports into actionable inventory intelligence through a multi-step pipeline:
| Field | What We Use It For |
|---|---|
Date + Time |
Velocity calculations, day-of-week patterns, seasonality |
Item |
Product identification, name fragmentation detection |
Category |
Category-level P&L, margin analysis |
Qty + Net Sales |
Velocity, revenue attribution, stockout detection |
Customer Name |
Customer concentration analysis, whale identification |
Discounts |
Markdown tracking, promotion effectiveness |
One of Orderacle's most valuable insights: detecting when you ran out of stock and missed sales. Here's how it works:
If a product was selling consistently, then suddenly stops for 2+ weeks, it's probably a supply gap — not a demand change. Customers wanted it; you didn't have it.
For each product, we calculate average units sold per week over the analysis period.
We only look at products that were selling at least 1.5 units/week consistently. Low-velocity items have too much noise.
Find products that went from active sales → zero sales for 2+ consecutive weeks → then either came back or stayed at zero.
Gap weeks × pre-gap velocity × average selling price = estimated missed revenue.
| Metric | Value |
|---|---|
| Pre-stockout velocity | 2.9 units/week |
| Stockout date | December 10, 2024 |
| Restock date | Never (through Dec 31) |
| Gap duration | 3 weeks |
| Avg selling price | $65 |
| Estimated lost revenue | $566 |
Key insight: December 10-31 is peak holiday gifting season. Wingspan is a top gift game. This stockout happened at the worst possible time.
POS systems like Square don't enforce consistent naming. Your staff might enter the same product three different ways. This destroys your velocity data.
You think you sold 15 "Dragon Shield Sleeves" and 12 "DS Sleeves" and 8 "Dragon Shield 100ct." Your top-seller report shows three medium sellers. Reality: you sold 35 of the same product — it's actually your #1 supply item.
Convert to lowercase, expand abbreviations (MTG → Magic, PKM → Pokemon, DS → Dragon Shield, TTR → Ticket to Ride, etc.)
All POS entries that normalize to the same canonical name get grouped together.
If a normalized name has 2+ different raw POS entries, it's fragmented.
// These all normalize to "pokemon surging sparks elite trainer box"
"Pokemon Surging Sparks ETB"
"PKM Surging Sparks Elite Trainer Box"
"PKMN SS ETB"
// These all normalize to "magic commander deck"
"MTG Commander Deck"
"Magic Commander Deck"
"MTG CMD Deck"POS exports include revenue, but not cost. To calculate margins, we need Cost of Goods Sold (COGS) for each product.
We maintain a mapping of product names to wholesale costs. For each POS line item, we look up the cost.
For MISC entries or products not in our lookup, we estimate 50% margin (conservative).
Sum revenue and COGS by category to calculate category-level margins.
| Category | Typical Margin | Why |
|---|---|---|
| TCG Sealed | 15-25% | MAP pricing, high competition, low distributor markup |
| TCG Singles | 50-60% | Buylist at 40-60% of market, sell at market |
| Board Games | 45-50% | Standard retail keystone |
| Miniatures (GW) | 45% | MAP pricing but better margins than TCG |
| RPG Books | 50-55% | Standard book distribution |
| Supplies | 50-60% | High velocity, good wholesale pricing |
| Events | 70-80% | Mostly labor cost, prizing is marketing expense |
TCG Sealed drives ~34% of a typical LGS's revenue but only ~21% margin. It feels like your bread and butter, but every $1 of sealed generates only $0.21 gross profit vs $0.50+ for board games. Push high-margin add-ons (supplies, events) to offset thin sealed margins.
The demo uses simulated data for "Dragon's Hoard Games," a fictional mid-tier LGS. Here's what we assumed:
| Annual Revenue | ~$650K |
| Location | College town, mid-market |
| Size | 1,800 sq ft retail + 600 sq ft play space |
| Analysis Period | Q4 2024 (Oct–Dec) |
When you connect your actual store data, here's what happens:
| Square | Supported | CSV export or API |
| Lightspeed | Planned | CSV export |
| Clover | Planned | CSV export |
The full Orderacle vision includes demand signals from:
A CSV export from your POS system covering at least 3 months. We need: Date, Time, Item Name, Category, Qty, Net Sales, Customer Name (if available).
A spreadsheet of your products and wholesale costs. Without this, we estimate margins based on category averages.
If you want community signal analysis, we'd need read access to your store's Discord.