Skip to main content

Current Features

Eaternity Forecast offers comprehensive demand forecasting capabilities for professional kitchens. This page details all currently available features.

Core Prediction Features​

Daily Demand Forecasting​

Description: AI-powered predictions for every menu item, every day

Capabilities:

  • Quantity predictions for each menu item
  • 7-day forecast window (next week's predictions)
  • 14-day extended forecast (for advanced planning)
  • Item-level granularity (individual dishes, not categories)
  • Service period breakdowns (lunch vs dinner predictions)

Accuracy:

  • Average MAPE: 12.8%
  • 25% better than human forecasters
  • Continuous improvement as model learns

Example Output:

{
  "date": "2024-01-20",
  "item": "Pasta Carbonara",
  "predicted_quantity": 52,
  "service_period": "lunch"
}

Confidence Intervals​

Description: Prediction ranges showing uncertainty and reliability

Every prediction includes an 80% confidence interval (10th to 90th percentile) to help you choose preparation strategies:

ComponentDescription
Point estimateMost likely quantity
Lower bound10% chance of selling fewer
Upper bound10% chance of selling more

Learn more about confidence intervals →

Historical Accuracy Tracking​

Description: Real-time monitoring of prediction performance

Metrics Provided:

  • Mean Absolute Percentage Error (MAPE) by item
  • Items within ±10% accuracy count
  • Large variance detection (>20% error)
  • Trend analysis (improving, stable, declining)

Time Periods:

  • Last 7 days accuracy
  • Last 30 days accuracy
  • All-time average accuracy

Dashboard View:

Item: Pasta Carbonara

Accuracy Metrics:
  Last 7 days:   95.2% (MAPE: 4.8%)
  Last 30 days:  94.2% (MAPE: 5.8%)
  All-time:      93.8% (MAPE: 6.2%)

Recent Performance:
  ✅ Within ±10%: 27 out of 30 days
  ⚠️ Large variance: 1 day (special event)

External Factor Integration​

Weather Integration​

Description: Automatic weather data integration for temperature-sensitive predictions

Data Sources:

  • Temperature (current and forecast)
  • Precipitation
  • Weather conditions (sunny, rainy, snowy, etc.)
  • Seasonal effects

Impact Examples:

  • Cold days → Higher soup and hot dish demand
  • Warm days → Higher salad and cold beverage demand
  • Rainy days → Lower overall volume (fewer walk-ins)
  • Extreme weather → Wider confidence intervals (uncertainty)

Coverage:

  • Automatic location detection based on kitchen address
  • 7-day weather forecast integrated
  • Historical weather data for pattern learning

Holiday and Event Detection​

Description: Automatic recognition of holidays and special events

Detected Events:

  • Public holidays (Christmas, Easter, New Year, etc.)
  • School holidays (for university/school cafeterias)
  • Local events (if data provided)
  • Day-of-week patterns (Fridays vs Mondays)

Effects:

  • Adjusted predictions for holiday patterns
  • Wider confidence intervals for irregular events
  • Historical holiday pattern learning

Manual Event Addition:

POST /v1/forecast/events
{
  "date": "2024-02-14",
  "name": "Valentine's Day Special Menu",
  "expected_impact": "high_volume",
  "notes": "Romantic dinner packages, expect 2x normal Saturday"
}

New Item Handling​

Description: Intelligent predictions for recently launched menu items

Approach:

  • Week 1: Low confidence, wide intervals (±30-40%)
  • Week 2-3: Rapid learning, confidence improving
  • Month 2+: Normal confidence, comparable to established items

Strategies:

  • Use similar item patterns as initial proxy
  • Conservative preparation recommendations
  • Accelerated learning from actual sales
  • Automatic confidence adjustment as data accumulates

Example Learning Curve:

New Item: Mushroom Risotto

Week 1: Prediction 22 (Range: 12-32, MAPE: 23%)
Week 2: Prediction 25 (Range: 18-32, MAPE: 17%)
Week 3: Prediction 27 (Range: 22-32, MAPE: 14%)
Week 4: Prediction 28 (Range: 24-32, MAPE: 11%)

Discontinued Item Detection​

Description: Automatic detection and handling of menu changes

Capabilities:

  • Automatic detection when item has zero sales for 7+ consecutive days
  • Graceful phaseout of predictions
  • Historical data preservation for future reference
  • Reactivation support if item returns to menu

Status Tracking:

Item: Summer Salad Special

Status: Discontinued
Last Sold: 2024-01-10
Reason: Zero sales for 14 consecutive days
Action: Predictions stopped
Reactivation: Available if item returns to menu

Seasonal Menu Changes​

Description: Support for rotating seasonal menus

Capabilities:

  • Learn seasonal patterns year-over-year
  • Detect menu transitions automatically
  • Adjust predictions for seasonal ingredients
  • Handle menu rotation cycles

Example:

Fall → Winter Menu Transition:

Phasing Out (Fall):
  - Summer Salad: Declining predictions detected
  - Grilled Vegetables: Demand trending down

Phasing In (Winter):
  - Butternut Squash Soup: New item, learning phase
  - Braised Short Ribs: Previous winter data applied

Reporting and Analytics​

Daily Performance Reports​

Description: Morning reports comparing yesterday's predictions to actuals

Includes:

  • Overall accuracy summary
  • Items within target accuracy (±10%)
  • Large variances requiring investigation
  • Top performing items
  • Items needing attention

Delivery Options:

  • Email (scheduled delivery)
  • Dashboard widget
  • Mobile app notification
  • API endpoint

Example Report:

Daily Performance Report
Date: January 19, 2024

Overall Accuracy: 91.2% ✅
Items within ±10%: 58 out of 65

Top Performers:
  1. Pasta Carbonara: 3.2% error
  2. Caesar Salad: 4.1% error
  3. House Bread: 2.8% error

Needs Review:
  - Grilled Salmon: 28% error (+8 portions)
    Possible cause: Unexpected warm weather

Weekly Forecast Summary​

Description: Comprehensive weekly outlook for planning

Includes:

  • 7-day predictions for all menu items
  • Expected busy vs quiet days
  • Recommended focus items for promotion
  • Suggested prep priorities

Use Cases:

  • Weekly buyer orders
  • Staff scheduling
  • Menu planning decisions
  • Promotional strategy

Format Options:

  • Excel export (editable planning sheets)
  • PDF report (print for kitchen)
  • API data (integration with planning tools)

Monthly Business Impact Report​

Description: Executive summary of forecast value

Metrics:

  • Food waste reduction (portions and cost)
  • Forecast accuracy trends
  • Cost savings (waste + time)
  • Service quality (stock-out tracking)
  • Environmental impact (CO₂e avoided)

Example Report:

Monthly Report: January 2024

Financial Impact:
  Food waste savings:        €1,045
  Time savings value:        €715
  Stock-out reduction:       €185
  Total value created:       €1,945

Operational Metrics:
  Average MAPE:              12.3%
  Waste rate:                7.1% (down from 12.8%)
  Stock-outs:                2 instances (down from 14)
  Time saved:                20 hours

Environmental Impact:
  CO₂e avoided:              285 kg
  Food waste prevented:      2,450 portions
  Water conserved:           32,000 liters

Trend: Improving (vs previous month)

Integration Features​

REST API​

Description: Full-featured API for system integration

Capabilities:

  • Submit sales data (daily or bulk)
  • Retrieve predictions (date range queries)
  • Override predictions manually
  • Access analytics and reports
  • Manage menu items and events

Authentication:

  • OAuth 2.0 (user-delegated access)
  • API keys (server-to-server)

Rate Limits:

  • 100 requests/minute
  • 10,000 requests/day
  • Custom limits available

View API Documentation →

Necta Integration​

Description: Native integration with Necta ERP

Features:

  • Automatic data sync from Necta sales
  • Native dashboard in Necta Planning module
  • Seamless workflow (no system switching)
  • One-click export to Excel from Necta

Exclusive Benefits:

  • Zero setup complexity
  • Automatic menu updates
  • Integrated recipe costing
  • Priority support

Learn more about Necta Integration →

Webhook Support​

Description: Real-time notifications for important events

Available Events:

  • predictions.generated — New forecasts ready
  • variance.large — Significant prediction error detected
  • data.quality_issue — Data submission problems
  • model.retrained — Model updated with new data

Use Cases:

  • Trigger automated workflows
  • Alert systems integration
  • Real-time dashboards
  • Inventory management systems

Example Webhook:

{
  "event": "predictions.generated",
  "timestamp": "2024-01-20T03:15:42Z",
  "data": {
    "date_range": {"start": "2024-01-20", "end": "2024-01-27"},
    "total_items": 65,
    "prediction_url": "/v1/forecast/predictions?date=2024-01-20"
  }
}

Export Capabilities​

Description: Multiple formats for data export

Formats:

  • Excel (.xlsx) with formatted tables and charts
  • CSV (comma-separated values) for import to other systems
  • PDF reports for printing and sharing
  • JSON via API for programmatic access

Export Scopes:

  • Single day all items
  • Date range specific items
  • Full weekly forecast
  • Historical accuracy reports

User Interface Features​

Dashboard​

Description: Web-based interface for non-technical users

Widgets:

  • Today's performance summary
  • Tomorrow's forecast
  • Weekly calendar view
  • Accuracy trends chart
  • Top items table

Customization:

  • Widget arrangement
  • Item filters
  • Date range selection
  • Metric preferences

Manual Override​

Description: Ability to manually adjust predictions

Use Cases:

  • Known events not in historical data
  • Special promotions
  • Supply disruptions
  • Operational constraints

Process:

  1. Select item and date
  2. Enter override quantity
  3. Provide reason (documented for learning)
  4. Optional: Adjust related items proportionally

Tracking:

  • Override history logged
  • Effectiveness measured
  • Reasons analyzed for model improvement

Example:

Override Prediction

Item: Pasta Carbonara
Date: Saturday, January 25
Original Prediction: 52 (48-56)

Override Quantity: 75
Reason: Conference booking (50 pax, 60% selecting pasta)

☑ Apply ratio adjustment to related items
  - Caesar Salad: 31 → 45
  - Tiramisu: 18 → 26

[Cancel] [Save Override]

Notification System​

Description: Alerts and reminders for key events

Notification Types:

  • Daily: Predictions ready
  • Weekly: Forecast summary available
  • Monthly: Performance report generated
  • Ad-hoc: Large variances, data quality issues

Delivery Channels:

  • Email
  • Dashboard notifications
  • Mobile push (coming Q3 2024)
  • Slack integration (coming Q3 2024)

Data Management Features​

Data Quality Monitoring​

Description: Automatic detection of data issues

Checks:

  • Completeness: Missing dates or items
  • Consistency: Item name variations
  • Accuracy: Outliers and anomalies
  • Timeliness: Delayed submissions

Quality Score:

Overall Data Quality: 92% ✅

Components:
  Completeness:  100% ✅ (no missing dates)
  Consistency:   89%  ✅ (minor naming issues)
  Accuracy:      95%  ✅ (few outliers)
  Timeliness:    85%  ✅ (some late submissions)

Issues:
  - 3 items with name variations (standardization recommended)
  - 2 late submissions in last 30 days

Historical Data Management​

Description: Tools for managing historical sales data

Capabilities:

  • Bulk import (CSV, Excel, JSON)
  • Data correction (fix errors in submitted data)
  • Item mapping (standardize inconsistent names)
  • Data export (backup and analysis)

Retention:

  • Minimum 90 days active (for training)
  • Maximum 365 days initially (expandable)
  • Archival storage beyond 365 days

Privacy and Security​

Description: Data protection and access control

Features:

  • Encryption in transit (TLS 1.2+) and at rest
  • Access control (role-based permissions)
  • Audit logging (all API access tracked)
  • Data isolation (your data never mixed with others)

Compliance:

  • GDPR compliant
  • Data processing agreement available
  • Right to data deletion
  • Data portability support

Platform Features​

Multi-Location Support​

Description: Manage forecasts for multiple kitchens

Capabilities:

  • Separate model per location
  • Shared learning across locations (optional)
  • Consolidated reporting
  • Location-specific configurations

Use Cases:

  • Restaurant chains
  • Corporate catering (multiple sites)
  • Hospital networks
  • University dining halls

Administration:

  • Centralized billing
  • Unified administration

For pricing information, visit eaternity.org/pricing.

User Management​

Description: Team access and permission control

Roles:

  • Admin: Full access, billing, user management
  • Manager: View predictions, override, reports
  • Kitchen Staff: View predictions only
  • Read-only: Reports and analytics only

Features:

  • Invite team members by email
  • Assign roles and permissions
  • Activity logging
  • Single sign-on (SSO) coming Q4 2024

Model Retraining​

Description: Continuous model improvement

Frequency:

  • Weekly retraining (every Monday at 4 AM)
  • Ad-hoc retraining (after significant menu changes)

Process:

  • Incorporate latest sales data
  • Optimize model parameters
  • Validate accuracy improvements
  • Deploy if performance improves

Notifications:

  • Email summary after retraining
  • Accuracy comparison (before/after)
  • New baseline established

Example:

Model Retrained Successfully

Date: Monday, January 22, 2024
Training Data: October 1, 2023 - January 19, 2024

Improvements:
  Previous MAPE: 13.3%
  New MAPE:      12.1%
  Improvement:   +1.2%

Items Improved:
  - Grilled Salmon: 15.2% → 12.8%
  - Daily Specials: 19.5% → 16.3%

Deployed: Yes
Status: Active

Support Features​

Documentation​

Description: Comprehensive guides and references

Available:

  • Getting started guides
  • API reference documentation
  • Implementation best practices
  • Troubleshooting guides
  • Video tutorials

Formats:

  • Online documentation (this site)
  • PDF downloads
  • Video tutorials
  • Interactive examples

Technical Support​

Description: Expert assistance for users

Channels:

  • Email support
  • Dashboard help center
  • API documentation
  • Phone support (premium tier)

Response Times:

  • Critical: 4 hours
  • High: 24 hours
  • Medium: 48 hours
  • Low: 1 week

Included Support:

  • Integration assistance
  • Data quality consultation
  • Accuracy optimization guidance
  • Feature usage training

Feature Requests​

Description: Community-driven product development

Process:

  1. Submit feature request via dashboard or email
  2. Eaternity team reviews and prioritizes
  3. Community voting (customers)
  4. Development roadmap updated
  5. Notification when feature ships

Current Top Requests (in development):

  • Orbisk visual waste tracking integration
  • Multi-location POS system support
  • Recipe-based ingredient forecasting
  • Advanced analytics dashboard

View Full Roadmap →

See Also​