AI Forecasting

Our forecasting system automatically predicts how much of each product you'll sell in the coming months. The best part? It figures out the right approach for each product on its own—you don't have to configure anything.

What You Get

What You Need

For the best forecasts:

• At least 6 months of order history (for AI-powered forecasts)
• At least 12 months to detect seasonal patterns (like holiday spikes)
• Don't worry about data gaps—the system handles them automatically

How It Works

The system looks at your sales history and automatically chooses one of four methods:

1. Simple Average

What it does: Looks at your average sales and assumes you'll keep selling at that pace.

Perfect for:

• New products you just started selling
• Products with steady, predictable sales
• When you have less than 6 months of data

2. Trend Forecast

What it does: Spots whether a product is growing or declining, then predicts the continuation of that pattern.

Perfect for:

• New product launches that are taking off
• Products being phased out
• Anything with clear upward or downward momentum

Bonus: We add a 10% safety buffer for growing products so you don't run out of stock.

3. Seasonal Forecast

What it does: Compares this year to last year to predict seasonal spikes (like holiday sales).

Perfect for:

• Products that spike during holidays
• Weather-dependent items (umbrellas, sunscreen)
• Any product with a yearly pattern

4. AI-Powered Forecast 🤖

What it does: Uses advanced machine learning to spot complex patterns—like a product that's both growing AND has seasonal spikes.

What it looks at:

• Monthly patterns (is December always bigger than July?)
• Long-term trends (are sales growing over time?)
• Recent momentum (what happened the last few months?)
• Historical patterns (what's normal for this product?)

Perfect for:

• Products with complex, unpredictable patterns
• Items that have both growth and seasonal swings
• Anything that doesn't fit a simple pattern

When it kicks in:

• You have at least 6 months of order history
• The product shows complex behavior the simpler methods can't handle

Understanding Your Forecast

When you view a forecast, you'll see:

• Average Daily Demand: How many units you sell per day on average
• Lead Time Demand: How much you'll sell while waiting for a restock
• Target Stock Level: The ideal amount to have on hand
• Suggested Reorder Quantity: How many units to order right now

Your forecast chart shows:

• Blue bars: Your actual past sales
• Green bars: Predicted future sales

Tips for Success

✅ Do This

• Keep your data current – The system gets smarter with more recent data
• Run forecasts after big events – Had a huge sale or promotion? Update your forecast
• Check forecasts regularly – Review predictions monthly or before placing large orders
• Trust the suggestions – The reorder quantities are calculated to keep you stocked without overbuying

❌ Avoid This

• Don't delete old orders – Historical data makes forecasts more accurate
• Don't set unrealistic lead times – If your supplier takes 2 weeks, don't say 2 days
• Don't ignore the forecasts – If a product is predicted to spike, stock up in advance

What the Method Means

When you see which method was used for each product:

Common Questions

"The forecast seems way too high (or low)"

This could mean:

• Check your sales history for unusual spikes (like a one-time bulk order)
• Verify the method being used makes sense for this product
• Make sure you have enough historical data

"Why is my product using Simple Average?"

Usually because:

• You don't have 6+ months of order history yet
• There's not enough data to detect a pattern
• The product has very stable, consistent sales

"The AI forecast looks unpredictable"

Some products genuinely have erratic demand—that's normal! The AI does its best to predict them. Compare the forecast to your actual sales to see if it's tracking well.

Quick Reference

---

The bottom line: Keep your order data up to date, and the system will automatically figure out the best way to forecast each product. Then just follow the reorder suggestions and stay stocked!