Predictive Delivery Windows, Even Before Shipping Begins
Predict uses historical delivery data to predict delivery time windows well in advance. Delivery service providers can notify their customers and consignees days before delivery, or display a reliable 2–4 hour window directly in the webshop. Predict gives customers true planability, reduces failed deliveries, and creates a real competitive advantage: our early delivery windows match actual delivery in roughly 80% of cases, with an average deviation of about 15 minutes.
Predict delivers predictive time windows through a simple API, which makes it easy to embed these predictions into checkout flows and customer notifications.
Predict supports both synchronous and bulk (all shipments) integrations. A request to the endpoint returns a delivery time window for the given address.
- Input: Address & estimated delivery date
- Output: Predictive time window (e.g. 9:00:00 – 11:00:00) for delivery
- Based on historical delivery patterns
- Broad coverage for most addresses
- Optimized for pre-arrival communication (not real-time)
- Provides delivery windows when real-time tracking data is unavailable
Built for: pre-arrival communication and planning, including early notifications, checkout messaging, and preferred time-slot selection.
Not for: real-time tracking on delivery day, or route and resource planning.
Use instead: for live updates once a parcel is out for delivery, use Better Tracking Shadow or Orchestrate. For route or resource planning, see Better Planner.
To calculate a Predict window, Bettermile requires historical data on all previous delivery events within your delivery network, as well as a live data feed with ongoing delivery event updates.
- Minimal viable data: Delivery address (or geolocation) and delivery timestamp
- For initial setup: Delivery events from the last 6 weeks
Most customers share their data with Bettermile through our secure Data Gateway (see detailed information here).
For an initial proof of concept, we can also process a sample set of historical data. To explore a proof of concept, contact us at customer@bettermile.com.