FourKites — a supply chain visibility platform — says it is able to predict freight arrival times more accurately than ever before thanks to a new DynamicETA algorithm.

The new approach draws on more than 150 data points associated with a single load, and draws other insights from millions of monthly shipments. The focus includes such things as shipper, carrier, lane, rest patterns, loads, traffic and weather.

The company says the new feature helped one food and grocery wholesaler to predict arrival times within a one-hour window 91% of the time, and within a six-hour window 97% of the time. And that involved more than 30,000 weekly deliveries.

FourKites says traditional accuracy rates are closer to 45% for day-of arrivals, largely because data is limited to things like standard drive times on a projected route and remaining hours of service.