TrackingMore website currently supports more than 600 couriers. Many couriers only provide logistics trajectory information, not logistics status. However, trackingmore still successfully divide all logistics trajectory into seven unified situations. How does TrackingMore manage to do it?
First, let’s see which are the seven logistics status classified by TrackingMore:
Not Found: None information on this tracking number
Transit: The package is in the transit
Out for Delivery: The courier is on the way to deliver parcels.
Failed Attempt: The courier was unable to deliver parcel due to unexpected reason.
Delivered: The parcel is delivered.
Exception: The parcel may be lost or damaged.
Expired: The parcel is not delivered for a long time.
One. Setup Keywords Libraries.
To automatically identify these logistics status, we need to make corresponding keywords libraries for each situation.
For example, when the word “delivery” appears in the USPS logistics information, we know that the parcel has been signed by the consignee; and when the keyword “unsuccessful delivery” appears, we can determine that the package is failed to delivery.
Two. Define priority of Each Status
Exception > Delivered > Delivery Failure > Failed Attempt > Transit
Although we already setup the keyword libraries, the keywords “waiting for delivery” and “successful signed by consignee” will show up in logistics information. So we will recognize the express delivery as “successful signed by consignee” according to the priority principle.
1. How to distinguish the returned parcel from the normal one?
Returned parcel’s logistics information often appears “customer rejection”, “return to sender” and other words, but also appear “out for delivery” and other words at the same time, we will identify the single as “maybe abnormal”, and then distinguish from the usual parcels.
If there are no keywords like “returned sender” in the single number, but it is indeed returned, we can judge the status according to the geographical location, such as the origin of the tracking number is “A street, Queen, New York”, and the signing place is “A street, Queen, New York”, then the express delivery is likely to be returned.
For international express delivery, if the original country is China and the destination is the United States, but the final place of signature and receipt is China or the words like “进口拆封” appear in the information section of China Post, then the express delivery must be a returned parcel.
2. The influence of misleading keywords
For example, “delivery” means package is signed by consignee, but “will be delivered in 3 days” means not delivered yet; and “return to sender” means return, but “will return to sender if failed to deliver within 3days” corresponds to the status of “Out for Delivery” rather than return to sender.
3. Keyword Recognition in Multilingualism
For international express, as there may be different languages in the logistics information, it is necessary to updates the keyword Libraries with different languages in order to recognize the status correctly.
For example, “投妥”, “delivered”, “zugestellt”, “お届け済み”, “ENTREGADO”, “Отправление доставлено” are Chinese, English, German, Spanish and Russian respectively.
4. Differences of Express company
Because the logistics information of different express companies is not uniform, so some times the same keyword mean different status in different express companies, so it needs to be amended separately for specific express companies.
For example, “Received” maybe mean that the courier already receives the package in some express company, but in some other express company, it means the consignee already receive the package.
In order to recognize the status of express delivery successfully, first of all, we should have enough data to establish a big vocabulary database, so that we can recognize the state automatically and accurately.
Secondly, apply the newest AI algorithm. Hundreds of express company each have a different expression. Why can an experienced man recognize the parcel’s status with only one glance? Our algorithm must think as a real man.
Ps: After three years of accumulation and trial-and-error, TrackingMore platform has achieved the best status-recognition function in the industry. If you find some tracking number still has a problem, please contact us.