• No results found

Big Data, Analytics, and IoT in Cross-Border Ecommerce

N/A
N/A
Protected

Academic year: 2021

Share "Big Data, Analytics, and IoT in Cross-Border Ecommerce"

Copied!
13
0
0

Loading.... (view fulltext now)

Full text

(1)

Big Data, Analytics, and IoT in

Cross-Border Ecommerce

James A Fairweather, PhD

Vice President, PB Technology, Experience and Ecommerce

(2)

How we enable transactions in commerce.

Our solutions help businesses manage customer information and turn it into data-driven insights

We help companies convert geographic data into better

customer experiences and smarter strategic decisions

We help transform the way our clients

engage with their customers – driving consistent interactions and targeted, orchestrated connections We help clients

increase the speed of the mail process while cutting costs

We remove barriers to cross-border shipping

(3)

We remove barriers to

cross-border shipping.

By uniquely guaranteeing the cost of

international shipping, our Global Ecommerce solutions make it easier for online retailers to ship parcels and packages to anyone in the world without surprises. We do this by:

Guaranteeing shipping costs for every single transaction

Removing the guesswork when calculating foreign taxes

and fees

Studying every aspect of customs processes to simplify

cross-border complexity

Empowering customers worldwide to easily and efficiently

track their parcels and packages no matter where they are

3 SOFTWARE SERVICES TRANSPORTATION SERVICES API / WEB INTEGRATION Product  Classification • Catalog  harmonization • Machine  learning Product Goods  Management • Import / export  compliance • Hazmat &  dangerous  goods Fully landed  costs • Duties &  taxes • Shipping Hub Operations • International  shipping lables • Commercial  invoice  preparation Carrier  Management • Organic,  optimized  carrier network • International  line‐haul Parcel Tracking • Consistent &  reliable tracking  visibility • API integration &  tracking portal

(4)

OUR

SITUATION

Executives received static, canned reports with limited information. Some analysis was done by populating manually in Excel

spreadsheets.

Operations teams and

product team heavily relied on reports for day to day operations.

Lot of manual efforts spent in analyzing the reports to drive

actionable insights.

COMPLICATIONS

Multiple data sources No centralized storage High volumes

KEY QUESTIONS

How can we integrate disparate data into a single view? How do we make available to multiple stakeholders?

DESIRED STATE

Centralized data store

KPI dashboard via Web, mobile & notifications Interactive, self serving drill down

Looking for insight into parcel status, SLA achievement, hub performance and more…

(5)

Parcel Process Flow

Parcels are transported to the PB Hub (single parcel or freight)

Inbound to PB hub:

• Retailer/shipper prepares parcels • ASN transmitted to PB

• PB generates a Universal Parcel Identifier (UPID)

• Retailer/shipper sends parcels

Parcel is received at the PB Hub

Goods are cleared through foreign customs via a PB Broker

PB coordinates with a contracted agent or a

commercial carrier, depending upon the destination country

PB coordinates and consolidates parcels for international line-haul

PB prepares the parcels for International line-haul

--• HS Classification

• Export Compliance (ECCN) review • Import Restrictions Screening review • Content Inspection

• Commercial Invoices • TSA Screening

(6)

Tracking Event Aggregation & Architecture

Belt scan, weigh and  dimension Ring scanners Central Parcel  DB Hub Operations Software Hub application, Hub application database Carrier 1 Carrier 2 Carrier 3 Carrier 4 Carrier 5 Carrier 6 Carrier 7 Carrier 8 Carrier 9 Tracking Data FTP, HTTP/XML, Web Services Carrier Tracking Services Parcel API, Adapters Reporting (operational,  analytics, etc.)

(7)

Strategic foundational technologies

Solution/ Business Outcome Physical SW/HW Products Digital (Web)

Convergence: Broader Digitization (People, Machines, Things, Logistics) Digital (Mobile) Evolution of Industry Technologies How  Technology  is  Consumed Information SaaS SaaS Mobile Mobile APIm APIm Big Data Big Data Machine Learning & AI Machine Learning & AI Analytics Analytics M2M / IoT M2M / IoT 7 Pitney Bowes | January 14, 2015

(8)

Drive Fact Dimensions to KPI Dashboards

APPLICATION 

DATA SOURCES DATA WAREHOUSE KPI DASHBOARDS

INSIGHTS

Carrier data

(9)

Data System Architecture

9 Pitney Bowes | January 14, 2015

(10)

Insights from Dashboards

• Dashboard of “In-transit,

shipped forward from hub”

• Current status within four SLA

categories

o Delivered In SLA

o Delivered Out of SLA

o Undelivered

o Undelivered, but still within

SLA

• Trends for the same over prior

months

• Can be filtered in real time on

partner, which outbound hub, destination country, and

(11)

Learnings

1. http://www.nytimes.com/2014/08/18/technology/for-big-data-scientists-hurdle-to-insights-is-janitor-work.html

11

“Janitor Work”1 in data cleanup and

aggregation

Data de-normalization and duplicate data resolution across application data systems Incremental data feeds from existing

systems (RDBMS) were challenging

Understanding of Hadoop cluster capacity planning and resource management within a shared infrastructure

We quickly adopted patterns to abstract the data ingestion layer, to accommodate ongoing changes in source data schemas

(12)

Conclusions

We have deployed a Hadoop based production big data solution based on a data lake architecture to integrate data from multiple data

sources, and built a data warehouse.

Data is made available at centralized store for responsive Dashboard queries and batch/real-time analytics.

We have built a series of web accessible Dashboards for Business and Operations teams to monitor key KPI’s (parcel status, SLA

achievement, hub efficiency) and make informed operating decisions. Dashboard users can interact and have self-service drill down

capability.

Data warehouse definitions are published so additional advanced analytics can be developed by our analysts as required.

Integrated Pitney Bowes Location Intelligence capabilities to cleanse and geocode global addresses, perform demographic segmentation and derive location based insights.

(13)

Thank you

James A Fairweather, PhD

[email protected]

References

Related documents

Census data was compared at three different points in the processing cycle: at Load (after the data had been captured and coded), after Edit and Imputation (where missing values

Players can create characters and participate in any adventure allowed as a part of the D&D Adventurers League.. As they adventure, players track their characters’

In- terestingly, both the removal of haem from haemopexin and HasA addi- tionally use steric hindrance to displace the haem group from its high- af finity binding site, while no

[r]

Whether grown as freestanding trees or wall- trained fans, established figs should be lightly pruned twice a year: once in spring to thin out old or damaged wood and to maintain

In conclusion, for the studied Taiwanese population of diabetic patients undergoing hemodialysis, increased mortality rates are associated with higher average FPG levels at 1 and

This course covers the major elements of logistics management including gaining competitive advantages through logistics and supply chain, the customer service dimension of logistics,

In order to capture the stochastic- ity of such retarded systems, stochastic differential delay equations driven by the standard Brownian motion have been proposed and