Skip to main content

Five Challenges of Managing Big Data in Supply Chains

By September 25, 2011Article

Logistics and supply chains are the richest data domains around. With data doubling every 18 months, how can companies manage the scale, quality and security of the data in the supply chain?
The rise of complex and global business networks means that a majority of the data will be generated outside a company’s firewall. The data of record will no longer be in a single company’s database. Traditional ERP systems and silo’ed IT infrastructure and systems will no longer have all the answers. The new economics of global commerce will force companies to move beyond their mission-critical systems of record towards global collaboration and innovation platforms, which enable business networks and partner ecosystems.
As we enter into this brave new world, businesses will not compete company to company but value chain to value chain. The best companies in the future will need to learn how to be a business network. Unfortunately, the current IT infrastructure and applications are woefully inadequate for this task. In this article, I examine the challenges companies face as they transition to a business network in the context of the supply chain and how and why cloud-based solutions are the only way to deliver on that promise.
Companies face the following five challenges as they manage big global data in the supply chain and strive to improve collaboration, automate processes and increase the visibility and efficiency of supply chains.
1. Eighty percent of the supply chain data set is outside the enterprise
“Inter-company commerce data is big and rich,” says Greg Johnsen, executive vice president of marketing and co-founder of GT Nexus, a cloud-based logistics and supply-chain platform company that hosts more than 15,000 organizations on its platform. “But only 20 percent of a supply-chain data set is internal and 80 percent is contributed by external partners,” adds Johnsen. “All of these data transactions are distributed across multiple enterprise systems in different companies with no easy way to determine the single version of the truth. What has been missing is a B2B inter-company infrastructure.”
Many large manufacturers and retailers such as Caterpillar and Home Depot need to collaborate with hundreds of partners and logistic service providers to track products, run purchase-to-pay programs, and automate myriad supply chain processes. However, without access to 80 percent of the supply-chain execution data, companies are starved for a detailed transaction record of the end-to-end life cycle of a transaction as it executes across multiple partners and suppliers.
2. Most companies are supply-chain data blind
Because 80 percent of the data is outside their own ERP systems of record, most companies are supply-chain data blind. Not surprisingly, therefore, variability is super common in many supply chains. While 95 percent of the orders get to their destinations more or less on time, the remaining uncertainty means companies have to carry more inventories. Companies that are global — which is where pretty much every industry has gone — have especially long lead times. These are complex and risky supply chains. This is the reason why companies find it difficult to determine with any level of certainty the root cause of that variability and thus are unable to reduce the cost of carrying excess
inventory.
Consider this all-too-common situation: Ninety percent of all goods and products in a supply chain move in containers today. The air freight or ocean carrier can tell you where this container is, but they can’t tell you what line items of a purchase order are in that container. The forwarder can tell you what line items are in that container, but they can’t tell you where the container is. The manufacturer knows the line items in the PO, but they can’t tell you which line items are in which container and what’s left in the PO that needs to go in another shipment.
As a result, companies lack a unified view of how the product is moving through the supply chain. They need to be able to triangulate across multiple constituencies. With data distributed across multiple platforms and ERP systems across many partners, it’s difficult, if not impossible, to get a unified view of the supply-chain truth.
3. Supply chains in different industries use radically different models
The supply chain of a major electronics manufacturer is very different from the supply chain of an apparel/footwear company. They have very different drivers, business models and business rules.
An apparel manufacturer may care a lot about what happens at origin in Asia as suppliers are building out orders in cartons. The manufacturer might want to tightly monitor these operations at origin and keep a close eye on metrics such as the time the order is  received / accepted / acknowledged, time for preparing the shipments and hitting the ship window, time to coordinate with a third-party logistics provider, etc.
On the other hand, a large heavy equipment manufacture cares a lot about outbound fulfillment of a major tractor shipment from a U.S. plant to a dealer in Brazil.
However, both these different supply chains use a common service provider community including truckers, carriers, customs brokers and consolidators. Because of the silo’ed nature of IT infrastructure, the benefits of fixing a data feed for one customer in one industry does not flow to another industry even though the data feed is common across both industries.
4. Garbage in, garbage out
The quality of data feeds in supply chains is a significant problem. There are two fundamental issues:

  1. Common data standards don’t exist. While most data is structured, much of it is proprietary and comes in a variety of formats including XML, flat files and spreadsheets. For example, one carrier may encode a shipment status message of a location like Hong Kong one way (HK,CN) and the next carrier in the very same industry and even operating out of the same country may encode it differently (CN/HK).
  2. The data source origins are not reliable across all customers. “The EDI file (shipment status feed) for shipment statuses hase been around for 35 years,” says Johnsen. “But as recently as four years, if you ask the question, over a course of a single shipment, ‘what percentage of necessary milestones did we receive,’ the industry-wide answer would be: ‘the average was 54 percent!'” In most cases, the issues are related to origin: the system simply didn’t send the signal.

There is lot of excitement around collecting consumer data at the point of sale and marrying that with profile information and predicting demand better. Enterprises have an equally strategic opportunity to do that within their own supply chains. If they have a quality data set that covers the transactional details of order fulfillment or the purchase-to-pay process, they can begin to glean actionable insights such as how their supply chain is operating, what they need to do to change it, whether they are sourcing in the right areas, whether they are working with the right providers, etc.
5. Lack of common platform for collaboration and building communities
With their current disparate systems and lack of a unified informational model, companies cannot enable massively scalable information-sharing or build collaborative communities across their network of partners.
Once communities are put together, they can participate in sharing best practices to improve data quality. They can start a blog, post a comment, and send a signal. This allows participants along the supply chain to collaborate around business topics.
Johnsen mentioned a shipper council (which collectively represents over $1 trillion dollars) whose member have a common interest in improving data quality by collaborating and talking about this topic independent of the structured processes that they execute every day.
Benefits of a cloud-based supply-chain platform
A cloud-based supply-chain platform can solve many of the above challenges including:

  • Automating supply-chain processes such as order fulfillment, purchase to pay and inventory tracking so they span across the entire value chain of partners. This in turn reduces:
    • Inventory carrying costs
    • Operational expenses
    • Transportation expenses and overhead
  • Once a supplier joins the community, it can easily work with multiple retailers. The work and contribution of any single partner or a logistics service provider — such as repair or improvement of their data connection for just once for one of their customers on the network — can be applied to  everybody in the community.
  • The cloud platform becomes a virtual supply chain platform — customers can plug and play with partners virtually without any rewiring of the nervous systems of these processes.
  • The physical flow of shipments, orders, inventory, packing lists, commercial invoices, etc. can be informationally represented and persisted just once. Entire communities gain access to a central repository of information — a single version of the supply chain truth — hence, they can collaborate and make decisions that improve operational efficiency, reduce waste and improve cycle times.
  • The cloud becomes a system of record for transactional processes around an order while it’s being executed and can provide a level of detail of the order status as it’s being executed to the merchants, inventory planners and procurement staff.

According to Johnsen, GT Nexus has a single multi-enterprise, multi-industry, cloud collaboration platform where all the partners responsible for executing orders in a supply chain can rally around and synchronize with a single version of the transactional truth. Their platform has the integration, rationalization and translation technologies to map all the non-standard data sources and organizational structures into a universal standardized format available to internal and external consumers.
“The event completeness metric for many companies was 54 percent for almost 30 years,” said Johnsen, “but it went to 76 percent in five months for customers on our platform. Many of our customers are at 90 percent today.”
That’s an incredible improvement, one that would not be possible without a cloud-based platform.
I thank Greg Johnsen, executive vice president of marketing, and co-founder of GT Nexus for sharing his thoughts on Big Data and cloud-based supply-chain platforms.
Kamesh Pemmaraju heads cloud computing research for Sand Hill Group. Follow him on Twitter @kpemmaraju.