There is a new generation of control towers, far more robust than what we have seen in the past. Covid-19 has served to highlight just how important supply chain agility is.

The key building blocks for the modern supply chain control tower includes data from key supply chain partners, robust supply planning, and a master data management/data harmonization layer that helps to normalize the data and can then feed accurate data to the planning engine or team members. With these building blocks in place, companies can navigate supply chain disruptions with an agility that they have never had in the past.

While working on ARC’s global supply chain planning market study, now complete, I got briefings from Infor on both their supply chain planning and supply chain collaboration network solutions. Infor, it turns out, is one of the few suppliers that can provide all three of these pillar solutions; in short, they can provide a modern control tower solution where the customer has but “one throat to choke” if problems arise.

Most of the writing done on the new generation of supply chain control towers has been focused on the need for concurrent planning. But the ability to get data from across the end-to-end supply chain, harmonize it, and use it to make better decisions is equally important. Infor’s end-to-end trading partner data platform is the focus of this article.

The Infor Nexus Supply Chain Collaboration Network Solution

A supply chain collaboration network (SCCN) is a collaborative solution for supply chain processes built on a public cloud – many-to-many architecture – which supports a community of trading partners. Infor Nexus is one of the better-known solutions in this market.  65,000 companies use the platform. $1 trillion in trade is managed over the platform and $50 billion in payments is transacted here.

A SCCN allows a company to get data useful for planning, sourcing, manufacturing, logistics, returns, supply chain risks, and supply chain finance. Basically, a SCCN that can get data feeds from across the plan, source, make, deliver, risk, and finance categories allows companies to have an end-to-end view of their supply chain. Being able to get these alerts with fewer data exchanges lowers IT complexity and speeds implementations. In Infor’s case, their SCCN model also contains a digital twin – an end-to-end model – of the supply chain.

Infor’s Matt Simonsen – a Director of Product Management for the Infor Nexus solution – explained that they use a knowledge graph to map parties, places, products into the digital twin. Every transaction writes to this in-memory digital twin.  What does this mean in practice? An in-memory solution allows for Big Data to be accessed very quickly.

What does a “knowledge graph” do? A knowledge graph does not start with a model of the supply chain, it creates that model based on what the data shows to be true. Mr. Simonsen said, “we don’t’ start by having a model that says the ports are here, the suppliers are located here, and so forth. It is a supply chain model built on the fly based on what is really occurring in the network.” A knowledge graph model is more accurate. That is because these types of models are based on what the data shows is really happening versus what a supply chain manager thinks is happening.

But the Nexus model can also provide the kind of data that digital twins that are based on a static planning model struggle to provide – true lead times (and other supply chain throughput/cycle time metrics). These parameters are critical for the creation of optimum plans. Lead times in planning systems are often measured, input into the system, and assumed to be true from that point forward. The Nexus system can measure the changes that are occurring and keep the lead times up to date.

The changes measured in the Nexus solution can be based just on the data of one company. For example, when my company orders from this supplier in Asia to be delivered by this Ocean carrier, the lead time is 24 days. This kind of data could be provided by a point to point data exchange solution.

But the network can do this and more. If a company wants to work with a new supplier, and that supplier is in the network, onboarding is easier.

Also, certain types of alerts are better if they are based on network data rather than a company’s own data. For example, how long does this Ocean carrier take to move goods from the port of Tianjin to Long Beach? If the answer is based on the data from all parties in the network, then the alert is more accurate. And the more data that is used, the better predictive alerting will be. Predictive alerting needs to be based on Big Data. “Earlier sensing leads to more agility,” Mr. Simonsen pointed out. “Many find out too late that a shipment is delayed. Their only choices are to expedite the shipment or admit to the customer it will be late.”

Visibility without Decision Support has Limited Value

Knowing that something bad is going to happen, if you can’t do anything about it, only serves to ruin your day. Supply chain personnel need to know what they should be working on. The Nexus solution allows various supply chain personnel – both internal to a company and among key trading partners – to get alerts that are pertinent to their responsibilities.

The work to resolve a problem is often executed in applications. The more applications that must be used, the more difficult it is to respond. In many cases, problems can be resolved in the Nexus solution.

Infor is building out the options on what a person with a designated job could do to resolve a problem or opportunity.  In some cases, the resolution can be executed on the spot.  In some cases, collaboration with partners or other internal departments is necessary to make and execute the best decision.  Infor enables collaboration and execution with internal and external partners in its SCCN.

When a solution to an exception situation is based on prebuilt rules in an execution solution, we speak of an “orchestrated” solution.  When a planning engine provides an answer, we have an “optimized” solution.  Because both SCCNs and planning solutions model the end-to-end constraints and costs, determining which solution to use to solve an exception situation depends upon timing.   SCCNs are better suited to address exception situations in the immediate and near-term execution window.  Planning solutions are better suited to address exception situations in the medium to longer term timeframe. Learnings from SCCN orchestrations are shared with planning solutions to optimize future planning with real-world feedback.

Infor is working to more tightly integrate their SCCN and planning solutions. “Our goal, Mr. Simonsen concluded, “is to update the supply chain planning model based on what is happening in the network model.”

On May 5^th and 6th, Blue Yonder, the former JDA, had their annual user conference. The pandemic led to its being held on-line. I’ve previously written about the benefits BD receives from their real-time control tower in terms of being able to serve their customers better – including hospitals desperate for medical supplies during this pandemic – and new supply chain efficiencies.

BD, with annual revenues of more than $17 billion, is a leading global medical technology company. They offer with a broad set of products. This includes solutions used in COVID-19 research, a product that tests for COVID-19, and many other products used for treating hospital patients. 90 percent of those in the US who go into a hospital will encounter a BD product.

Two speakers from BD spoke at the Blue Yonder ICON conference. This article is based mainly on the presentation of David Baitman on May 6^th. Mr. Baitman is the BD Control Tower Lead. The control tower was built using the Blue Yonder platform.

The Scope of the Control Tower Project

In any significant IT project, the scope needs to be clearly defined. The requirements were gathered from various business units and supply chain functional areas including logistics and planning. Effectively managing inventory is at the heart of supply chain management. BD bucketed inventory management into three categories: inventory at rest, inventory in motion, and hot items.

Inventory at rest was based on products located at BD manufacturing plants and BD’s distribution centers.  This type of visibility includes the current stock picture across the network, what customers want to purchase, what stock is expiring, what supply orders are outstanding.

Inventory in motion includes what is moving between, in, or out of each location, what is on each truck, ocean container, airplane, or rail car as well as whether the shipment is early, on time, on time but not received, or late.

When it comes to hot items, Mr. Baitman pointed out that the trick is to identify what is “news” and what is “noise.”   Perhaps a truck is late, but it contains product where BD has plenty of stock. On the other hand, an ocean container might come in early but contain products that are in a low stock position. “So, we want the system to tell us which to receive first.”

The company is building their control tower in three phases. The first phase was to support visibility from a BD plant to a BD distribution center, the second phase is from the BD warehouse out to the customer, and the third phase will be visibility of supplier shipments into BD facilities and internal manufacturing processes. They have finished phase one and are well into phase two.

An Agile Implementation Process

Using an agile implementation methodology, they do work in short sprints. Sprints are small manageable blocks of work that deliver capabilities sought by their users. Mr. Baitman believes that the increased frequency of collaboration and feedback at the heart of this methodology, results in two advantages. Those advantages include more rapid implementation of functionality valued by users and faster detection of issues and defects.

Further an agile implementation process increases flexibility. Because they employed this methodology, they were able to pull functionality needed during this pandemic up in the project timeline.” This feature set allows BD to track critical pandemic medical products shipped via FedEx to customer locations.

The project plan involved an understanding of what problems the company was trying to solve and tackling high value problems first. The requirements definition and project plan were based on an understanding of which requirements helped solve high value problems and then tracking the value associated with solving those problems. Value capture was built into the design of the project because BD needs to be able to tie actual savings to what was promised.

Data, Data, Data: The Heart of a Control Tower

BD built a data lake that consumes all the critical data from the ERP databases. This cloud-based database is hosted by Microsoft Azure. BD chose to replicate the data using database logs rather than pulling the data from the applications. They used a product called Qlik Replicate (formerly Attunity Replicate) for this.  The data lake approach also provides a degree of future proofing.

There was significant work with the business users to generate a harmonized data layer. The critical master data for supply chain management includes sales orders, shipments, inventory, and lanes. Different business systems can define all the fields associated with these forms of master data differently. But BD took these objects and asked their business users what data points they needed for each of these objects. The BD team did not care what various business systems, like SAP, thought the object looked like. Harmonizing data is a difficult task. Companies need to give themselves plenty of time to work through these issues.

Two sets of core functionality include being able to understand what BD’s current inventory position is and tracking shipments. To understand the current inventory position, BD took the initial stock position and then calculated how that inventory position is changed by things like quality inspections, purchase orders, production orders, and intercompany demand. This generates visibility to in-stock inventory at a specific location.

In terms of tracking shipments, carriers needed to be on boarded. Mr. Baitman said “all carriers are not created equal. Some are more willing to provide data.  Some are more able.”  In order to track shipments, carriers reference numbers needed to be tracked. If it is not possible to track a carrier’s shipment with a tracking number on the carrier’s website, it almost certainly can’t be tracked in the control tower. Further, BD itself may need to change processes to make sure those tracking numbers are captured in a business system.

It is not viewed as the control tower team’s job to clean up all data issues. The core systems may need to be fixed or a regional team or business unit may have to change processes to improve the quality of the data.

Control Tower’s Need to Provide Actionable Information

BD also built key performance indicators (KPIs) to be used in the platform. The KPI work included deciding which key performance metrics were needed, and then aggregating data from all the different forms of master data – deliveries, inventory, orders, etc. – and then applying the right math to that data. By building KPIs in this way, users have a view not just into performance, but this architecture allows users to have the ability to drill into the data to better understand the root cause of an issue.

To make the information actionable, BD built playbooks. Playbooks are based on common scenarios – for example late shipments or an unusually large order from a customer – and then figuring out the steps involved in rectifying that problem. BD is also looking to build playbooks for uncommon scenarios – what the company should do in the event of a hurricane, or a port strike, or a pandemic.  Building these playbooks is an ongoing activity. Even though it is not known what the completed playbook will look like when the system is being designed, for the system to work well, playbooks do need to be part of the initial design.

Not all problems can be solved by one or two planners. The Blue Yonder platform also allows for the building of situation rooms. In the case of a crisis, like caused by a pandemic, you might even call these ‘war rooms.’ These situation rooms allow for stake holders across an organization to collaborate to solve problems.

Conclusion

The kind of Control Tower that BD has – that integrates planning with logistics data, has real-time visibility to important supply chain events across the end to end supply chain, and provides playbooks and situation rooms for problem resolution – seems almost magical. This kind of holistic solution was just not available a few years ago. This is the technology stack needed to much more tightly tie planning to execution.