Network Design and Optimisation: Maze mastery

LinkedIn +

A supply chain network can turn into a maze without a proper Understanding of its true objectives. Malory Davies looks at the key factors in ensuring the success of a design and optimisation project.

The process of evaluating, designing and optimising a supply chain network is one of the most complex that any organisation has to deal with. Even ensuring you have the right starting point can be a challenge.

This article first appeared in Logistics & Supply Chain, July 2016.

This article first appeared in Logistics & Supply Chain, July 2016.

Dominic Regan, senior director for EMEA logistics applications, at Oracle, says: “Organisations looking to design and optimise a supply chain network should ensure that the approach taken aligns with the over-arching corporate strategy. Before undertaking this process, businesses must consider things, such as plans for international expansion, growth expectations, external developments from competitors or regulators, sustainability initiatives and where automation is appropriate.

“There also needs to be a clear understanding of what the true objectives of the network are; be it increasing profit by reducing overheads, enhancing customer service, or both. This will give businesses the peace of mind that the network best meets the demands of its customers. Networks must also be tested before implementation to ensure that they are resilient and able to cope with spikes in demand or unpredictable macroeconomic factors,” says Regan.

Lucy Larkin, managing director in Accenture’s retail practice, points out that a big part of the challenge is that this isn’t something the supply chain team does every day and there are lots of potential pitfalls, so evaluation of your supply chain network can be tricky.

“Aggregating and analysing the relevant data, modelling the network, and evaluating future scenarios are all infrequent but detailed activities. Firstly the level of resource required may not be available from within the day-to-day team, so extra support may be needed. Next, some of the information needed may not to be routinely reported, so you’ll need to extract data and validate this before it can be worked on. In our experience getting the base data right is critical as this forms the foundation for your decisions about the future.

“Modelling your network will be complex, as this will reflect all of the activity across your entire supply chain. You then need to think about the potential future options you might want to evaluate. It can be difficult to decide on the right scenarios to model as there’s always a limit to how much time and resources you can apply. Understanding which options are likely to yield best results takes experience and is key to getting to the right answer. This answer will be about operational cost, but most importantly about service, particularly with the growth of digital, for example the demand for faster delivery and tighter time-slots.

“The final challenge is delivery, not only making changes to improve your supply chain but actually realising the full benefits from this – especially when the future will always be different from the one you predicted,” says Larkin.

The importance of being specific on the business question you are trying to address first is highlighted by Will Lovatt, VP sales, EMEA at LLamasoft.

“If engagement from the wider business has been obtained, supply chain design projects can become the opportunity for everyone to share their pet theory on what they would like to understand. While the engagement is welcomed, priorities need to be documented and agreed, and the purpose of the initial exercise ring-fenced. Momentum is built from a crisp first project with valuable insight more easily than a project that takes many months and addresses many questions averagely.



One of the key steps is to validate and calibrate the initial as-is network model against reality to ensure that once the modelling process commences, variants and scenarios will be properly represented. Many modelling processes miss a critical stage, which is to clearly define the business questions to be asked. It’s too easy to slavishly follow the data and create a monster process which delivers back little of practical value. Never start modelling without a clear idea of the business question that is being addressed,” says Lovatt.

Regan points out that the team responsible for the modelling process should have clear objectives and responsibilities to make sure that the project moves at a good speed and all parties are aware of their goals.

“At certain times in the process, there will be a need to obtain approval from the business, which can really slow things down if the right people are not on board from the beginning or if there is a lack of executive sponsorship. When starting the network modelling process, it is important to understand what data is needed to complete the design, and establish if the data is available. If it is available, then they must understand how that data can be extracted effectively and quickly, and then analysed efficiently and objectively. The KPI’s for comparing different scenarios should be prepared and agreed well in advance.

“Upon building the network, organisations should have a forward looking plan in place, so that any maintenance on the network, rebuilds or modernisation requirements can be completed with minimal impact. By having this process in place, businesses can be certain that all parts of the network are aligned.

And says Regan: “Additionally, I would advise having a contingency plan in place, which outlines how to handle any situation which could potentially go wrong. By having this prepared from the offset, organisations can mitigate the impact of network builds running over schedule, or safeguard against the risk of the network not performing as per the original plan.

Larkin highlights some of the important planning questions which need to be addressed before launching into modelling.“First, you need to set the scene with clear scope and project objectives, as well as the overall project timescale expectations. Knowing what needs to be delivered and by when will undoubtedly shape your thinking about how you approach the modelling and what level of detail you work at.

“Planning the approach and team is vital, so you know ‘how’ and ‘who’ before you start. It’s good to look at options and decide the level of detail needed in advance. Do you have the necessary systems to execute this? Do you have the required capability and availability or is other support needed? It’s a really good idea to develop a detailed project plan, as this helps with sequencing, resourcing and addresses dependencies. But a perfect plan needs contingency too to cope with the unexpected.



You have to bear in mind that it’s a strategy project, which is about setting direction rather than being a future budgeting exercise,” she says.

There are some increasingly sophisticated technologies available to support the supply chain network design and optimisation process.

“While the traditional spreadsheet approach remains viable for simple and relatively static modelling initiatives, leading-edge companies are making use of innovative and flexible technologies which are now proven in a wide variety of supply chain business scenarios,” says Lovatt.

“The transition from occasional strategic review to more dynamic, near-operational-level supply chain design has driven many organisations to invest in an in-house capability. This equips them with the technology and skills to deliver massive tangible benefits, using the supply chain to gain competitive advantage.

Larkin says: “An alternative to generic well-known software is the bespoke modelling approach. Bespoke modelling allows you to specify where you want the detail and where you don’t, how much time and effort you want to put it, as well as accommodating limitations you might have (missing data for example). A tailored approach to design is part of the benefit, there is also the ability to drill down into the results to see exactly where and why these are generated.”


Advanced analytics hold the key to new insights

Supply chain networks bring large amounts of useful data into an organisation and big data analytics are now making it possible to gather more insightful data from a wide range of structured business and unstructured data sources like social media. This data can optimise supply chains by being able to identify market trends and build forecasts around them.

Oracle’s Dominic Regan says: “With supply chain cloud technologies, organisations can obtain all of the information of orders in real time. This data will provide insights into the cost associated with each product, allowing the business to understand how truly profitable individual products actually are.

“Along with business data, social media provides a lot of valuable unstructured data. If brands can collect and analyse tweets, likes, public posts and other indicators they can build an accurate picture of where public taste is heading in the lead up to seasonal periods. Once the season is in full swing, consumer goods companies still need to monitor their supply chains to ensure actual results are matching predictions and flex distribution accordingly and keep customers happy.

“This on-the-fly supply chain optimisation demands that brands are able to match supply with demand accurately and change supply flows in real-time. In this regard few trends are as important as omni-channel fulfilment. This allows companies to expand fulfilment options to include stores, inbound supply, manufacturing capacity and supplier availability. This leads to decreased reliance on their customer distribution centres to meet demand – an important way of ensuring profitability.”

Share this story: