Checks and balances

Demand Forecasting & Planning - S&OP Technology Report

Manufacturing & Logistics IT spoke with leading industry spokespeople from the analyst and vendor community about how modern Demand Forecasting & Planning/Sales & Operations Planning (S&OP) solutions and other planning-related technology can help to provide a more efficient supply chain, manufacturing and distribution regime.

One thing’s for sure, 2021 has been the year of supply chain lessons learnt. As Shaun Phillips, product & market director, QAD DynaSys, points out, we are now talking about and implementing technologies and processes that were off the radar pre-pandemic. In this regard, Phillips cites some key examples: 

• For more effective supply chain agility, agile planning acts as a navigator through these disruptive times. Agile supply chains detect actual changes and anticipate potential changes in current demand and supply signals. 
• Adaptive planning uses supply chain digital twins to automatically and continuously analyse, predict and tune supply chain model parameters using actual observed data. Adaptive planning re-calculates lead-times, run-rates, capacities and production yields.
• Supply chains are focusing more on resiliency. This is presenting in various ways. More local or nearshoring sourcing, more days of inventory and other redundancies, more agile production and distribution practices, stronger reliance on short-term demand sensing techniques, and more comprehensive risk assessment of tactical demand and supply plans.
• The rising prominence of Corporate Social Responsibility initiatives is bringing an increased focus on supply chain sustainability, and a big part of this will focus on the circular economy.

In terms of some of the key drivers behind these changes, Phillips cites:

• Agility – Driven by continuous disruption.
• Adaptive Planning – Driven by a need to simplify planning, while improving outcomes and leveraging more available real-time data.
• Resilience – Driven by disruption on a large scale where even the most reliable sources of supply and demand were challenged.
• Sustainability – There are so many independent forces driving the cause for supply chain circularity; such as scarcity of raw materials, consumer buying preferences, regulatory controls, rising commodity prices, as well as CSR & ESG mandatory disclosures.

Accurately managing and predicting demand

The efficiency of supply chains and more accurately managing and predicting demand is absolutely critical in today’s fast-changing and sometimes highly unpredictable world. During the pandemic, managing stock and forecasting of essentials – particularly around groceries and medical supplies – has become incredibly important. “Who would have thought toilet rolls would have become a worldwide talking point?” says Manhattan Associates’ UK managing director, Craig Summers. “So, from a supply chain and demand perspective, pressure on being more efficient is greater than ever, and omnichannel becomes even more important because if people are isolating and can't go out to brick-and-mortar stores to the extent they did before the pandemic struck their only way of ordering is through other channels – typically online. This naturally can result in significant increases in demand, which again puts added pressure on the supply chain.”

So, how does Manhattan Associates’ Demand Forecasting software offer a means to meet these types of challenges and deliver tangible benefits for customers? “Even the most regular-selling item can be difficult to forecast.” Says Summers. “Add in slow sellers, intermittent items, new item introductions and promotions, and forecasting future demand can be extremely arduous. Multiply that problem by a massive network of SKU locations, and the problem becomes even more daunting. However, predicting future demand is the first step in any stock management strategy, regardless of the industry or type of items being managed.”

He adds that Manhattan’s solution provides visibility into network demand and combines proven forecasting techniques with demand cleansing, seasonal pattern analysis and self-tuning capabilities to accurately anticipate demand even in the most complex scenarios. Using machine learning to constantly evolve and adapt the science of demand forecasting, our customers benefit from higher degrees of forecast accuracy, without heavy user intervention.

“Manhattan simplifies the complex science behind demand forecasting by focusing the analyst on managing just those key exceptions that the system itself cannot reconcile,” says Summers. “It becomes easy to manage an infinite combination of locations and products with differing time horizons and aggregation to enable range, financial and merchandise planning, in addition to replenishment.” Summers explains that the software offers a raft of demand and forecasting benefits; including improved forecast accuracy, the ability to model demand for slow and intermittently selling items, the ability to forecast demand for both enterprise planning and daily replenishment activities, and the capability to scale to meet the needs of very large networks.

Summers makes the point that a lot of companies are adopting a cloud-first policy because they realise that owning their own hardware and all of the resources that go with it is not core to their business. “Also, if you then think about in-store type capability the less in-store type hardware you need the more it can come in through a cloud device,” he adds. “This can also result in less need for support. Also, what many customers like to be able to do is size their hardware for, say, the one day a year when they are at their peak. And that day’s peak might only be for couple of hours, but everything needs to be sized to be able to cope with that most critical time. When you have something built on cloud architecture in such a way that it can scale up very quickly to these types of requirements as and when required the demand and supply issue is no longer such a worry for companies. Indeed, through the pandemic we have seen some customers running efficiently at Black Friday levels of demand more regularly.”

Digital supply chain twin

Tim Payne, research vice president, Gartner, considers that within the world of planning the rise of digitisation and more automation resulting in less manual work by planners is definitely a current trend. “This was the case before the pandemic, but in the wake of COVID-19 many companies have been pushed to double down on that effort,” he says. “This relates to the digital supply chain twin, a term Gartner coined three or four years ago.

Planning is about decision-making; deciding, for example, what you're going to make and put in inventory and if you make high-quality decisions you create more value for the business and you help to prevent value loss. These high-quality decisions can be measured not just in terms of the accuracy of the decision or the plan but in term of the timeliness, the relevancy, the granularity, the time horizon and so on. However, if the planning model isn't a very good representation of the real world you end up with a plan that works in the model but doesn't work in reality. So, the digital supply chain twin has gained a lot of ground among end-users.”

Payne adds that users often say I think I need a digital supply chain twin but how do I get one, where do I go? “So, the market has picked up a lot on that term and quite a lot of vendors are increasingly moving in the direction of being able to provide a digital supply chain twin – the idea being that you can utilise low latency, more granular data that is becoming available from many different data sources; internal, external, signal data and so on. Then, it is about being able to make some sense of this data and being able to use it to better represent the uncertainty and variability of the real world within the models you use for planning.” 

In terms of drivers for these types of changes, Payne comments: “You can point to particular events such as Brexit, COVID-19, trade wars, major weather events or volcanoes and so on, but what underlines them all is that they are disruptions and there is uncertainty involved. So, I think there's more realisation within planning where companies are saying we aren't very good at being able to manage planning in disruptive and uncertain operating environments. Some forms of disruption will be different than others, but the point is it's not likely to get back to a steady state situation.

So, I think that's what's driving the interest in things such as the digital supply chain twin in order to better model and react to disruptions in the real world. If you’ve got AI and machine learning you can better predict and react to probabilities of events happening or not happening. If you’re agile and responsive you can spin about quicker in terms of taking advantage of an opportunity or mitigating disruption. And if you have resiliency, you can absorb more of that uncertainty and that disruption more effectively. So, when you peel it all back, I think it still comes down to companies saying I'm just seeing more are more examples of uncertainty and variability and disruption and I've got to get better at being able to manage my supply chains in those types of environments.”

With this in mind, Payne explains that users want to use more data – lower latency, more granular data, internal and external data, structured and structured data etc – and that can help them to get better supply chain visibility. “However, with so much data you have to be able to make sense of it and relate it to what’s important to your business,” he says. “We now often hear about control towers, but there is some confusion about what a control tower is. It tended to grow out of the business agility world and of course there are benefits in knowing there is, for example, a shipment delay that you wouldn't have known about. The data can start to paint a pattern that maybe points to a high probability that your shipment on this ocean freight is going to be delayed so you can immediately start asking the question ‘what can I do about it?’

This question gets into the territory of what Gartner refers to as respond planning – how can you respond intelligently to what you see happening, what's being predicted to happen in execution? And with COVID-19 factories and DCs have closed for certain periods or have only been half open, demand is shooting up for certain products and plummeting for other products. So, there are all kinds of demand and supply events and it's all about been able to very quickly figure out what that means for your business and your customers and what should you do as soon as you've got all the relevant data to hand. So, you’ve got to make effective decisions, which means you’ve got to be planning; albeit it can often be very granular and short term but it’s still planning. Therefore, the rise of respond planning is something we've seen a lot over the past year or so.” 

Extremes of variability 

Bryan Ball, vice president and group director, supply chain, ERP and GSM research practices, Aberdeen Group, cites the good industry as a good example of disruption and its effects. “If a large percentage of food supplies is being supplied to restaurants and suddenly restaurants are shut down because of the pandemic, that food is now sitting idle and it's in the wrong form and in the wrong place,” he says. “If it’s going to be saved it has to be repackaged, but there will also be shelf-life concerns and possibly refrigeration issues. Grocery stores, on the other hand, are deemed to be essential and are going to remain open.

So, you have an abundance of demand in one area and a shortage in the other although, historically, you have been used to a certain balance between restaurant supply and supply to the grocery store. The demand signal in terms of how much is needed and where it’s needed has been totally disrupted. So, you need to get produce repackaged and reprocessed for the use in grocery stores, but do you know what mix to put the chicken or ground beef in? Ideally, you need a demand forecasting system that is able to deal with these extremes of variability.”

When restaurants start to open back up, suppliers have the same problem in reverse,” as Ball explains. “Now, they need to move much of their goods away from the grocery stores and direct to the consumer deliveries and prepare to supply restaurants again. So, they need to allow for more demand going back through their previously normal channels. But if restaurants close again they end up with this back-and-forth routine again, with all the fits and starts.

Then to complicate things even further there might be a vaccine passport requirement, which might exclude a number of previous regular customers. I'm not arguing the point politically but simply saying that because they open back up but add these extra conditions much of the demand for food doesn’t bounce back to where it had been previously. So, despite all the back and forth the demand signal is going to continue to maybe isolate or deviate and coalesce to a new level of stability. And there's going to be some oscillation until the whole issue is resolved one way or the other. If we can eliminate much of the noise and get a more accurate signal of how people are going to shop and where they going to shop this will provide a better level of stability.”

Ball adds that during all the disruption, come companies decided to manually override the signals of their systems, thinking they could make better decisions themselves because these were unprecedented scenarios. And, explains Ball, in many cases it took a few weeks for people to realise that it would have been best to trust system with the right type of algorithms rather than rely on communication via a virtual conference room with spreadsheets trying to sort things out for the best.

Disruptions will continue to varying degrees, maintains Ball, whether it’s the lockdown of a whole country that companies might supply to, or whether it’s more local or whether it’s certain markets within countries. “So, you need to be prepared, and if you don't have the tools to accommodate this – to give you the demand visibility you need – you're going to be flying blind,” he says. 

As Ball points out, many companies over the past year or so had to embrace omnichannel; to support retail, for example. He adds that cloud is a good methodology for order management, so more companies sought this type of system. “Many also wanted better inventory visibility tools or visibility suites, and functionality that could provide better upstream supply chain visibility of suppliers and transportation networks,” he explains. “All these individual things can certainly add real benefit. Even though the demand signal itself may not have improved a whole lot at least these companies can see more, make better decisions and start adjusting as opposed to scrambling after everything in manual mode and spreadsheet mode.” 

The advice Ball gives to companies is to put in place the most advanced algorithms they can. He also points out there are eight main things they need to make a schedule: A part number or SKU number, a quantity, a start date and due date, a router, a bill of materials, a unit of measurement and a lead time so they can predict what their supply is going to be. “So, they need those to come up with an effective plan for any kind,” he says. “Then, they can look at each one and say, well, the routing, the bill of materials, the unit of measurement and the start and due date are probably going to be kind of fixed, but then they get back to the demand signal itself – the quantity – and they can start asking how good is that number. That's when you get literally into what the demand signal is and what the quantity should be and where it needs to go.”

Big Data

Michael Clarke, research analyst, VDC Research, considers that the integration of artificial intelligence (AI) to strengthen predictive analytics for demand forecasting and inventory management is a key improvement for these software systems. “Enterprises have historically underutilised Big Data to navigate complex retail challenges due to a lack of comprehensive software capabilities,” he says. “In this sense, whether though proprietary software platforms or WMS tools, the implementation of AI enables enterprises to better utilise Big Data to address key business challenges such as inventory optimisation, picking and packing maximisation, route efficiency in the warehouse, dynamic pricing and go-to-market strategy.”

Clarke adds that predictive modelling with AI today has become stronger and more interconnected across business ecosystems. “Taking into account historical transaction data through POS systems, partner data, consumer demographic data, regional landscape data, weather trends, macroeconomic metrics, competitive pricing, and website search trend data, among several other metrics, AI systems can create stronger forecast scenarios for strategic decision making,” he says.

“Not only have AI systems created stronger predictive accuracy for inventory, but also display with more intuitive dashboard analytics designed for C-level executives to leverage for complex business decisions such as, rolling out new product launches, optimising new brick and mortar locations, and regionally designing supply chain networks. In today’s retail landscape where margins are increasingly thin and omnichannel strategies are rising freight costs, predicative modelling with AI has the power to improve margin profitability through dynamic pricing, route optimisation, stock replenishment accuracy, labour optimisation, and warehouse and supply chain design.”

Major cost reductions 

In terms of drivers for change, Clarke reflects that retailers have historically struggled with inventory management and warehouse optimisation where the largest enterprises are losing $500 million in wasted inventory annually. “Sharper AI software can help reduce this to $50 million with better predictive analytics and inventory optimisation capabilities that capitalises on transaction data, consumer demographics and macroeconomic trending,” he says. “A stronger need for cost savings, labour optimisation and margin improvement within the framework of growing ecommerce and omnichannel shipping demand are key catalysts for investment into AI systems to improve margin profitability.”

Clarke believes the pandemic has only exacerbated the need for accurate stock levels and inventory accuracy across warehouse and retail locations, as omnichannel and ecommerce operations have accelerated warehouse turnover and parcel volumes amidst the need for contactless fulfilment options. “As legacy software tools have proven less accurate in demand forecasting and predictive modelling, AI systems have filled the gap with underutilised enterprise data,” he says. “As 50K+ SKUs are handled at the largest warehouses, the smallest errors can have the largest consequences. Thus, there is a strong need across enterprise ecosystems for inventory optimisation, storage efficiency, route and transportation management, supply chain design, error tracking and performance monitoring – all of which AI systems can help address through synthesising large data.”

Karen Sage, chief marketing officer, Syncron believes AI and machine learning (ML) should be crucial tenets to a company’s Demand Forecasting/Planning/S&OP software. “Together, AI & ML are intelligent are workhorses capable of going beyond what any excel spreadsheet, homegrown solution or even direct human involvement could perform on its own,” she says. “Renowned AI researcher and former Ivy League professor, Dr. Darrell West gracefully describes AI as a wide-ranging tool that enables people to rethink how we integrate information, analyse data and use the resulting insights to improve decision making.

Indeed, AI is intentional, intelligent and adaptative – foundational attributes of any effective decision-making entity, and particularly for tasks like inventory management which involves a lot of complex data, both historic and real-time, and a need to leverage that data whilst capturing and optimising multidimensional objectives such as balancing costs, service-levels and risk. The benefits of AI layered with the strength of ML, which analyses data for underlying trends and anomalies, give manufacturers a reliable foundation to make the best decisions for their business and their customers depending on their situation. The bottom line is, AI and ML brings intelligent, automated decision making to reality and scale.”

Additional promise of reward

In terms of drivers for change, Sage considers that AI and ML have begun to play major roles in everything we do to simplify and automate our lives when it comes to being able to operate and successfully manage a supply chain. “Because these technologies have only begun to be deployed in a large scale recently but have already shown immediate and substantive advantages even in their most rudimentary of forms, AI and ML hold out additional promise of reward because it is easy to anticipate there is room for the AI/ML models to be refined, advanced and matured,” she says. “Without a doubt, AI and ML will remain essential tools and will grow in use and sophistication. Indeed, there is a strong appetite amongst technologists for automation to become smarter, more efficient and better at solving inventory optimisation problems at scale.”

However, adds Sage, one thing the pandemic has taught us is that AI models work best when faced with normal operating events and common directives such as to drive cost efficiencies under constraints of that goals defined by metrics used to ensure customer expectations are met consistently and predictably. “Under uncertainty and different, varying objectives, simple AI/ML models can fail to meet expectations and can become unpredictable in their recommended consequences,” she says. “For example, in the event there is a tremendous storm forecasted, there is likely a run on basic staple good in the local grocery store. The AI/ML algorithms not having a line of sight into things like weather or a broader view into surrounding areas and inability to transport goods over the road, might recommend reactive responses that might not only result in even worse inventory stockouts but extra fulfilment costs.

During the storm mentioned above, a run-on basic staple goods might deplete inventory to a point where a new inventory ordered is triggered. Even worse, because of the state of the inventory and estimated delays in normal transit times, the models might recommend expediated shipments, the raising of allowable spend for goods and unintentionally creating a worse situation and sub optimisation of spend. Fulfilment forecasting is particularly challenging in predicting demand – even more than supply. That said, anomalies come from both supply and demand data, since there will be unexpected fluctuations in each.”

Sage also makes the point that the recent fluctuations of the supply chain have put a spotlight on how we are all impacted by our shared global economy. “Container shortages, extended lead times and cash-strapped suppliers have pushed manufacturers to re-examine their legacy systems and make critical determinations on if they are the right tools to meet the demands of the future,” she says. “Today, it is integral to business success to move beyond relying only on historical data from past sales or stock outs, but to instead introduce complexity to demand forecasting using technology developed specifically for this purpose.

To succeed in the aftermarket especially, you must master the delicate balance of managing sporadic, non-traditional supply chain demand, while meeting the expectations of your customers.” Sage adds that implementing a software solution built specifically for optimising and streamlining your inventory ecosystem offers residual benefits to the manufacturer as a byproduct is more reliable and comprehensive data, which can then be analysed to further enhance business operations. “Syncron has the unique distinction of focusing only on the aftermarket space, allowing our customers enjoy the extra benefit of leveraging insights from peer companies across varying industries,” she says. 

Integration of Market Intelligence into Demand Planning

Based on recent engagements with prospective clients, Jonathan Ogg, senior solution architect, Sofco, comments that several key current themes come straight to mind, and the talking point depends on an organisation’s maturity in the use of Demand Forecasting/Planning/S&OP software. “For mature users of Demand Planning software, the integration of Market Intelligence into Demand Planning both in terms of data and process is seen as the next step forward,” he says. “Examples of this can be in relation to Market Growth/Decline, Market Share, Market Innovations, Promotional impact, and Weather Forecasting.

There are organisations that can provide this data at a cost and clearly there must be a cost/benefit balance. The data also must be relevant so that it can be used in conjunction with the Demand Planning system and influence the Forecast outcome in a positive way. This may be as a result of human analysis and collaboration, but also algorithmic analysis and application. In considering this approach the technical challenges cannot be ignored in terms of what format the data is in, what transformation is required and what cross referencing is needed in order to ensure that Demand Planning and Market Intelligence are ‘Like for like’ in their identification of products, categories, markets etc.”

Ogg adds that there are also many companies who have existing Planning software whose organisation is going through substantial change. “So, the questions are, does the software and software vendor have the capability to remodel, how long will it take, what will it cost, how much time will the business teams need to take out of daily operations to support this, and what is required from IT if there are changes to infrastructure policy?” he says. “These and other questions need to be answered by the business and software vendors if successful remodelling is to be delivered.”

Ogg reflects that another regular talking point for companies who have yet to deploy any Demand Forecasting/Planning/S&OP software is do we need to do this, who is the best vendor, what will it cost and crucially, will it deliver on the benefits? “These are just the starting questions as any business going down this route embarks on a journey where there are many factors aside from software,” he says. “It can bring major change to processes, roles and responsibilities, IT deployment along with all the challenges of delivering a project on time and in budget. Key to this is not just choosing the right software but the right vendor as well who has the sector specific consultancy experience and technical expertise. Other key partners may also be considered e.g., Change Management specialists. Whilst this may not be the latest blue sky thinking or innovation it is a vital topic for the many established or up and coming organisations who are considering this journey.”

Looking for the next step in improving Forecast Accuracy

In terms of the topics discussed above, Ogg considers some of the main drivers for change. “Businesses who are long-term users and have bedded in Demand Planning solutions that are providing the benefits they were deployed to achieve are looking for the next step in improving Forecast Accuracy,” he says. “The route is enhancement by the use of Market Intelligence. They believe that they are getting the best possible results out of their existing software systems and processes, and they still believe additional benefits are available. Given that forecasts are never 100% accurate there are always opportunities as even the best businesses still have issues with inventory levels, customer service and the cost of customer service. As ever these are key drivers.”

Ogg also makes the point that businesses go through significant change for a variety of reasons. “This be through consumer buying patterns, COVID-19, acquisition, personnel turnover e.g., retailers moving from the High Street to online, huge increase in Promotional activity, supply chains dealing with new constraints to name but a few. Whilst these things are not new and will always be with us a focus on remodelling is required and this cab also bring with it next step improvements if managed correctly. Therefore, the best ways of doing this are keenly discussed as the circumstances are often out of the control of the business or those who have to deploy the change.” 

Finally, adds Ogg, there are still many companies who are using Excel or older ERP-based planning modules to address often complex planning problems. “As a result, they have to face inherent issues which often result in significant data manipulation and manual activities which can also be open to human error. Where this reaches a critical point, they then consider Demand Forecasting/Planning/S&OP software, and the drivers remain fairly constant around customer service, inventory and costs.” 

The Coronavirus outbreak has exposed the vulnerabilities of current global supply chains. Many have unravelled, unable to withstand intense demand-supply shocks. Against this background, Dr. Julia Saini, associate partner & vice president – mobility, Frost & Sullivan, and Sven Thiede, vice president – energy, sustainability and mobility practices, Frost & Sullivan, cite Shell as a notable success story. They explain that the company, one of the world’s leading oil and energy suppliers, has stayed resilient. “By leveraging data and digitalisation to drive new ways of working and collaborating, focusing on a more holistic approach to safety, nurturing future-focused talent, and pushing forward on sustainability and business continuity agendas, it has ensured that its global supply lines have remained robust, reliable, and most importantly, competitive,” they remark.

More digital ways of thinking

In terms of how Shell has continued to successfully produce and deliver the lubricants and greases required by its customers amid the upheaval, and how it has kept its sprawling supply chain network spanning assets, suppliers and partners running strongly and seamlessly, Saini and Thiede recently spoke with the man tasked with ensuring the supply chain for Shell’s lubricants business stayed operational even during the worst of the crisis – Singapore-based Richard Jory, vice president, lubricants supply chain, Royal Dutch Shell. He explained that across the company’s supply chain, it is building a strong digital backbone. Shell is simplifying complex data to improve productivity and better know and serve its customers. However, Jory emphasises that it is not just about hardware. “I’m also encouraging the 3000 people who work in and with the lubricants supply chain organisation to embrace more digital ways of thinking,” he points out.

Jory believes data’ is fundamental to digital. “It’s about moving towards a digital business using data to make better decisions,” he says, adding that Shell has access to large amounts of data that cover every aspect of the business. “Today, with computing power being what it is, we can leverage the scale and data to make better decisions in everything that we do. That is the core of digitalisation for me,” he says. One good example is BlendRight, an award-winning data analytics tool developed in-house to improve decision-making. “We have been able to combine our huge legacy dataset with advanced analytics tools to make better blending decisions – and in doing so reduced costs by millions of dollars,” explains Jory. Another example is Smart Plant+, a manufacturing execution system at Shell’s Singapore plant that helps the company to predict where there is going to be a failure, a bottleneck or maybe even a quality incident on the shop floor, even before it happens. “This is something that motivates our frontline operators because there is nothing more frustrating than having to spend a lot of time correcting an error,” says Jory. “There is a lot of waste involved. Using data-driven insights helps us get it right the first time, saves time, removes frustration, and just makes for more satisfying work.” 

Revised customer expectations 

Have changes in transportation legislation (either locally or globally) influenced the development of Demand Forecasting/Planning/S&OP systems over the past year or so, in your view? Phillips makes the point that this has been reflected in transportation capacity and lead-times. “This has led in some cases to revised customer expectations and increased inventory,” he says. “The problem is not new. In some cases, opportunistic road freight capacities have arisen via platforms such as Uberfreight.” 

What are some of the key differences regarding the different brands and types of Demand Forecasting/Planning/S&OP systems/devices currently available? According to Phillips, there is a big difference between digital and traditional supply chain systems. He maintains that digital challenges everything in a supply chain:

• Digital supply chains are widely connected to both internal and external systems and data sources. 
• Digital supply chains are ultra-responsive due to their ability to receive near-time demand and supply signals and rapidly anticipate risks or opportunities.
• Digital supply chains are more inclusive and collaborative. Collaboration occurs between stakeholders that transcend functional silos and even business entities. 
• Digital supply chains are highly intelligent. In a digital supply chain data is an asset to be leveraged to extract value, not an expense to be minimised.  

Phillips believes there are no longer any notable remaining security or confidentiality concerns at the more ‘mobile’ end of the Demand Forecasting/Planning/S&OP solutions space. 

Are there any Brexit-related issues that have expedited the need for better Demand Forecasting and Planning? Phillips reflects that Brexit is another form of disruption, as has been the COVID-19 crisis and the trade difficulties between the US and China. “We should expect some uncertainty to exist for the coming 18 months or more as the dust continues to settle,” he says. “However, Brexit is just one example of disruption that it is part of the new normal. This is why supply chain agility and resiliency is important to navigate the disruptions.” 

What might be the next innovations/developments to look out for over the next year or two? “I don’t want to give away any secrets,” says Phillips, “but the digitisation of supply chains is a long journey, and we have some way to go.” He adds that we will continue to see opportunities from emerging technologies. “We are yet to see the mainstream adoption of Blockchain or other forms of distributed ledger technology,” he says. “We should expect the adoption of 5G communications, and the advancement of quantum computing will have a game-changing impact on supply chains.” 

In terms of further changes regarding end-user requirements over the coming year or two, Phillips considers this will change significantly, especially in the area of supply chain collaboration. “End users will require a more inclusive, simpler and more intelligent planning experience,” he says. “Embedded capabilities such as conversational analytics, online risk management and Robotic Process Automation (RPA) will ensure end users have fact-based analytics at their fingertips to make decisions or approve system suggested decisions.”

The future 

What might be some of the key areas of further development to look out for over the next couple of years or so? Sage believes we will start to see even more innovation, faster. “Customer expectations have been unmistakably shaped by their interactions outside of the B2B space,” she says. “We have seen an astronomical rise in online shopping, groceries being delivered to our doorstep, all with consistent communication throughout the delivery process. These experiences have set a standard for customer service, and we are challenged to help our manufacturers build the necessary infrastructure in their journey towards digitalisation. It is the responsibility of the OEM to empower their supply teams with the tools they need to be successful with this ongoing digital transformation. It’s unreasonable to expect any one person, or group of people, to keep up in their head or with spreadsheets with daily sales for many SKUs. Digitisation is the future. Implement intelligent systems that support the work of your inventory managers, and free their bandwidth to tackle problems that computers can’t solve.”

Another upcoming focus, in Sage’s view, will be on increasing end-to-end visibility within the supply chain and distribution networks. “The ongoing pandemic has exposed the flaws of relying only on internal data and have propelled companies to begin thinking more holistically and factoring such data into the models,” she says. “By not taking a comprehensive account of market trends, cross functional and industry specific insights, companies are handicapping themselves, and are in a weakened posture when it comes time to outmaneuver emerging threats or unexpected risks.” 

Ball believes the value of having a very robust S&OP discussion – where companies talk about the contingencies and address not just financial risk but also supply chain risk concerns – are going to be extremely important going forward. “This is because until we return to some level of stability without the openings and closures and full shutdowns and partial shutdowns, we need to be as flexible and ready for the unexpected as we can be, both business and operationally,” he says. And to facilitate this, Ball maintains that a demand forecasting & planning system with the best algorithms for the task is an important consideration.

Payne reflects that we are still at a fairly early stage with AI and ML, and still at a fairly early stage with digital supply chain twins. “At the moment, that’s more for larger leading companies but this will certainly percolate down through the ranks,” he says. Payne also thinks resiliency still has a way to go. “There's a lot of hype around that in the market at the moment and it can be quite misunderstood, particularly when it comes to planning. So, I think we’ll be seeing more development and uptake around resiliency over the next couple of years.” 

Ogg considers that while Machine Learning is becoming more widely available from software vendors it is not always clear what it is doing under the covers, and what logic it uses in getting results. “There are also different viewpoints on it benefits e.g. Demand Planning Machine Learning versus more traditional Forecasting algorithms,” he says. “There are many uses in the area of Production, Distribution and Inventory Management and these often require additional sources of data to widen out the context from more traditional parameters so that the Machine learning can be effective. Use of Machine Learning will become more widespread because of the additional benefits it can bring when used in the right way to solve the right problems. It is likely to become more standardised, particularly as businesses become more confident in its application and deployment.” 

Will there be further changes with regard to end-user requirements over the coming year or two? Ogg believes the fundamental requirements will probably not change as they are all about supporting the planning requirements of the business in terms of delivering to customers the right product of the right quality at the right time and right cost. “Users will continue, however, to want to work more efficiently, more intuitively and from any location at any time without feeling isolated,” he says. “Many planners have workarounds to do related activities and analysis and will want this to be all in one system. It is not uncommon for a business to have a central data warehouse and reporting solution which does not refresh until overnight, whereas a user would want key planning analysis and reports available within the planning system on demand. Whilst there are many automated processes and algorithms that support the planning process, user still require different degrees of manual intervention and collaboration with other users and functions within the business. Therefore, the more that can be automated with users being presented with specific exceptions along with recommended actions will be of benefit. 

“Many algorithms, particularly in the area of production planning and scheduling, can appear to be like a ‘Black Box’. A plan is presented where the user cannot easily follow the logic that the algorithm has gone through to deliver the result. Users want this to be clearly presented on screen so they can easily see what has been planned were, when and why if it is not the result they were expecting. Easy navigation of exceptions, particularly when complicated Bill of Materials and Routings are in place where the cause of an exception can be many levels away. Altogether users will want a simpler and more efficient experience with as much as much work as is reasonably possible being done by the system.”