Managing the demand curve - Demand Forecasting/Planning/S&OP Technology report


Manufacturing & Logistics IT spoke with a number of experts from the vendor and analyst communities about recent developments and key discussion points within the Demand Forecasting/Planning/S&OP space – including those concerning planning unification, best of breed, integration, Cloud/SaaS and Big Data.

Major developments, subtle enhancements and even more disruptive innovation continues apace within many technological disciplines on a very frequent basis, and the world of Demand Forecasting/Planning/S&OP solutions is certainly no exception. Indeed, there have been some sizeable steps forward since our last special technology report on the theme.

So, without further ado, Richard Watkins, managing director, Delos Partnership, begins our discussion by emphasising that Integrated Business Planning (IBP) now should embrace all parts of a business into a single plan, which can be reviewed through simulations to help manage the increasing levels of uncertainty faced by all businesses. "The unexpected is now the new expected," he remarks. "This means integrating financial plans, people plans, IT plans and engineering plans with the simple integration of Sales and Operations.

However, the real challenge is to ensure that the performance measures which drive business behaviour are correct. Over and under forecasting – which cause inventory and shortages respectively – are normally driven through bonuses which encourage this behaviour. Thus, over selling and under-forecasting gets congratulated. Sandbagging forecasts are caused by lack of trust of the supply side. Hence, Sales & Marketing should be rewarded on the basis of the accuracy of their forecasts. Where this is done, inventory levels decrease and customer service improves."

Nick Castellina, vice president and research group director, Aberdeen Group, references the theme of connected, collaborative and holistic planning. "Individual aspects of planning should not exist within a silo; they all need to be connected to make plans that are going to allow an organisation to drive profit margins and drive everything to work in lock-steps," he says. "If you are forecasting demand and you want to ramp up production but don't have access to the material you need to meet that production, you are still going to have gaps. So, organisations are trying to drive their planning tools to more users within the organisation and rely on data to make plans that can change in an agile manner on-the-fly."

In terms of the main drivers for these types of changes, Castellina believes it is really about users having access to more user-friendly tools than they have had in the past. "I think this requirement derives from the typical challenges that come along with planning today, using the tools that organisations have had for a long time," he says. "When you are using spreadsheets to manage your plans, there are many challenges – particularly within larger organisations that are using multiple spreadsheets and multiple tabs that are supposed to be connected but often aren't connected all that well. This makes it difficult to work to a plan that is based on business reality – you may be taking a long time to create plans, waiting on people to provide their inputs, and dealing with older versions and different types of formulas that can be broken. So, there are challenges that come with the types of tools. The market's newer planning tools are allowing organisations to get away from those challenges."

Nick McGrane, managing director, K3 Syspro, considers that the planning, scheduling and execution piece that forms part of the enterprise solution is key at the moment. "ERP solutions like SYSPRO now ship with extensive forecasting, inventory optimisation, advanced planning and scheduling and MES functionality," he explains. "This enables the quintessential 'one version of the truth' maxim we talk about so often, as well as assuring maximum efficiency in both planning work and executing it on the shop floor, with real-time feedback supporting the supply chain and the financial well-being of the business."

Online analytics

McGrane adds that there is an eye on the future too. "Online analytics may well replace traditional statistical algorithms in predicting future sales and stockholdings," he says. "That smart, connected 5-axis machining centre may have already told the tooling supplier it has run out of bits by the time you get back to your office, or told the machine manufacturer it has developed a vibration when doing a rough pass – on aluminium only..."

McGrane makes the point that many of the things happening now have been available for some time. "SYSPRO has had fully integrated forecasting and inventory optimisation functionality for nearly 10 years, and advanced planning and scheduling functionality for nearly 20," he says. "I do, however, think that the talk of Industry 4.0, Big Data and the like have caused people to look around a bit more at the extant systems and data they have, and are looking to extract value from those circumstances."

Pascal Garsmeur, product manager for demand & supply chain planning, DynaSys, comments that, in terms of what will be increasingly important from a technology standpoint will be the added value coming from:

  • IMD (In Memory Database) that provides almost real-time simulation planning to assess options and plan more rapidly as the environment changes.
  • Advanced algorithms allowing user interactions only when exceptions are detected, prioritising activity for those areas that require it.
  • Enhanced mobility provided by the latest HTML user interfaces which allow users to view data and collaborate seamlessly across any device in any location.
  • Advanced analytics with the ability to slice and dice data to make informed decisions.
  • Cloud strategy helping to leverage best in class technology whilst enabling end user organisations to be business focused rather than spending their time supporting operational systems, patching servers and other non-value add tasks.

In terms of the main drivers for these changes and developments, Garsmeur comments that DynaSys's prospects are expecting mobility and Web 2.0 as a standard part of their planning applications. "What was considered cutting-edge months ago is now increasingly seen as the standard for being invited to the table by organisations looking at implementing supply chain systems," he remarks. "Whilst ROI has always been an important part of supply chain projects, the current unsettled environment has increased the focus on this aspect greatly. How rapidly a prospect will recoup their outlay and start to drive 'actual' savings is becoming a key decision-making driver for prospective clients."

Garsmeur adds that change is the only constant, but the pace of that change is accelerating and this is leaving manufacturers and distributors stretched and without the ability to respond rapidly to these ever-shifting requirements they struggle to manage their supply chains and to satisfy their customers' requirements. "This unsettled environment has led vendors to focus on technological advances that deliver the ability to plan more rapidly and accurately, whilst deployment of solutions becomes quicker and simpler," he says. "Advances such as IMD for rapid 'real-time' planning, templated deployments based on years of best practice and many hundreds of implementations to streamline deployment, increased collaborative features and web access are all focused on allowing customers to more rapidly recoup their spend whilst improving their supply chain operations."

So, what methodologies should be embraced to ensure a more accurate forecasting and planning regime? Garsmeur makes the following points:

  • Implement a structured Sales & Operations planning process that is supported by the Executive team.
  • Be driven by exceptions.
  • Use different planning levels and levels of granularity.
  • Accurately classify products (Production Segmentation).
  • Use appropriate forecasting methods dependent on the product life-cycle phase (launch, maturity, decline).
  • Focus on making it easy for non-planners (Sales, Marketing) to add their intelligence to the forecast. This is a combination of the solution design (a nice and simple GUI) and the business process (don't make it too onerous – let the planners be responsible for creating the base forecast and Sales to add and review)
  • Invest in the training and development of planners. For example, CPIM to provide a broad understanding of supply chain.
  • Build a cross-functional culture to ensure all stakeholders work together, specifically planning, sales, manufacturing & finance.
  • Include forecast accuracy as part of bonus for the sales team to ensure they have 'skin in the game'.

And with regard to effective production planning & scheduling; how can manufacturers better balance orders with machinery and workforce capacity on the shop floor, and how can they ensure the right materials and right volume of suitably skilled people are ready for speedy dispatch to the right production area without any expensive production hold-ups? Garsmeur focuses on a number of key areas:

  • Ensure there is proper co-ordination of demand and supply using a Master Production Schedule (MPS). This ensures demand is matched with the capacity of machines, labour and material availability.
  • Combine efficient production planning with Lean strategies.
  • Execution should be properly prepared with anticipation of capacity, resources and inventory level. This anticipation is one of the main benefits of best in class production planning solutions.
  • To define the best plan, balancing demand & supply, powerful simulation capabilities are required. For instance, evaluating the infinite capacity plan versus the finite capacity plan (including stock policy, production constraints, service level...).
  • Evaluate the risks arising from inventory anticipation due to capacity smoothing.

S&OP and IBP

Moreover, how can companies roll-out best practice planning & scheduling methodologies throughout multiple plants? And how can both the business and operational arms of a business better manage production and demand planning for the good of the whole enterprise? Garsmeur believes this is the main objective of the planning process known as S&OP (Sales & Operations Planning) and IBP (Integrated Business Planning) to provide global supply chain visibility and the ability to create a balance between the demand and supply not just of one plant but for the whole supply chain (multiple plants, multi distribution centres).

"Ideally, this balancing act should include both value and volume to allow informed decisions," he says. "Once this balanced global supply chain has been created, the actual scheduling and execution should happen more rapidly and with fewer hiccups or unexpected surprises. By building the planning process as a 'core model', each plant can benefit from best practice and not having to re-design the process for every site. Using a planning solution that is deployed through the cloud, each site can utilise the core model with a rapid and stable roll-out of the solution."

'Full truck load'

Also, with regard to logistics professionals, how can they best plan and schedule their fleets in order to best serve their customers while maintaining 'full truck load' and scheduling the most fuel-efficient routes wherever possible? Garsmeur believes one way to achieve this vision is to run a distribution requirement plan that will take into account full truck loading, logistics cost optimisation and the simulation and definition of an allocation model shared between distribution sites and production sites.

Stuart Hall, sales director, Epicor Software UK and Ireland, observes that ERP systems, like most other software supporting the manufacturing industry, are changing. "What was considered innovative in effective integrated manufacturing management 20 years ago is now considered clunky, out-of-date and not up to speed with the frenetic demands and pace of the modern industrial sector," he says. "That's why the move towards advanced planning and scheduling software is vital. This software can, if used properly, be particularly effective in reducing human error and the capacity for customer order mistakes – giving those in control advanced warning whenever an order is placed and highlighting when they're running low on a product necessary for that order."

Hall adds that many manufacturers find themselves guilty of over-committing to meet customer demand; and, as a result, find themselves struggling to meet customer expectations. "This scenario can be avoided by having the right planning and scheduling software that provides managers and business leaders with up-to-the-minute business data helping to predict future bottlenecks, staff and resourcing issues, and possible order conflicts," he explains.

Artificial intelligence

Hall also makes the point that the artificial intelligence (AI) element of more recent planning and scheduling software frees up workforce capacity with many tasks being handled automatically. "This allows time to be spent more effectively within the business and help to keep costs down," he says.

Andrey Reiner, expert in asset management, Capgemini, considers the Internet of Things (IoT) and greater levels of connectivity – and about data being collected and meaningfully translated into information that can then produce more accurate forecasts and better plan demand in the supply chain. He adds that this type of methodology can also be used to better manage the workforce; those people who execute the manufacturing operations or those with responsibilities for asset management etc.

From a planning perspective, Reiner considers that better connectedness of data concerning the workforce, machinery, the supply chain and third-party logistics etc. can provide organisations with a better, more fully rounded real-time view. For example, he explains that there are now concepts whereby all of a company's key operations can be viewed in real-time on multiple panels. So, there is now greater real-time visibility and everything is becoming more connected.

In the case of the more technologically mature organisations and how they use their planning tools, Reiner explains that, in the case of workforce scheduling software, for example, people and tasks can be automatically scheduled on the basis of qualifications or of time availability etc. – so, there can be a considerably high level of automation within the system. However, Reiner points out that there will still be a need to interact with the system in instances where there is an exception; something that occurs that is not part of the normal routine. So, management by exception – interacting with the system only when something out of the ordinary or unforeseen occurs – gives more freedom to the company with regard to a greater level of automation. However, adds Reiner, it also requires more trust of the system on the part of the company.

Unification of planning

Tim Payne, research VP, Gartner, points out that one of the big trends Gartner has seen develop over the past three years is the unification of planning. "This type of planning model moves away from the practice of putting together a demand plan, then preparing a network plan and then the production plan, and from these plans putting together a schedule," he says. "People used to talk about integrated planning, but it's really about the unification of planning. Certainly, many leading companies recognise that breaking up planning degrades the quality of the plan. Planning likes to be joined up."

Payne then spoke about the importance of both horizontal- and vertical-type planning. "We often talk about horizontal integration – demand, inventory, replenishment, production planning, production scheduling and maybe even order promising and workforce planning. The idea is to integrate planning across the different domains of manufacturing, logistics, workforce, finance etc.; doing that internally within the enterprise but now also increasingly wanting to do that outside of the organisation – what we call multi-enterprise planning that involves customers and suppliers.

"There is also vertical integration to consider, because companies have planning happening at different levels of detail and granularity. It could be very short-term event driven; orders coming in on the demand side, things happening with the schedules in the plant, deployment decisions to be made in the network, delays from suppliers or in terms of shipments between facilities – this all needs to be planned. However, you also want that linked to your tactical plan. You have a number of resources and can flex them to some degree – the people, the material, the machines, capacities etc. – but you also have an optimal plan; maybe a blend regarding cost, service, profitability and so on, as well as a higher-level plan in terms of what I might want my supply chain to look for in the future. And all these things increasingly have to link together. So, both horizontal and vertical planning are important.

"A company might have a business plan and a strategy that links, for example, to its plans to build a new factory, new production line or the idea of bringing in a new product for a new market. This should then link down to when these things will be available and how they could be used, because you will need to execute on that strategy. For many years, the scenario has been about creating a plan and then handing it to execution. But things happen and there can be a lot of firefighting. So companies want this to be more intelligent and more responsive in a way that keeps them closer to their corporate goals.

"So, there is this vertical integration that has to happen, and if you look at some of the planning solution vendors' roadmaps they are trying to do the horizontal integration end-to-end, unifying their solutions. They are also working on more of the vertical integration. If they are in the middle they might be trying to go down and get more granular – something that we call respond planning. If they are already there they might be trying to go up and get into the optimised planning, which involves better use of resources. And there are vendors that are involved in the long-term functionality, which is what we call configure planning. So, many vendors are trying to build out more of those layers that allow for that type of vertical alignment; because that is what the market is asking for."

Single version of the truth

In terms of other drivers for change, Payne points out that end-users increasingly do not want lots of different solutions from lots of different providers. "They are fed up with the fact that they can't plan in an integrated way, they can't run scenarios across their supply chain, they can't collaborate adequately if everybody is using different solutions," he says. "They might also be using different data models, process models and analytical models, and so interfaces are firing all over the place. It makes getting a single version of the truth – a unified plan that people can work on – that much harder. Also, there are often gaps between the applications, which are filled with spreadsheets. So, there is that kind of unification drive."

Payne also addresses the topic of digitalisation, and what that might mean for individual companies. "We see two major groups here – first, many companies that are at a lower stage of maturity have fragmented planning and different parts of the planning process happening in different places, with spreadsheet trying to oil the wheels between those things. Digitalisation for them could mean getting to a unified planning solution; getting to something that everyone can work in to have a single version of the truth. Strictly speaking, this isn't really digitalisation; we call it a planning system of record. It's really about putting a robust foundation in place."

Payne adds that if a company is already at stage 3 maturity or above, digitalisation then often means automation of planning. "Many of those companies will say they have initiatives for automated planning, lights-out planning or people-less planning," he says. "We have found what that means is not automation of everything, but rather automation of different aspects. Planning is like a decision-making process – what am I going to make, where do I need to make it and how much inventory am I going to have to hold?

So, these companies are starting to think a little bit more about those types of planning decisions and considering which aspects can they automate. Maybe there are certain types of planning decisions they can fully automate. Then, they will look for machine learning and other capabilities and advanced analytics to enable that. Maybe in terms of other decisions – such as where they will need to collaborate on something and get some buy-in on what-if scenarios – they could possibly automate the preparation of those scenarios and then feed that into an appropriate collaboration environment."

The best of breed debate

Is there still an argument for best of breed planning & scheduling/demand forecasting-type solutions? McGrane points out that he has seen and used great best of breed products, and great built-in ERP functionality. "Given its proximity to the coal-face and multi-faceted functionality though, APS is particularly difficult, and requires strong leadership to adhere to a market-relevant roadmap," he adds.

Just like the manufacturing sector in general, and ERP systems, Hall reflects that the world of planning & scheduling is changing. "Unlike previous production and planning systems, today's modern ERP software automatically plans production based on several factors such as plant capacity, and available materials and personnel," he says. "In most cases, this will render best of breed solutions unnecessary.

Hall adds that some of this change of course comes down to a couple of relatively simply tenets. "Manufacturers need more than ever to ensure their ERP systems and Advanced Planning & Scheduling (APS) processes are the latest possible versions," he says. "At the same time, prospective purchasers need to properly research and analyse the system before buying it. It is crucial that manufacturers pick a system that will be suitable for their specific needs and challenges. Picking a tailored ERP system with the right sort of planning and scheduling capabilities might take longer, but will bring greater efficiency and practicality in the long run."

For example, Hall explains that while it might take some time to find a system that will support planning and scheduling across multiple factories or plants, this is a core requirement of many manufacturers – so it is essential that the chosen system incorporates this need. "Again, this comes back to choosing a system that is modern, agile and flexible," says Hall. "These qualities are vital so the software can manage several different sets of data and operations across different locations."

Castellina believes best of breed still has a role to play because in many instances the ERP tools may not have the required capabilities dedicated to planning. "If organisations are using ERP for planning it's probably because they think it's good enough, but I think they can get more value from specific planning tools in many instances," he says.

Reiner believes the progress that has been made by the best of breed planning solution providers is good in terms of the systems' ability to integrate with ERP. "When we talk of scheduling software, these planning systems will typically come with an adapter that connects to an ERP system," he explains, "so there is some minimal integration effort involved, but we know how the integration works so the integration effort will be minimal." Reiner has also witnessed some of the key players within the ERP systems space catching up in terms of the planning capability they offer; in some instances, through the acquisition of a best of breed planning solutions provide and in other instances through a rigorous R&D regime.

Garsmeur observes that most of the major Tier 2 ERP companies already use a best of breed solution approach with regard to Demand & Supply Chain Planning (DSCP). "A number of the larger ERP solution providers have acquired best of breed solutions (for example, QAD, Infor and Oracle) and developed interfaces with these solutions to their ERP solution(s)," he says. "Other ERP providers had developed best of breed solutions incorporating functionality from several acquisitions, such as SAP. However, the majority of mid-market vendors either partner with a specific specialist provider or provide weaker forecasting functionality contained in their core ERP products."

'Complete' suite

Watkins comments that many ERP systems providers now try to provide a 'complete' suite of products to cover everything in a company; from clocking on into the People and Payroll system to working out the Vision and Strategy of the business. "However, not all ERP systems are strong – from a functional point of view – in all areas," he says. "In the area of Forecasting, Promotional Management and Finite Scheduling, there are a number of smaller and specialist software providers who understand their specific area, and it makes sense to integrate these with standard ERP systems, which will carry out the basic transactions necessary to run a business."

Payne considers that the traditional distinction of saying do I go with my ERP vendor or do I go best of breed has really changed. "With regard to non-ERP vendors in the planning solutions market, roll back a few years and they would say we do scheduling or we do demand planning etc. However, there has been such an extension that they either are able to be what we call a planning system of record or it's on their roadmap. This is because end-users increasingly don't buy just a demand planning solution or just another type of solution.

They might think doing it with spreadsheets or trying to do it with their ERP system isn't good enough, and therefore want a demand planning tool. So, they want to buy that capability but that's only the first stage on the journey towards integrating or unifying their planning. Because they want to join up their planning if they are going to choose vendor A for demand planning, vendor A better also be able to provide the inventory, replenishment and production planning capability that the user is going to need further down the line in order that the user doesn't have to go with vendor A for demand planning, vendor B for something else and vendor C for something else."

Payne adds that companies are becoming smarter, realising they need a certain type of functionally today while also knowing this will need to be able to fit the next piece of the jigsaw and the bit after that if they are really going to build a solid foundation upon which they can become more differentiated or more innovative in the market. "So, the best of breed argument sort of goes away when you are trying to build that foundation – the planning system of record," he says.


Have ways of best integrating these types of systems with other solutions developed to any notable degree over the past year or two? Castellina believes that the development debate really depends on the particular software vendor in question, and the ERP system that a particular customer is using. "I think better integration is becoming more and more available over time as the planning vendors continue to develop their tools," he says. "That type of progress is something that will always continue."

Watkins considers that integration is getting easier, as the formats are simple and the speed of interaction can be much quicker. McGrane believes the changes are subtle and apply to feedback. "For example, it may have been acceptable a few years ago to dump MRP data out to feed an APS, schedule away and execute," he says. "Now, it is correctly proposed more often than not that the APS will perform a schedule, and these data (particularly dates) will be fed back into the MRP to drive it – perhaps as MPS entries – thus ensuring that supply and demand are lined up appropriately."

Garsmeur maintains that end users are looking for mobility and Web 2.0. "That is the reason why we are offering more and more Cloud-based full web solutions," he says, adding that integration tools such as Dell Boomi now allow communication between almost any systems – whether on-premise or Cloud.


Continuing the Cloud theme, has the Software as a Service (SaaS) model and the Cloud concept in general had any notable level of impact on the planning-related software solutions market so far? "Yes, definitely," remarks Castellina. "As I mentioned when I spoke about holistic planning and connected planning, you have to work across geographic boundaries and across functions. Having a Cloud solution makes it easier for your employees to access the system, so I think Cloud technology certainly has a significant utility within the planning and forecasting function."

Strategic choice

Garsmeur also believes the emergence of the SaaS/Cloud model has had a huge impact on the planning related software solutions market. "Moving to the Cloud is more than a deployment decision, it is a strategic choice and goes to the heart of how companies define their business," he says. "DynaSys Cloud DSCP provides the freedom to companies to focus on their products and customers without the distraction of having to administer their Demand & Supply Chain Planning solution. The Cloud vendor supports and optimises the hardware and systems allowing clients to free up their valuable in-house resources to work on more valuable strategic projects and accelerate their journey towards the Effective Enterprise."

McGrane comments that SaaS has made software much more accessible. "According to the annual manufacturing report, funding is the primary barrier to 67% of manufacturers, even if the return is guaranteed," he says. "By moving the Capex to Opex, they can mitigate the cost barrier."

Richard Watkins: The Cloud concept is helping companies support their global ambition to trade globally. There are still problems with accessing some countries using the cloud, but, on the whole, this is getting less of a problem. With Sales and Customer Orders happening all around the world, the cloud is making things easier.

Under threat?

Is SaaS and the Cloud a threat to the on-premise model? "I wouldn't commit to saying that Cloud is going to take over," says Castellina, "because there are organisations that prefer an on-premise tool and there are IT departments that prefer an on-premise tool. But Cloud is certainly a viable option and is something that can add tremendous value."

Payne reflects that the on-premise planning model is certainly under threat from Cloud, but he doesn't think on-premise will disappear from the market as a deployment model. "Some vendors will continue to offer an on-premise option and a Cloud option," he says. "However, I don't think the hybrid model is so viable. If organisations want more of a unified planning model they either need solutions that are all on-premise or all in the Cloud. They don't really want their demand planning in the Cloud and their supply planning on-premise. The two things need to be together because companies want a unified data model. As mentioned earlier, it really shouldn't be broken up."

Payne observes that all vendors talk about Cloud, but says it is important to bear in mind that there are different types of Cloud options available. "You may have been a traditional on-premise solutions provider and now you offer them as Cloud solutions. But this might simply mean you are hosting the solutions; it may be the same software and it may be the case that you haven't really done anything different – so you are essentially offering the same capabilities that you were with your on-premise solutions.

"On the other hand, if they are true Cloud solutions actually built for the Cloud, like the hosted Cloud models they will be Cloud with subscription pricing, but will usually come with better user interfaces, more social collaboration capability and will probably have an in-memory database. So, it will likely be faster and perform better. They will probably have different planning paradigms and will usually be much better at running multi-enterprise because there will be a range of integration protocols built in as part of the Cloud platform. Also, companies will be able to use the elasticity of the Cloud to scale up when they need to for greater performance etc. They don't really get that with on-premise solutions that are hosted in the Cloud – they don't perform in the same way."

One foot in the grave

McGrane's view is that on-premise has one foot in the grave. "I think it's a bit like electric cars versus oil powered cars," he says. "The advantages are there for both oil and on-premise but the companies that produce software and cars are pointing consumers in only one direction. I've heard through the grapevine that SYSPRO is building future versions for the Cloud and I know many other software companies are doing the same – just making Cloud software. Despite the advantages on-premise may have over Cloud, on-premise is quickly becoming outdated. Cloud is much more agile and suited to IR4 and the SaaS subscription model suits a lot of businesses. Operational expenses are much easier to work with than capital expenditure. With SaaS being on a cost-per-user model, businesses can forecast costs much easier and there are no spare licences going to waste. On-premise is usually a cheaper cost per user, but just less convenient."

Watkins considers that Saas/Cloud is definitely in competition with on-premise solutions, and is being presented as such. "But, if you look at the cost of a Cloud solution based on monthly payments, then the cost of on-premise is always very competitive/cheaper," he says. "This must be related to software companies trying to keep prices high for early adopters."

Garsmeur believes there is a threat for on-premise systems as more and more customers are looking for a SaaS/Cloud solution and the flexibility this gives them. "The Opex versus Capex model appears to be gaining significant traction within the supply chain sphere," he says. "A good illustration of this is the increasing number of vendors who are only offering Cloud-based solutions. Most of DynaSys' net new customers are opting to deploy their SCP solutions in the Cloud as they look to concentrate on their core competencies. Conversely, we don't see a strong trend of moving to a hybrid SaaS/on-premise solution."

Big Data

Is Big Data having an effect on the development and benefits of today's forecasting and planning solutions? "Not yet, in my environments," comments McGrane, "but you would have to think it is coming." Garsmeur makes the point that Big Data is having an impact through the use of new forecasting methods such as predictive analytics. However, he adds that there is not a direct link between Big Data and a good Demand Planning tool. "One of the reasons is the quality of data and the massive amount of non-qualified information available," he says. "As with all data, this is another input to the planning process and there is still a need for that to be reviewed by the planner."

Reiner believes Big Data has potential in the planning and scheduling world. "Big Data is really about understanding and contextualising information that is coming through in large volumes, and in terms of applying that to workforce planning I can't see at this point of time it having a great effect," he says. However, he recognises that there is a desire to evaluate what it can offer, pointing out that it can involve asking fundamental questions such as is the data correct and invalid, who owns the data, who understands what the parameters mean when it comes into your data warehouse and what can the data be used for?

"The potential is definitely there," he says. "In the case of smart metering, for example, you can class all the messages that come from the smart meters as Big Data and consider what meaningful questions can be asked of that data in order to inform our decisions going forward." Reiner adds that there is a role for organisations such as Capgemini, together with systems integrators and business consultants, to point out where and how user organisations can potentially derive major value from Big Data.

Castellina makes the point that Big Data isn't a technology as such; rather, it is a term for saying we are dealing with more data more quickly today as a business community. "The planning tools that are being built are designed to handle larger amounts of data and make that more consumable. So, I think on the BI side of things that is certainly an aspect that is being taken care of, as organisations like to be able to plan in a more agile manner rather than having to run reports overnight. I think faster computing options certainly do have an impact on planning."

Payne considers that Big Data is predominantly a maturity play. He elaborates: "You can increasingly use the data sources you have inside the company and get more granular. In terms of demand sensing, for example, you would really be looking at the orders rather than aggregating and saying I have a demand for 1000 this month and next month, and for the previous month I had a demand for 1200.

Instead, you can create a statistical forecast by crunching at the order line level coming – seeing the demand patterns. For example, there might be more demand on a Monday than on a Tuesday, it might drop again on a Wednesday and come back up on a Thursday. This can really help the short-term decision-making process – you will probably need to put more stock in the warehouse on the Friday ready for the up-kick on the Monday if you are going to meet your service levels. So, there is a Big Data element in saying there is a lot of structured data that you have inside the enterprise, which you can crunch."

New and external data sources

Payne adds that there are also new and external data sources that are starting to come to the fore. "You could, for instance, start looking at weather data and that could have a big impact on where and how you sell products. A lot of products are weather sensitive and different demographics react in different ways to that. Also, you might have social data coming through in terms of a sentiment that is being shown on rating websites. So, if you have your foundation working fairly well and you are running the business adequately, you can start to look at some of these other sources of structured and unstructured data and see whether there are other cause-and-effect patterns that you could unearth to help you to improve your plan."

AI and retail

Frost & Sullivan also states that exponential progress in artificial intelligence (AI) and machine learning, fuelled by the combination of Cloud, Big Data and new algorithms, is transforming the retail industry. Frost & Sullivan comments that as AI leverages Big Data to automate, predict and personalise, retail is testing and implementing these applications to garner robust competitive advantages. According to Frost & Sullivan, the key focus for AI in retail is customer relationships. In times of concerns for the retail sector in the UK where sales posted biggest quarterly fall since 2010, the refashioning of this industry comes as a breath of fresh air, with many opportunities to come.

'Global Artificial Intelligence Opportunities in Retail, 2017', new research from Frost & Sullivan's Connected Industries Growth Partnership Service, offers an overview of AI and its relevance to business in 2017. The study assesses the commercial viability and impact of retail applications for AI, either through integration with existing workflows or by creating new ones. It also explores strategies for navigating AI as a retail or information technology (IT) vendor, and the imperative for both to adopt a data-focused mind set. Key market participants include Amazon, Ocado, IBM and Softbank.

"Improving and refocusing the customer experience online and offline must be the guiding principle for all retail businesses," says digital transformation research analyst, Vijay Michalik. "Large tech-driven firms and retail tech start-ups are leading growth, particularly in e-commerce, and data network effects may mean that laggards never catch up. Older e-commerce and brick-and-mortar retailers must urgently adopt these technologies to regain their competitive footing."

Retailers and tech industry players are investing in AI and creating opportunities that will disrupt incumbents. Emerging use cases include:

  • Chat Bots and Virtual Assistants: These AI tools of direct customer engagement allow for a seamless experience when ordering products. Chat bots have question-answer and recommendation capabilities that make it a highly scalable yet personal sales channel.
  • Marketing and Segmentation: AI models can use data sets to predict and prioritise the most successful campaigns and channels, and provide these insights to decision makers.
  • Inventory and Supply Chain Optimisation: In addition to increased accuracy and timeliness over traditional systems, AI tools can predict future supply-demand scenarios.

"In the AI age, data is digital gold and data inequality will prove a major battleground," says Michalik. "Optimisation across all business functions will require an ever-growing pipeline of data collection that will demand new hardware, software and networking investments. The market must also pay attention to security requirements for AI and customer-analysis data collection."

Transportation: Planning and route optimisation

With regard to a specific solution within a specific vertical sector, Frost & Sullivan recently recognised BestMile with the 2016 European Frost & Sullivan Award for New Product Innovation, Based on Frost & Sullivan's analysis of the fleet management software market. BestMile is one of the first providers of fleet management solutions for autonomous vehicles. Its open source integration protocol means that its solutions are compatible with vehicles manufactured by various automakers, and its software with any type of hardware. It delivers on-demand and fixed route planning services through a B2B2C Cloud platform that enables the real-time automated dispatch of vehicles, route optimisation, and energy management. The interface for its platform is accessible to both operators and passengers, providing real-time updates and functioning as a booking portal.

"With no driver required to man the vehicle, fleet management becomes of strategic importance in order to improve capacity utilisation, reduce operational and marginal costs, optimise routes and vehicles utilisation." says Frost & Sullivan research analyst, Ankita Mukherji. "Improving connectivity between all modes of transportation, BestMile's solution will play a strategic role in journey planning services." The Award for New Product Innovation is part of Frost & Sullivan's Best Practices Awards which recognise companies in a variety of regional and global markets for outstanding achievement in areas such as leadership, technological innovation, customer service, and product development. Industry analysts compare market participants and measure performance through in-depth interviews, analysis and extensive secondary research.

Planning in the near future

What are likely to be some of the key developments in the world of Demand Forecasting & Planning solutions over the next year or two? Watkins makes the point that in the FMCG arena in the USA there is more around the development of forecasting consumer demand at EPoS level and then working out replenishment to the store from DCs and back to the factory using the usual algorithms for replenishment. "This has yet to take off seriously in Europe, but will do over the next five years," he says. "It requires a genuine commitment to be collaborative, and in turn it needs trust."

Garsmeur comments that the continuing focus on 'new' planning processes such as IBP and DDMRP seem to be creating a lot of interest at the moment within the marketplace. "How much of this is long term and how much is due to them being the current 'in vogue/trendy' areas within the Supply Chain market will be determined over time," he said.

Garsmeur adds that, longer term, IoT and smart things will interact to make decisions within certain parameters of acceptance. "If there are standard deviations of exception, then you'll want human intervention, but if there's not, then these things can just interact and essentially run it for you," he says. "Ultimately, it's about how we leverage these different and better data sets to improve forecasts, to better understand our forecasting and how demand and supply interact."

Hall believes Cloud-based systems and the continued proliferation of Cloud migration will no doubt have a large effect on planning-related software solutions in the future, and indeed in some cases are already starting to do so. He adds that there are several reasons for this; including ease of communication and ability to pull down your business's scheduling data at a moment's notice to any device.

Hall also comments that planning and scheduling software is a constantly shifting and changing landscape. "It can be hard for manufacturers to know which system can align most closely with their business needs," he says. "Because of this, systems that can be responsive and demonstrate flexibility and agility will always be sought after and in high demand."

Castellina anticipates more connected planning; connecting the individual aspects of planning to each other. He also foresees the continuing development of Cloud solutions from a planning perspective, and believes agility in planning and forecasting will develop further – "being able to make changes on the fly to ensure that you are doing things related to realistic business conditions".

Expanding the time horizon

Reiner anticipates more ongoing developments regarding solutions that can help companies to better understand their operations; not only today in real-time but also in terms of better understanding what will likely occur tomorrow, next week, next month and next year. "So, it's about expanding the time horizon and the solutions that support this need," he says. For example, he makes the point that if I am a service organisation and I am providing services to a number of companies and a company asks if I can service a new fleet they have, or all these stores that they have, as a service organisation I need to be clear about whether I have the capacity to handle this request. It shouldn't take me three months to answer the question; it should be a very quick response because otherwise I am likely to lose the business to somebody else.

So, I need to think about visibility in terms of longer term capacity – whether it's the capacity of my own workforce or the capacity of my supply chain; because third-party logistics companies have limits – they are asked to do all sorts of things that can affect the margins of the organisations they are supplying services to. "At some point, the supply chain can start to fail, so there needs to be a long-term view of what your demand will likely be and you need to be able to give organisations within your supply chain a view of what they need to plan for," he says.

"They have limited capacity and need to plan in advance. If the supply chain starts to break down it can have a knock-on effect sometimes involving liability and health & safety issues from a workforce management perspective. This is largely why long-term planning is so important." So, with regard to technology, Reiner explains that companies want a very quick and simple way of communicating between all the parties involved in the supply chain. "This need will probably result in more application development, more information being available to the people who execute work orders, and more information being available to people who supply services to you."

Payne considers that, in terms of algorithmic supply chain planning, we will see more predictive and prescriptive analytics. He also believes we will see a lot more machine learning, neural networks and deep learning. "This will help with digitalisation and with degrees of automation," he says. "A lot of end-user companies today are worried about the productivity of their planners. Many of the planners are doing lots of manual tasks and slipping in and out of spreadsheets, and as companies embrace new business models and look at digital products and services and new markets they are increasingly realising they can't go on like this; they can't have this low level of planner productivity where they are nurse-maiding the technology a lot of the time. That all needs to be done at a higher level of productivity. The planners can then do more value-added things such as looking at the different trade-offs and scenarios or developing their relationships with other key stakeholders within the company."

Payne adds that there is also an issue regarding talent; having the right people for planning. "If you have more machine learning and you are encoding more of the knowledge of your senior planners, then you are less likely to lose that knowledge when they retire or move on. Another issue is planner turnover. It can be hard to keep younger planners in the role – many don't want to stay with one company, they want to move on to something else. So, encoding more of that best practice knowledge into the algorithms means companies can hold onto that knowledge more effectively, and this helps them to retrain and be far more value adding that they often are today.

Add a Comment

No messages on this article yet

Editorial: +44 (0)1892 536363
Publisher: +44 (0)208 440 0372
Subscribe FREE to the weekly E-newsletter