First law of forecasting

LinkedIn +

Is your guess as good as your competitor’s? Supply chain planning relies on the accuracy of the forecasting process. Malory Davies asks what it takes to improve the process.

The first law of forecasting is that forecasts are always wrong. The important thing is to understand how wrong the forecast is, and how to improve the accuracy to a point where realistic planning can be achieved.

The development of IT systems that can measure and manage every metric in the supply chain has opened the way to better forecasting. But, it would be a mistake to think that IT alone is the solution, according to Oliver Wight partner, Liam Harrington.

He argues that the critical factor in producing accurate forecasts is not IT, but people.

“Ownership and competence are vital. Many organisations still mistakenly treat forecasting as a supply process but true demand management should be the fundamental driver for business growth and strategy, and the needs and wants of the consumer must run through the bloodline of the entire supply chain.

“Effective forecasting or demand planning, hinges on a clear understanding of where value is created, from the customer’s perspective. By definition, forecasting must be led and managed by those closest to the customer and those who have the greatest understanding of what is happening at the point of consumption; that’s the sales and marketing team.

“Process comes before systems too; excellent process design is essential if you are to avoid the pitfalls. Not only does a consumption-driven demand management process remove unplanned issues, it improves the customer experience and optimises cost-to-serve, providing distinct competitive advantage. Without such a process, companies will find operating in the current business environment incredibly challenging.


“Equally, businesses
need to understand that not all customers
and consumers are the same. Market segmentation allows an organisation to recognise and anticipate different future consumer needs but often this link between segmentation and demand management is overlooked, especially when forecasting runs in parallel to sales and marketing, rather than as an integrated process. Sales and marketing have to lead the segmentation process, as they do demand management, as it is their insight upon which segmentation depends,” says Harrington.

Improving the accuracy of inputs into the forecasting process is clearly an important element in strengthening the process.

Mark Morgan, regional vice president of EMEA at JDA, identifies two measures: “Firstly by improving the input at a data level through history cleansing and management capabilities such as lost sales calculations, event identification or outlier detection. Secondly, by improving the input at process level through increased accountability of relevant departments such as sales or marketing. This can be achieved through using an integrated business planning tool. This will, among many other things, ensure that every participant in the forecasting process can be held accountable.”

And Jonathon Ogg, senior business consultant at TXT e-solutions, highlights a number of issues: “The team is important, along with defined and mature processes and clean data. Visible collaboration, supported by appropriate and integrated systems that are as automated as possible – let the planners do the planning and not have them moving tedious amounts of data into spreadsheets.

“Capture more information from the customer. Lots of companies have access to excellent data from their customers, but they often do not use this in the planning process – often due to technical challenges in integrating customer/supplier systems.

“A platform of companywide KPIs (Key Performance Indicators) that support the business strategy. Remove the inter-department conflict and reduce the fire fighting. Focus on the key products and customer areas that add the most value to the business.”

Nevertheless it is easy to overlook factors in the forecasting process. Stephen Szikora, IT director at NFT says: “As I see it the key factor overlooked is data translation. Most forecasting tools work from a base of sales units, which need to be translated and fed back up the chain to production and sourcing. While forecasting may give reasonable output and hence input figures, generally the physical handling element in the middle is overlooked or given only shallow functionality. All too often we see incompatibilities in data, which affects optimisation of plan to the forecast, for example inaccurate dimensions, poor case – or even pallet – configuration management, and complexity in different handling units in the different channels.”


And Phil Davies, solution consulting team manager at ClickSoftware, reckons that the iterative quality of planning is most likely to be overlooked when it comes to the forecasting process. “A good forecast is never static; it’s always changing. Forecasting is significantly improved with the more data you have. And the closer you approach the day of service; it increases in both granularity and accuracy. For example, you might be looking at data quarter by quarter next year, but as you get closer to the date, you can look day by day and so on.”

Ogg says: “Interestingly, it is the statistical forecast that is most often overlooked. A little more analysis and effort at the beginning will increase confidence. We often find that the first statistical forecast produced doesn’t meet expectations and is therefore overlooked. This reduces the on-going benefits of automating large parts of the product portfolio from a forecasting perspective.

“Another thing we are hearing is that metrics are so different. An account manager measured on volume will quite rightly adjust volume to meet target. The impact on the financial forecast can be very significant and margins can be eroded. If the forecast is not tied to execution this can result in costly changes in manufacturing or shipping at the last minute and result in reduced customer service.

And, he says: “It is important to understand the impact of one-off events – such as promotional activities, or large contracts – by either analysing previous similar events, or close collaboration with the client. Detailed attention to these areas is vital. Some businesses can have up to 80 per cent of their demand driven by promotional events, especially food businesses.”

Christine Hansen, product marketing manager at Epicor Software Corporation, highlights the importance of understanding and responding to product velocity.  “With shortening of product lifecycles, as consumers increasingly look for new innovation and thirst for the latest capability, forecasting needs to consider and plan accordingly.

“Not only are lifecycles becoming more volatile, the complexity of building and managing the supply chain for products has increased as well.  Yesterday’s product may have had ten components from three different suppliers in tow, but today’s has many more options for consumers to consider and therefore more product components and suppliers to manage. Forecasting these products in tune becomes more complex,” says Hansen.

One area that is frequently discussed, but less frequently applied is collaboration with partners in the planning process. “Collaboration encourages greater transparency and integrity throughout the supply chain, from the planning processes and beyond. It’s becoming more commonplace to work in this way in our industry,” says Szikora.

“Early warnings can benefit all parties as earlier planned resources can be smoothed. More volatile and poor plans will cost more to execute.”

And, says Davies: “There are many parallels when it comes to the service and supply chain. For the service chain, collaboration is of primary importance, and this is just as important for the supply chain. The service chain is all about the six ‘W’s’; Who does What, for Whom, with What, Where and When. Some of these factors are managed by different partners in the service chain, so without collaboration the chain will break down, and what use is the engineer on the wrong site with the wrong parts?”


The past few years has seen the development of increasingly sophisticated technology to help in the planning and forecasting process. So it is perhaps surprising that many organisations still rely on spreadsheets. Ogg says: “Even in companies with quite sophisticated systems, we see massive use of disparate spread-sheets, access databases and home grown systems that are reliant on the knowledge of a few individuals. Having one platform to manage the supply chain and forecasting processes can result in improved efficiency and customer service benefits. Deciding which system is appropriate to promote collaboration is also key. At TXT, we give the familiar Excel user interface with the power of a system that is secure, controlled and enhanced to support planning activities.”

And Davies points out that one advantage of using a solution to facilitate planning is to ensure a standard process is implemented, that’s repeatable and quick to apply.

“Also, if someone leaves the company, the same solution is still in place that everyone understands.”

Having put an appropriate system in place, it is important to be responsive to rapidly changing market conditions. It’s up to companies to ensure that their systems can respond quickly, says Morgan. “This applies to both detecting changes in demand as well as the ability to react to these changes. The latter, however, needs to respect the overall business strategy. It is here where many so called rapid response solutions struggle. A fast and feasible response is easily produced. However, for it to be aligned with the business strategy requires deep domain expertise in optimisation technology.”

Szikora says: “In terms of business systems, supply chain planning using alternatives and flexible, shared-risk based contracts helps to aid responsiveness to rapid changes in demand. In terms of IT, this can really only be achieved through selection of the right tool, configuration and real time data feeds, as modifications may take far too long.”

Hansen points out that modern ERP systems include many advanced planning and forecasting features and take into account supply of materials, the state of the supply chain, and orders being placed at the very moment the forecast is run.  “This level of real-time visibility across the entire business means that the systems accurately reflect the real world leading to much more accurate planning and forecasting.  It also means they can be very dynamic as when orders are placed or cancelled the planning systems automatically adjust their schedules.” 

Executive sponsorship

Ogg emphasises the importance of the organisation – especially to avoid silos that may have existed before, all having different goals. “The importance of executive sponsorship cannot be overstated. Also, a single platform for the planning information along with common KPIs that are shared throughout the business will enable responsiveness.

“Often, companies cannot react to the changes quickly, as there are layers of decision-making; or systems that make it difficult to see what we can buy/make/ship to meet the fluctuation. Can we offer a comparable level of quality/service/cost? Are we all pulling in the same direction?”

Davies says: “Weather is the biggest rapid change factor but there are other factors which can cause significant changes. For example, Satellite TV companies can experience large spikes and drops depending on when certain campaigns run throughout the year. Aligning planning in different departments (such as marketing and sales) and having company-wide visibility can help companies to cope when such times occur, and allow for proper staffing that can scale up or down with anticipated demand.

“Finally, automating as many processes as possible allows systems to deal with situations without the need for human intervention. For example, using a system allows companies to create a scheduling policy to cancel any scheduled maintenance work and concentrate on customer outages.”

Ten point plan

The ten step approach to achieving better forecasting:

Demand Solutions has prepared a ten point plan for improving forecast accuracy. It says it is an essential step to lowering inventory levels and the associated carrying costs and scrap. “Better forecasts also ensure that the right product is in the right place at the right time, delivering improved customer service levels and reducing the need for fire-fighting and expediting. In these days of complex and extended supply chains, increasing forecast accuracy is a competitive necessity to achieve top-line growth as well as bottom-line profitability.”

1. Forecast as close to the customer as possible
2. Forecast using demand — not sales
3. Identify and separate different demand streams
4. Understand and modify history to remove the effect of exceptions
5. Identify and remove company-inflicted distortions of demand and supply parameters
6. Forecast at aggregate levels and explode to lower levels
7. Identify and manage records with limited history
8. Communicate, share, and use the data to drive your sales and operations    planning process
9. Make forecasting a key business step with clear accountability
10. Measure performance, publish it and continuously improve
Source: Demand Solutions

Case study: Direct Wines aims to align marketing and supply chain

Direct Wines, the online wine retailer, has deployed cloud-based planning, promotion, forecasting and fulfilment software from Predictix, to improve alignment between marketing, planning, merchandising and supply chain operations.

UK-based Direct Wines sells more than 40 million bottles a year to 750,000+ customers in Australia, Hong Kong, and Germany as well as the UK and via partners in the US. It is an omni-channel business, selling via catalogue, online and through an estate of retail stores under the Laithwaites brand. It also provides wine clubs to customers of The Wall Street Journal, Zagat, Virgin and the Sunday Times.

The retailer generates high sales volumes for products that are often produced in limited quantities by brand or vintage, and then sold individually or in a multitude of mixed cases. This, together with dynamic and sophisticated direct marketing, created the need for a particularly flexible solution that enabled the company’s supply chain to respond in real-time to changes in actual and anticipated demand.

It turned to Predictix for a suite of cloud-based applications covering: merchandise financial planning; creating and managing mixed wine offers; promotion planning; offer planning; and demand planning.

“Predictix has helped us start to break down the walls and eliminate the silos between marketing, merchandising and supply chain,” said Direct Wine’s chief executive officer Simon McMurtrie.

“For the first time, we have an end-to-end solution that fits our unique business model. We set a number of ambitious goals for this initiative, and it has been a great team effort between Direct Wines and Predictix.”

It reckons it now has better alignment between marketing, planning, merchandising and supply chain operations, with a structured approach to range planning and the ability to plan offers that will meet revenue targets in the overall marketing plan.

“Direct Wines also has greater visibility into forecasted demand and projected stock requirements across all channels. This will help drive greater inventory efficiency and reduce the need for case and SKU substitution by ensuring that Direct Wines’ buyers purchase the right wine in the right quantities at the right time on a global basis.”


Key difference between S&OP and Integrated Business Planning:

While ownership by sales and marketing is fundamental, integrating forecasting with financial planning is equally important and this too is often overlooked.

Harrington says: “In many cases, forecasts are merely a volume-only statistical process based on overall demand history and projections. Integrated Business Planning has become the process of choice for smarter organisations, since one of its key differentiators from traditional S&OP is the financial integration it provides, allowing the business to operate with a single set of numbers.

“At the same time, forecasts must be based on real demand history rather than just shipments, and any non-repeating historical events must be edited out.”

He points out that although forecasting isn’t itself a supply process, integrating demand management through the supply chain is a must. “Planning in isolation is debilitating while collaboration between supply chain partners in the planning process is transformational. The key is to establish and prove the relationship, with multiple touch points throughout the partner organisations (not just between sales and purchasing), making it resilient to any future changes in personnel.”

Share this story: