Despite the importance of forecasting, demand planning software continues to be misunderstood by many businesses. Industry research has revealed many of us are left unsatisfied with our software, even though demand planning tools can bring major improvements in business performance.
Unfortunately, many companies continue to labor under the illusion that an ‘almost’ match is a good enough option. They either fixate on the most expensive products that has all the bells and whistles or they over simplify and assume any product will provide a decent forecast. The real focus should be on tracking demand streams and processes and finding the right technology that provides the right kind of support for the right operations.
Knowledge is power when it comes to choosing software solutions. Therefore, it’s a good idea to do your own research and read up on the most popular providers. It is also important to do your due diligence and come prepared with what exactly you need and understand the process and how it should work for you. To help with this, over the next few blogs I will provide insights and the key points to help you identify the kind of demand planning and forecasting system you need.
What Do I Need The Software To Do?
This depends. I know it sounds like a cop-out but it would be truly irresponsible for me to tell you this software will fit your needs without knowing exactly what your needs, maturity, budget, value chains, resources, data etc. look like. To this point, it is also irresponsible for any vendor to sell you anything without understanding your process and business. If they tell you a particular product is perfect for your needs, be very wary. And contrary to popular belief, it is not good enough if a system provides 80% of what you require. You do not need to settle -there are enough options out there for you to get what you need.
The first step any company should take when looking for a new demand planning software is determining your process, technical and system needs; your budget and other constraints; and a detailed list of requirements. Most systems will have the usual industry standard time series forecasting algorithms but maybe you also need some of the more advanced tools on the market. These include advanced analytics like machine learning, AI, Monte-Carlo simulation and Stochastic Optimization.
Going through the design process first gives you a systematic way to work from abstract concepts down to specific technical details and physical designs for the solution i.e. you start with the benefit you want yo gain from the system and work backwards until you get down to specific features.
Map Out Your Demand Streams In A Specification Document
To ensure that you select the right technology for your operation, it is critical to map the demand streams and the demand drivers. Consider a process map that helps visualize each steps including input and output, as well as the technology or information you already rely on that will need to be integrated. Techniques like causality, seasonality, top-down and bottom-up forecasting, and forecast-value add analysis are essential to the selection of the technology because whatever system you choose must support them.
Think about your customers. No not necessarily the end users of the company product but the intermittent internal users of your forecasting and demand plans. This can be operations, supply planning, finance etc. Each internal customer may look at your forecast at different levels of aggregation and different time horizons and may need different levels of forecast accuracy.
From here it becomes easier to develop a detailed functional specification document that provides direction and guidance for what you need and the must haves for when you start comparing solutions. In this document, simply list out what you require. Remember to include the models you use, data and input requirements, and process steps it needs to support, and the metrics and information you need out of it. Address collaboration and process, functionality, as well as technical requirements. Your company’s list may be 10 items or 50 items long. Take your time and include the must haves and the like to haves.
Typical requirements may include, but not limited to, the follow:
- Full spectrum of statistical forecasting methods. Demand plans may use more advanced methods and a holistic approach to statistical forecasting (e.g. ARIMAX, econometric, cluster analysis etc.)
- Forecast generation using dynamic multiple hierarchies and complex event or pricing models and use of what-if scenarios
- Overrides performed at varying levels by different users, with hierarchy reconciliation
- Manages demand planning consensus work flow
- Measurement of demand forecast error with limited accountability or understanding of causes (e.g. MAE, MAPE, WMAPE, Bias)
- Performance measurements to gauge the value of the demand planning forecasting process instead of just the output (e.g. FVA, tracking signals)
- Segmented and graphical results based on differentiating lags, hierarchy, segments, and inputs or process steps.
- Ability to utilize POS and external data signals
- Ease of integration and reliability, and high degree of support available
- Languages/global presence and support
- Supports multiple calendars (production, fiscal etc)
Staff Checklist: The People You Need To Run This Software
Another critical question in preparation needs to be do I have the right staff in place both from a planning and technical perspective to get the most out of this application? This part is really critical to the success of planning to prepare to look for a tool.
Typical internal resources may include:
- Project Sponsor/manager (40% – 70% FTE). Provides executive oversight and guidance. Works with software provider to develop project plan and manage day-to-day priorities. Reviews deliverables for quality and completeness.
- Functional core team members (15% – 45% FTE each): Represent end user and voice of customer, validate business requirements, define user workflows, validate functionality, provide testing.
- Technical core team members (5 – 15 days IT resources): Develop automated programs to extract and import data, act as SME as required in support of core team.
- Functional end-users (3 – 10 days per member): functional end user testing and training.
Note that we have highlighted demand planning and forecasting software only. For an Advanced Planning System that does forecasting as well as supply planning or inventory optimization, you will need to look at other options that have more features. With demand planning being more than just a supply chain process, and with advancements in technology and integration, do not assume it must be part of a supply chain or ERP system to get what you need. You no longer need a single solution and may be able to meet your organization’s needs with a standalone forecasting system. Predictive analytics and forecasting once again offer benefits as a separate system as long as you know what you need and understand how it will benefit your company.
Which brings us to the next question of what kind of return on investment (ROI) can I expect and how much will all this cost me? Subscribe to our newsletter or check back soon to see how I will answer this question.