Demand Planning Panaceas

Planner, Heal Thyself!

Demand Planning Pain Points are legion and can be found described in some detail in the article below but to summarise, they can include spreadsheet dependency, data inconsistency, forecast inaccuracy, high licensing costs, limited functionality, slow system speeds, chaotic cycle times and unhappy planners.

Panaceas = Get Fitter & Faster, Healthier & Smarter!

Fitter & Faster means obtaining, loading, creating, manipulating, allocating, reviewing, approving and issuing of data more efficiently; a faster forecast cycle. It means using collaboration and exception management, resolving existing workarounds, automating tasks and encouraging less intervention. It means making good decisions quickly and being able to re-forecast anytime.

Healthier & Smarter means improved accuracy at the right levels and adding insight to the business rather than treating forecasting as a task. It means knowing your limitations and how to overcome them with the tools and techniques at your disposal. It means understanding what better practices to apply and what to abandon, being flexible and adapting as data, functionality and capabilities change. It means observing what your competitors are doing in order to gain advantages over them.

How to achieve these fabulous goals? How to go from Planning Zero to Forecasting Hero? What are the Panaceas to make this happen? The answers can be found in

Information Technology, Process, Strategy & Resource

That's right, not just Technology, though frequently it is the main catalyst for change. The priority and mix of those elements can vary greatly depending on business type, culture and maturity but to start with you will need good clean data being managed by enthusiastic and well-trained planners, working efficient and structured processes.

Information Technology

Reach for the Sky!

Yahoo Mail, Dropbox LinkedIn, Microsoft 365... Cloud computing has been with us for years and is evolving all the time. Solution Providers like Oracle, SAP, Microsoft & Sage are all offering sophisticated Cloud business solutions. The benefits of moving include Performance, Cost, Security, Flexibility and Access. It is feasible to build Hybrid Cloud Solutions using the best-in-class solutions from the myriad of providers.

Cloud solutions utilise subscription evolution rather than static license so that as the product develops, you get the changes included. Cloud applications tend to be leaner and meaner in order to be quick and easy to use. Cloud solutions have virtually zero down-time and can easily be turned into virtual solutions for testing and tuning.

Any Cloud Planning Solutions worth its salt will have plug-and-play approach to Application Programming Interfaces; all the Planning providers are keen to connect with any Legacy or competitor ERP systems. The first stage Panacea to planning problems is a Cloud Demand Planning module to replace those creaking Spreadsheets.

Unity, Standardisation & Exception

Hark! I hear you cry. What about my unique custom requirements that won'/t be allowed in a Cloud solution? Well, your local and specific needs may well be what is causing the problems in the first place. It's time to buckle-up and properly assess what you do and why you do it:

Why Forecast? How to Forecast? When to Forecast? What to Forecast? Where to Forecast? Who Forecasts?

The best cloud solutions will easily allow you to define your own user interface functions from Worksheets and Graphs to data streams and KPI calculations. Make sure that you define your solution on strong foundations.

Unify your Planning Structure (Dimensions, Hierarchies & Levels) to reflect and connect all the big business forecasts. When levels and data can be referenced, allocated & aggregated and reported against from Annual Budget to Production Plan your planning will be faster, more accurate and a lot easier to communicate.

Configure worksheets, graphs and dashboards and align them to your planning process steps. Build scheduled workflows & automate exception calculations. Apply colour coded life-cycle information, enable filters by Demand Type, ABC Classification, Criticality, Promotion & Causal. Use alerts, information tiles and sort functions with Pareto of Volume, Accuracy, Cost, Revenue and Margin.

Structural unity, User Interface standardisation & exception management helps in training and communication - especially when Roles & Responsibilities are clearly defined. It makes overrides, notes, and approvals easier to apply and check. It turns data into information and enables planners to work smarter by analysing by impact rather than searching for it in endless rows and columns. Already your planning is getting fitter and healthier.

Distributed Data

Distributed data over fully Integrated (where all planning is done inside ERP) still make sense unless your dataset is small enough to not slow Finance screens to a crawl every Month-end. Distributed will mean collecting data either in small bits every day and/or in bigger chunks every week. This is why the adverts and specifications usually say "near real-time" data but if your forecast cycles are monthly or weekly this shouldn't be a hindrance.

The Right Classification

Lighten the load and run your forecasts with Modules, Cubes, Plans or Profiles within your solution that are segmented according to data type, horizon, purpose or criticality and then, organise planners by the segregated plan types so that their analysis becomes concentrated, their methods more specialised and the outcomes faster and more accurate.

Consider processing your forecasts by data, horizon and interval needs. Smooth demand with consistent forecast can be managed with greater gaps between planner assessments so that the erratic and critical combinations can be looked at more often. Don't do the same thing cycle after cycle if there is no need!

Get faster and smarter by utilising cannibalisation settings where old and new combinations link automatically, set automatic lifecycle activation & deactivation and use a hybrid of Push (Forecast) and Pull (Point of Sale or Order Demand) to maximise efficiency. Automate the archiving of data streams and the calculation of KPI's.

The Shock of the New

New Combinations are always difficult to manage whether it's products, customers, organisations, distribution centres, planners or salespeople. Mergers and Acquisitions, new categories, attributes and data streams are all problematic things to add into an existing solution.

There are two basic ways to create combination that to be forecasted: A/ Create every possible combination regardless of whether the intersection of data is feasible or not and B/ Create combinations based upon data (such as Shipment History).

Option A means creating a massive data set and will likely create terrible performance and management issues. Option B means that you can only forecast combinations that have been shipped, which is hardly ideal when you need to forecast something new. New can also mean old or existing (see table below):

Most Demand Planning solutions will provide functions to administrate new combinations and if you have a high volume of NPI's be sure to properly evaluate how they are created in any tool you are considering since the solutions can be labour intensive. My recommendation would be to make new combination creation part of Master Data Management and not Demand Planning.

Statistical Forecasting

A Statistical Engine with models and settings that can be accessed and tuned is a very powerful tool to have in your planning box and is often considered the main driver to resolve forecast performance and accuracy problems. Blended Models with Flexible Hierarchies are the most sophisticated solutions to reach for but beware the setup and maintenance of these engines!

Being able to use multiple plans & forecast profiles reflecting different types of business can help reduce overrides which in turn speeds the forecast cycle and increases accuracy. Engine run times should be fast (hours) and can include promotions, causals & seasonality and highlight outliers to create baseline and promotional forecasts. Simulations and instant re-forecasting have the capability to transform your forecasting capabilities.

The reverse is also true! Be sure to evaluate the suitability of your data for statistical forecasting. Consider the possibility that Demand Driven Management (where demand signal rather than forecast determines what to make & buy) is a leaner and smarter way to manage Supply Planning activities but not all Planning Solutions efficiently enable the execution of DDMRP.

At the very least though, a statistical output can be an invaluable guide and comparison to your manually created plans.

Process

Process Automation

Use workflows to automate as much of your planning processes as possible: Data Collection and Cleansing, Segmentation, Forecast Creation, Plan Comparison, Exceptions, Approvals, Uploads & Archives can all be reduced in time and effort through scheduled automation.

Consider this: A data collection function that conducts analysis accuracy and if a combination is acceptably accurate it flows in to the 'Good Forecast Cube' and if accuracy is not acceptable it flows into the 'Check Forecast Cube' and if a combination is repeatedly inaccurate (say, three cycle in a row) it flows into the 'Difficult Forecast Cube'. Fitter & Faster, Healthier & Smarter!

Cycle Time

Cycle time reduction has to be one of the most desired change after improved forecast accuracy. Getting from Months to Weeks or Weeks to Days can be done but it is rarely achieved though one element. Transforming from Spreadsheet to Planning System will make forecast generation faster but the checking, adjusting and approving in a tighter timescale will likely require significant reorganising the business calendar.

Meetings will need to be realigned with all stakeholders; identify the start and end points of the process and then re-align everything inside and be prepared to really shake things up in order to squeeze into the tighter timeline. If S&OP is in place, the changes will impact a lot of people and processes. If S&OP is not in place, the process can probably be shortened easily and then you can implement S&OP with it.

Tips to get faster and fitter: consider moving away from static saved forecast to a constant live forecast (with snapshots for archiving). Reduce approvals. Throw away the checks-in-boxes; empower and trust your planners. Think about cycle time by data segment because maybe not everything has to be forecasted every cycle.

Strategy

Accuracy improvement comes from being smarter and knowing how to apply those smarts. Assess your planning maturity (know your capabilities) and go for targets that are within your reach. Once the easy fruit has been picked then move onto the next category. If planning education is needed (maturity growth) then go, get it!

Frequently re-assess the level of your maturity, the difficulty and impact of your challenges. High Impact and Easy to Do first! Regular evaluation allows you to validate your planning journey and achieve improvements of cycle time and forecast accuracy more efficiently.

Be the Change

Forecasts are all about managing an uncertain future but how many companies manage their technology, resources, processes and strategies with the same effort that they apply to their forecasts?

If you haven't yet got to data like The Internet of Things, Biodegradable & Sustainability, Ethical & Social Responsibility, Country of Manufacture, Distribution & Reverse Logistic History and so on, then start preparing, because this data will soon be very important to measure.

How will this type of data fit to your past and how will you wish to see it recorded and managed? It could be by Hierarchy, Data Stream, Attribute, Plan, Forecast Profile or Planner to name a few options. Prepare to change what you do and how you do it in the future. Create placeholders inside new solutions with a view to utilise them when the data and capability arrives.

Resource

Forecasts get completed quicker and more accurately not just because of a fancy tool but because of motivated planners with a passion for improvement who buy into and deliver on planning maturity growth plans. Start simple and unified but develop into complex, segmented and sophisticated approaches to identifying and resolving planning conundrums.

Unity and Standardisation is the starting pointy but greater accuracy will come from increased maturity and being able to create on-the-fly worksheets, graphs and dashboards. Drill-in and out functions allow planners to follow their instincts and search out issues and solutions. Breaking out from standard process and analysis to create unique and creative data streams, calculations and scenarios.

A Cloud tool that will allow this is clearly ideal, but planners need more than a technical platform; they need the training and permission to try new things. Consider research and development of ideas and approaches inside your planning team.

Communication is the key

Cloud applications include in-built instant messaging functions, and some with full internal social media capabilities. Use these elements to aid virtual collaboration across teams. What-if scenarios, writing notes with uplift explanations etc. become much easier to share. The Demand Review and Approval steps can be done on-line - no more paper printouts!

Create Customer Portals so that external parties can add their forecasts directly into the system. Start building the Annual Budget in the tool from the last approved Forecast. Collaboration across business, customers & third parties become achievable. Apply these functions and approaches and yes, you really can become leaner, fitter, faster, healthier and smarter with your planning.

A small selection of Cloud providers as a starter for ten:

Anaplan, B2Wise, DemandWorks, FuturMaster, Infor, John Galt, Logility, Lokad, Oracle, o9, SAP, Summence.