Making Forecasting A Team Sport

Collaborative forecasting as a term has been in vogue for quite some time. However, in reality it is seldom seen in action except in the organizations with a very disciplined demand planning process. Even then the collaboration is a highly manual process relying on emails and Excel Documents.

Demand planning is one operation that works directly on achieving the company’s objective of selling, and is the first business function to flag any unexpected variations from the plan. Since all functions in a company should align with the company objectives, demand planning should ideally be the lynchpin holding all the spokes together.

One of the reasons why this outcome is not effectively achieved is the organizational positioning of demand planning function in the supply chain, whose main purpose is the production and deployment of product. Ideally demand planning should be a strategic function working to give an unbiased view of the expected unconstrained demand and then working as a team with the other value chain functions create a final one-number constrained plan.

Let’s us discuss about what role each function should play to make this a winning team.

Demand Planning initiates the process and should strive to make the baseline statistical forecast as robust as possible. They have the widest access to the statistical tools to create the most accurate estimate of demand by SKU. They should clean up the history through outlier correction and events marking and if required make the override changes to the numbers using known business knowledge.

Next, the Sales Team should provide unconstrained validation of the demand for their respective product(s), customer(s) and geography. The operative notion here is ‘unconstrained’ demand. Its been widely observed that during the collaborative process, the sales person has the tendency to give the biased view of the sales bases the available inventory. Its seldom made clear to them that what is expected is the ‘demand’ validation and not the expected ‘sales’ necessarily – hence the unconstrained view. The expected sales at times may be less than the demand due to inventory constraints. However, its important for the functions responsible for the supply to view the real demand specially for the forward periods. Constraining the demand for supply by the front end functions, distorts the real demand view for the supply planners.
Sales can do the overrides at any level in the hierarchy depending on the information they have on the demand.

sales
Planamind view of the planning and sales overrides

Following sales comes the Marketing for the unconstrained forecast. It should be noted that the constraints referred here implies the internal supply constrains only. Marketing team brings in a strategic level information such as promotions planned, category marker growth, competitor activity etc and can constrain the forecast if it is expected due to external factor such as competition activity. They are also to-go team to help bridge the gap between the plan and forecasts to help shape the demand favourably.

Marketing team shall usually offer information at the product or the geo level and not at the SKU level.

Beyond marketing, the collaboration moves to the functions that need to make the supply meet the demand. This means what we need from them is validation of the demand from the supply perspective and whether or not it will be met.

Supply Chain would ideally validate whether there are any logistics constraints on moving the stock that can impact the demand. This can be done at any level of the hierarchy but should ideally be expected at a product – geo level.

The Operations Team should provide inputs in terms of product availability. This is usually done for the periods beyond the current period and serves as a heads up to the sales team. This is valuable information to the sales team to plan for any substitutions and managing customer relationship. This should ideally be expected at a SKU level.

unnamed
Planamind view of the operations overrides

The Finance Team in an ideal scenario should flag any budget related constraints. This also means that they need to see the forecast in value terms. Beyond budgetary constraints the finance teams shall be interested to know if the forecasted revenue is in line with the planned revenue and more importantly what is causing the difference. They would also operate at the higher level of the hierarchy and the first one to see the numbers in totality at the highest level.

Finally, it’s of course the Management that gets the final word on the numbers. The changes here also usually are at the higher levels of the hierarchy. As a process these final changes would be out come of the executive S&OP meeting. As an input to this meeting, the planning should service the summary of all the changes done by the collaborators along with their reasoning, apart from the forecast summary, as part of the best practices.

unnamed (1)
Override Report From Planamind System

Its important to note that to be able to effectively collaborate with various stakeholders who would provide input at various levels of the data hierarchy, the planning system should be able to support this working and also handle disintegration of numbers from higher to lower level. In addition it should support the sequential process so that no two functions make changes at same time. Finally, it should make the system transparent and facilitate information exchange. So all collaborators should be able to view the inputs given by others along with the reasoning. We believe that this is the hallmark of a great consensus planning process. However, this is not possible to achieve using the excel sheets.

To summarize the collaboration process

Function


Consider supply constraints


Data hierarchy input level


Planning


No


Lower


Sales


No


Mid – Lower


Marketing


No


High – Mid


Supply Chain


Yes


Lower


Operations


Yes


Lower


Finance


Yes


High


Management


Yes


High


 

Collaborative forecasting if deployed effectively can make the system highly effective and contribute to the forecasting accuracy beyond the statistical models. In addition, as seen above, it makes the different functions work as single team to achieve a common objectives, but playing different roles – a team-sport!

Leave a Comment

Your email address will not be published. Required fields are marked *