Once you have decided to implement planning you will have the choice of various approaches
- spreadsheets
- a black-boxed package
- an adaptable package
Differentiation between these product types is found in the degree to which they are specifically adapted to the planning process. There are several areas where planning specific functionality can be built in. In this article we are going to look at one of the most important ones, namely, the facility to do both bottom-up & top-down planning.
Bottom-up planning is what we normally do when we use spreadsheets. We create a plan at the lowest level of detail and then use formulae to aggregate the results. An example of this would be a standard Weekly Stock, Sales & Intake model. Most retailers run this sort of model at category or line level and add the results to give high-level totals.
Top-down planning is what we do when we impose a number at a consolidated level. For example we might look at a Total Product level sales plan and decide to impose a high level change because there is a gap between this number and the previously agreed strategic targets. This change would then cascaded down to all component members of Total Products. This is normally, but not exclusively, done on a pro-rata basis as shown in the example below:
| Bottom Up Sales Plan | Strategic Sales Plan | Top Down Sales Plan | Calculation: | |
|---|---|---|---|---|
| All Products | 5000 | 6000 | 6000 | Input |
| Product A: | 1000 | n/a | 1200 | (1000 x 6/5) |
| Product B: | 2500 | n/a | 3000 | (2500 x 6/5) |
| Product C: | 1500 | n/a | 1800 | (1500 x 6/5) |
This pro-rata approach assumes that data exists to provide a pattern, but what do we do if no data exists? There are a couple of alternative approaches that we can take. Firstly we may use the pattern from another member. For example if we wish to create a weekly sales budget for a new store we might first “seed” the store by copying data from a store that we expect to have similar sales pattern. Once we have done this we can impose a seasonal total and pro-rate it downwards.
| Store | Sales Wk 1 | Sales Wk 2 | Sales Wk3 | Sales Wk4 | Month 1 |
|---|---|---|---|---|---|
| Store 1 | 100 | 125 | 125 | 150 | 500 |
| Store 2 | 0 | 0 | 0 | 0 | 0 |
Now we seed Store 2 with Store 1’s sales pattern …
| Store | Sales Wk 1 | Sales Wk 2 | Sales Wk3 | Sales Wk4 | Month 1 |
|---|---|---|---|---|---|
| Store 1 | 100 | 125 | 125 | 150 | 500 |
| Store 2 | 100 | 125 | 125 | 150 | 500 |
and impose a new total …
| Store | Sales Wk 1 | Sales Wk 2 | Sales Wk3 | Sales Wk4 | Month 1 |
|---|---|---|---|---|---|
| Store 1 | 100 | 125 | 125 | 150 | 500 |
| Store 2 | 120 | 150 | 150 | 180 | 600 |
Alternatively, we may accept an average approach. Thus, inputting a sales forecast at the annual level of 52,000 would result in a weekly budget of 1,000 per week. With seasonal merchandise this is not likely to be acceptable, but there are cases where this approach is appropriate.
So, should you plan top-down or bottom up? The answer is that you probably need to do both. Many plans are built bottom-up by planners working in isolation. Approval of these plans requires aggregation followed by gap-analysis with strategic plans and subsequent top-down amendment.
Equally, a very common requirement in merchandise planning is the ability to plan at any point in the relevant hierarchies (which are normally Time, Product and Location).
We may, for example have a plan that is by Product Category by Week. However we may want to change a plan at the level of Product department by Month. Both Product department and month are mid level members of their respective hierarchies so this entails cascading the numbers to their individual components and then re-aggregating them to the totals for the plan.
So as we have seen the ability to create a top-down budget is one of the defining characteristics of an effective merchandise planning system. In future articles we shall go on to look at some of the other functionality that helps to differentiate them.
.jpg){kind=link}