Decide which elements of your supply chain to include:

Sites (Distribution Centers, Warehouses) – Which facilities will be modeled?

Customers – Will you model every customer or a representative subset? Modeling all customers is most accurate; modeling a sample is faster for initial testing.

SKUs (Products) – Modeling all SKUs is ideal often necessary, but starting with high-velocity or strategically important SKUs is a reasonable first step to begin modeling.

Suppliers – Include all suppliers that serve your network, or at least those serving your modeled sites and SKUs.

Choose a historical time period that will form your baseline. Consider:

Length – A longer period (12-24 months) captures more variability; a shorter period (6-12 months) is easier to manage initially.

Recency – Choose a period recent enough to reflect current operations, but old enough that you have reliable historical data.

Seasonality – If possible, select a period that includes normal seasonal patterns for your business.

Write down your scope decisions. You'll refer back to these as you gather data. For example: "Baseline model includes 4 DCs, 150 customers, 250 SKUs, covering Jan 1 – Mar 31, 2024." This clarity prevents scope creep and keeps your model manageable.