FORECASTING
Inventory forecasting: the right stock, in the right place
Inventory forecasting with TIMi estimates the stock level needed at each store or warehouse for a future period, to avoid both stock-outs and overstock, and shows the factors behind it.
Stock per location
A recommended level for each reference and each storage point.
Lead times included
Supplier lead times and logistics constraints are inputs.
Explained levels
Each recommendation shows why the safety stock is what it is.
HOW IT WORKS
How inventory forecasting works.
Three steps, in one engine, on a standard server with no cluster.
1. Prepare the data
Anatella gathers the history of sales or stock outflows, replenishment lead times, current stock levels and logistics constraints (storage capacity, delivery frequency), then structures them into variables.
2. Build the model
Modeler projects future stock consumption from historical trends and seasonal factors. The recommended stock level is then derived, given replenishment lead times.
3. Recommend and explain
Each recommendation comes with its factors, such as the consumption trend, the supplier lead time or a safety margin linked to demand variability.
Anatella
ModelerHOW IT DIFFERS
From a demand forecast to a stock level.
Demand is only the first half of the problem.
Demand, then lead times
Demand forecasting estimates a future consumption volume. Inventory forecasting goes further and turns it into a recommended stock level, given replenishment lead times and logistics constraints.
Lead time sets the safety stock
Replenishment lead time is an input of the model: the longer it is, the higher the recommended safety stock for the same accepted risk of stock-out.
Handle constraints explicitly
When constraints matter (storage capacity, delivery frequency), Anatella includes an LP/IP solver for large-scale optimization problems.
EXPLAINABILITY
A level says how much to hold. The factors say why.
A stock recommendation alone gives a level to maintain, not what justifies it. A white-box model shows the factors, so logistics teams can tell whether a high safety stock comes from demand variability, a long supplier lead time or a particular seasonality.
For logistics and purchasing
A recommendation can be questioned and adjusted with business context, instead of being followed without justification.
For holding costs
Understanding the factors of a safety stock helps to adjust it precisely, instead of applying a uniform margin out of caution.
For model owners
Modeler’s built-in Drift Report shows when a model loses accuracy, so you update it when needed instead of on a fixed calendar. Recompute recommendations at regular intervals (weekly or monthly, depending on rotation speed), or as soon as a significant change in demand or lead times is detected.
DATA
What data a stock model uses.
Anatella connects to more than 80 sources, so this data can be gathered without prior custom development.
Sales and stock outflows
By reference and by storage point.
Lead times
By supplier or by reference.
Stock and capacity
Current levels and storage constraints.
Seasonality and events
Promotions, periods of high demand.
PROOF
Closest reference.
We do not yet have a published customer story on stock-level optimization itself. Bavaria is the closest: its goal was to avoid stock-outs and simplify the work of store managers.
Demand models for 300,000 points of sale
The models cross sales, weather and neighboring stores, and score 2.4 million records per second.
FAQ
Common questions.
How is it different from demand forecasting?
Demand forecasting estimates a future consumption volume. Inventory forecasting goes further and turns it into a recommended stock level, given replenishment lead times and logistics constraints.
Can it handle slow-moving references?
Yes, but the precision depends on the amount of history available. For references that sell very little, a model can rely on comparable references to build a first recommendation.
How are supplier lead times taken into account?
The replenishment lead time is an input variable of the model: the longer it is, the higher the recommended safety stock for the same accepted risk of stock-out.
How often should stock recommendations be recomputed?
Ideally at regular intervals (weekly or monthly, depending on rotation speed), or as soon as a significant change in demand or in supplier lead times is detected.
SEE FOR YOURSELF
Test it on your own data.
Run TIMi on your own data with a free trial, or talk to us about your case.
