The goal is to find patterns within the customer orders and then tune warehouse processes
to be a more efficient match for these patterns. Customer orders are available to us from
a sales history. This is extremely valuable information and companies are understandably
reluctant to share it.
We must be able to guarantee protection. This can include asking specifically that No
financial information be shared — we do not need it since our concerns are operational
— and that product descriptions can be omitted if necessary.
Plan for data transfer
Client will prepare two files:
sales.txt, which lists all customer orders picked on during the recent past,
preferably the last year.
skus.txt, which describes the stock-keeping units shipped from this warehouse.
The files will be prepared in two stages: First, skus.txt together with one
week’s worth of sales.txt, which we will validate and then return a summary to
the client. (One week’s data is small enough to be easily handled. For example, it will be
viewable by a spreadsheet program, while the full data set may not be.)
Then the client prepares the full files (all SKUs and all lines). We will
then validate the full files and return an exception report.
Structure of data files
Format: ASCII text
Field delimiter: a single Tab character
Record delimiter: end-of-line
Which convention? Cr (unix)? CrLf (Microsoft)? Lf (Apple)?
No missing (empty) fields unless meaning agreed upon in advance
Date format: YYYYMMDD
Records of lines.txt sorted by date and, within date, by order
number so the lines of each order are consecutive.
Description of customer orders
Unique order number
Unique SKU identifier
Description of SKUs
Unique SKU identifier
Is this case sensitive?
Does it encode any special information?
Have there been any recent changes?
SKU description (text, for reality checks)
Current storage location
Pieces per case
Piece length, width, height, weight
Case length, width, height, weight
Stack height of cases, pallets
Active? New product date?
Description of Warehouse
This data may be unstructured and typically requires discussion with warehouse personnel.
Addressing scheme (how are addresses assigned to storage locations?)
Map giving layout of warehouse. (This does not have to be of AutoCAD level of detail;
it can be a drawing in a spreadsheet.)
List of types of storage and their dimensions (example: shelving, with openings
18 inches deep, 36 inches wide, and 16 inches high; 5 shelves per bay.)
Explanation of key warehouse processes (receive, put-away, pick, check/pack/ship).