A common reason layout decisions get made on instinct is the belief that modelling the site would be a project in itself — months of data work before you see anything useful. In practice, a workable warehouse twin needs less than people expect. Three ordinary inputs get you most of the way, and you almost certainly already have two of them.
The three inputs
A layout drawing, for geometry and scale. A racking or floor-plan drawing — a DXF, or even a PDF — carries the two things a model needs to be real: where things are, and how big they are. A drawing with its scale intact lets the model place racking at true dimensions instead of guessing. This is the input that turns a flat plan into a navigable 3D space.
A location list, for structure. An export of your bins or locations — typically row, bay, level and position, with a type for each — tells the model how the racking is divided and where the pick faces are. Most sites can produce this from their location master without much trouble.
Pick history, for the activity. A history of picks — item, location and time — is what lets the model run a representative day rather than show a still picture. It is also what weights everything that follows: the fast movers, the heat map, the routes that actually matter.
With geometry, structure and activity, you can build the model, see where the work concentrates, and compare alternatives. None of it requires a live WMS connection to begin.
You do not need it perfect
The other common blocker is the belief that the data has to be clean first. It does not. Bin masters are routinely a little messy — a mistyped location, an untyped shelf, a zone that moved and never got updated. A sensible onboarding process expects that, flags the gaps, and works around them rather than stalling. You learn more by building a rough-but-real model and correcting it than by waiting for a perfect dataset that never arrives.
If you have no usable location data at all — a paper site, or no WMS — you are not stuck either. You can sketch the layout directly and let the model derive the pick paths from the geometry. It is less precise than working from real history, but it is enough to compare options.
What you get back
The output of this is not a 200-page report. It is a navigable 3D model of your site and a short, plain answer to the question that prompted the exercise: what is the layout costing you in walking, and what would the main alternatives do about it. A one-page summary — current versus optimised, the free re-slot win separated from the capex re-rack win — is usually enough to move a decision that had been stuck.
Keep the figures honest
Any pounds-per-year figure a twin produces is an indicative, worked example: calibrated to a real dimension on the drawing, tied to your own pick volumes, and clearly labelled as illustrative rather than a quote. The value is in comparing options on a consistent basis, not in a single headline number. A model that is modest about precision but honest about direction is the one worth bringing to a budget conversation.
See the shape of it
The fastest way to understand what a twin gives you is to try one. WalkBill builds this kind of model in the browser on a synthetic demonstration site — run the live demo to see the 3D, the day-run and the compare. When you want the real answer for your own warehouse, traced from your own drawing, that is what the Options Pack is for.