The model comes before the data
Plant managers often hear about live forklift tracks, material positions and spatial dashboards first. Those layers only work if the 3D geometry underneath is already clean and usable. That preparation is the part that rarely gets airtime, yet it sets the pace and the cost of everything that follows.
Twinzo applies a 100 MB limit on 3D models uploaded to the platform. The ceiling is deliberate: the application is built to run on cheaper Android devices as well. Larger models can be accepted, but they raise the hardware requirements for every viewer. Automatic processing covers .fbx and .obj files. Other formats—CAD, IFC, glTF—need manual or enterprise handling.
.fbx remains the preferred format because it moves through the pipeline with the least friction.
BIM and scans rarely arrive ready
BIM models can be brought in, but they almost always need work. Polygon reduction, floor division and object-count reduction are the usual steps. Without that cleanup the model becomes too heavy or too cluttered for smooth use inside the operational twin. Point clouds cannot be ingested as-is either; they must be meshed first. Photogrammetry or lidar scans can supply the raw material, yet the conversion and simplification still sit on the project timeline.
A basic layout with overall dimensions is enough to start. Higher fidelity often brings in 360-degree camera video or scanning methods. Every extra detail adds reduction and validation effort before the model is accepted.
Why the calendar stretches to months
Deployment typically runs from one to three months. A large share of that window is spent getting the spatial base right before any live data is overlaid. The work is iterative: import, reduce, check alignment against the physical plant, adjust again. It is not glamorous, but it is the foundation.
If the geometry is inaccurate or overweight, the later positions of forklifts, people and materials will not sit correctly in the 3D view. Spaghetti diagrams, heat maps and zone analysis lose reliability. The operational twin is not a photorealistic simulation or a design BIM; it is a spatial context layer. The model only needs to be good enough for that purpose, yet reaching “good enough” still takes focused effort.
Geometry first, live data second
Once the 3D model is prepared and accepted, the platform can ingest data from RTLS, MES, ERP, WMS and SCADA systems. Twinzo functions as an umbrella layer that adds spatial context; it does not replace those systems. The sequence is deliberate. Live data on top of a messy or oversized model produces limited value. Live data on a clean, correctly scaled base produces usable operational insight.
The real cost of an operational digital twin therefore begins with the geometry work—time, specialist attention and repeated checks until the model sits under the size limit and matches the plant well enough for day-to-day use. Only after that step does the live layer become worth adding.