Most mine designs do not fail because of bad engineering. They fail because of bad data.
Wrong file. Wrong version. Wrong folder. It sounds trivial, and it is where an enormous amount of rework, frustration and lost time actually comes from. Good design can be undone by bad data, and autosave quietly makes it worse — saving into the wrong file is data loss you may not notice for weeks.
You do not need perfect systems. You need control.
A folder structure that survives contact with a site
Keep it simple and keep it consistent. Four folders will carry most operations:
- Survey — raw inputs
- Design — working files
- As-built — what actually exists
- Archive — old and superseded
They are separate on purpose. The moment design files and as-built files share a folder, someone will design in the as-built.
Names should tell a story
A good file name carries its own hierarchy: site, then region or area, then pit or location, then the design, RL or shot. Something like WAN_DEPB_Pit4_580RL, WAN_Pit3_Ramp_v02, or PitA_Shot12_Block3.
Use one separator consistently. And avoid the FINAL trap — the moment you have FINAL and FINAL_v2 in the same folder, the naming has stopped working.
Version control without a system
You do not need software for this. Use v01, v02, v03, or use dates in the form 2026_05_01. What matters is that you never overwrite without tracking. Saving over a working file loses history; saving over an as-built creates a much bigger problem.
Keep a design register: file and version, date, author, and what changed. If it is not tracked, it did not happen.
Working data is not issued data
Working data is editable, in progress, and expected to change. Issued and as-built data is locked, approved, and for reference only. Protect it accordingly — read-only folders, controlled access, backups that actually run.
Never design in an as-built folder.
The mistakes everyone has seen
Working in the wrong folder. Saving over issued data. Autosaving into the wrong file. Using old topography. Duplicate files everywhere, and no record of what changed.
None of them look like much on the day. All of them are expensive by the end of the quarter.
Structure drives outcomes. How much of your design history could you reconstruct if you had to?
The carousel
This piece began as a LinkedIn carousel. The full deck is below — use the arrows, or click any slide to enlarge.
