The compliance gap between a demo and a deployment
The distance from “it works on stage” to “it is legal on your floor”.
Payback in months, not slides: reading a vendor ROI claim
How the maths is built, and which line usually goes missing.
Every payback slide is a model wearing the clothes of a fact, and a model is only as honest as the assumptions loaded into it.
Every ROI claim is a model
A vendor’s payback-period claim compresses labour cost, uptime, service cost and install time into a single number. Change any one input and the number moves, sometimes by months — which is why the same category of machine can carry a short payback in one deck and a much longer one in an operator’s own recalculation.
None of that makes the claim dishonest. It makes it a model, and a model is worth reading for its assumptions before it is worth reading for its headline number.
“A payback slide with no assumptions listed is not a calculation — it is a marketing number wearing a spreadsheet’s clothes.”
The gap between a vendor’s model and an operator’s outcome usually sits in one predictable place.
The line that usually goes missing
Integration time, spare-parts logistics and the labour needed to supervise a machine through its first months rarely appear in the same slide as the payback number. A worked example published by an industry body — dated, sourced, and built from real deployment data rather than a single vendor’s projection — will usually show a longer runway to payback than a sales deck does, precisely because it counts the lines a sales deck leaves out.
None of that is a reason to distrust every ROI claim. It is a reason to ask which costs were modelled and which were assumed away, before comparing one number to another.
WHAT TO TAKE AWAY
Ask which inputs built the payback number — labour rate, uptime, shift length — before trusting the headline.
Integration time and spare-parts logistics are the lines a sales deck most often leaves out.
A dated, sourced worked example from an industry body is worth more than an unpublished vendor projection.
Rebuild the calculation with your own site’s numbers before comparing it to anyone else’s.
The maths only becomes useful once it is rebuilt around a specific site, not a generic one.
Rebuilding the maths on your job
A vendor’s model is usually built on average shift lengths, average labour costs and an assumed uptime that may not match a specific floor. Rebuilding the same calculation with the buyer’s own numbers — their shift pattern, their labour rate, their site’s realistic uptime — is the only version of the payback period worth acting on.
The rebuilt number is rarely as clean as the one on the slide, and that is exactly the point: a payback period that survives contact with a real shift is worth more than one that only survives a pitch.
EDITORIAL RULE
Sourced or silent. Every figure on this page carries an attributed, dated source. Where we could not source a number, we left it out rather than estimate it.
SOURCES
- 01IFR World Robotics 2025Oct 2025
- 02Grand View Research 2025Market sizing & methodology
- 03Manufacturer price listsJan–Mar 2026
- 04Operator interviewsMilan, London, Abu Dhabi · Q1 2026
A QUESTION
One person answers — from the first email to the room.
EMAIL
WECHAT
nh_alliance
LINKEDIN
