TAWhy it paysREF / THE CASE AND THE SOURCES

Reduce time to competence: why 3D equipment training pays

Why it pays.

Time to competence is the time from a new starter’s first day to being signed off to work a role unsupervised. Gaps in equipment know-how lengthen it, and show up as downtime, incidents and slow restarts. Here is what published research says, and a way to put your own numbers on it.

What the research says.

Published figures from industry bodies, researchers, analysts and two equipment makers, Siemens and ABB, not our own results. Each note below gives the source and its limits.

44%

of 183 major accidents in US and European process plants were linked to maintenance.

Okoh and Haugen, 2014 · note 1

36%

of the most serious (Tier 1) upstream process safety events put down to human performance involved control of work and isolation.

IOGP data, IChemE Hazards 32, 2022 · note 2

81 min

on average to restart a large plant after an unplanned stop, up from 49 minutes five years earlier. A skills and knowledge gap is one of three causes named.

Siemens, 2024 · note 3

$125,000

an hour is close to the typical cost of an unplanned outage, as reported by 3,215 plant maintenance decision-makers.

ABB Value of Reliability, 2023 · note 4

2.4 to 1

energy workers nearing retirement for every worker under 25, across advanced economies.

IEA, 2025 · note 5

73%

of 100 maintenance leaders surveyed find new technicians hard to recruit, yet only 55% have a formal system for sharing know-how between technicians.

McKinsey, 2023 · note 6

76%

higher learning gain for university students taught one lesson with a desktop 3D simulation rather than a lecture, still held 40 days later.

Bonde et al., Nature Biotechnology, 2014 · note 7

9 vs 18

months to operator sign-off: the target of an LNG terminal’s simulator-based programme, against traditional peer training.

ARC Advisory Group, 2017 · note 8

Put your own numbers on it.

Two places the value shows up on one machine: people ready sooner, and downtime that better know-how prevents. Change any number.

Currency People trained on this machine a year Weeks to get a new person ready on it today Cost of a week in training, per person (pay, overheads, cover) How much sooner they are ready with 3D training on the model25% Unplanned downtime on this machine, hours a year Cost of an hour of that downtime Share of that downtime better know-how could prevent10%

Value of training time saved0

Downtime avoided0

Estimated value a year on one machine, before costs0

An illustrative estimate from your own numbers, not a quote or a promise, and before the cost of the build. The starting values are examples: the hourly cost starts close to the ABB median above, and the 25% is half the cut in time to sign-off that one LNG terminal set as its target (note 8). None of the research above measures how much downtime better know-how prevents, so the 10% is a placeholder for your own judgement. Change any of them. On your first machine we measure the real figures with you.

How we prove it with you.

Industry figures make the case. Your own figures close it.

Baseline first

Before we start, we measure how long it takes today to get someone ready on the machine, and how your people score today on the assessment the explorer will use.

One machine, fixed scope

The first build is one machine to a scope agreed up front, so the result shows quickly and the cost is known.

Your people judge

We agree the measures with you, and your own people judge the result. We claim nothing we have not measured with you.

Get started Try the explorer live

Sources

Okoh, P. and Haugen, S., Process Safety and Environmental Protection (2014): 80 of 183 major accidents in the handling, processing and storage of hydrocarbons and hazardous chemicals in the US and Europe, 2000 to 2011, were linked to maintenance. Linked to maintenance is broad: it includes a lack of barrier maintenance and deficient planning, not only mistakes by maintainers. Read the source
Analysis of IOGP member Tier 1 process safety events, 2014 to 2021, presented at IChemE Hazards 32 (2022). The 36% applies to the 223 events classed as operational human performance, not to all events. Upstream oil and gas only. Read the source
Siemens, The True Cost of Downtime 2024: survey of large industrial firms in automotive, FMCG, heavy industry, and oil and gas, April 2019 to March 2023. Siemens gives three causes for the rise: lost skilled labour and a skills and knowledge gap, supply-chain delays for spares, and harder faults. A vendor study. Read the source
ABB, Value of Reliability survey (2023), run by Sapio Research across energy, oil and gas, utilities, chemicals and other industries. A median of self-reported costs, not audited accounts. ABB sells maintenance services. Read the source
International Energy Agency, World Energy Employment 2025. The IEA also projects that two in three new hires to 2035 will be needed just to replace retirees. Advanced economies, whole energy sector. In emerging and developing economies the ratio is about 1 to 1. Read the source
McKinsey and Company (2023), survey of 100 senior leaders responsible for equipment maintenance in oil and gas, energy and three other asset-intensive sectors. A small, self-reported sample, not specific to the GCC or the UK. Read the source
Bonde, M. T., Makransky, G. et al., Nature Biotechnology (2014): 91 university students, one lesson, a desktop 3D laboratory simulation compared with a lecture and group exercise. Students rather than plant staff, and the authors declared competing interests. Read the source
ARC Advisory Group (2017), reporting an LNG terminal operator’s training programme built around an operator training simulator. A programme design target at a single site, not an independently measured outcome. Read the source

Last updated 24 September 2026

Questions about the case.

How to read the figures, and how the result is measured on your own plant.

Are these figures Upskill Energy’s own results?

No. Every figure on this page is published research from industry bodies, researchers, analysts and two equipment makers, not our own results. Each one has a note giving the source, what it measured and its limits. We will publish our own figures only once a pilot has been measured with a client.

Where do the figures come from?

They come from peer-reviewed papers, an IChemE Hazards conference paper on IOGP data, the International Energy Agency, a McKinsey survey, ARC Advisory Group, and surveys by Siemens and ABB. Each figure links to its source in the notes at the foot of the page, with the year and the caveat that travels with it.

How does the value calculator work?

The calculator puts your own numbers on two effects for one machine: people ready sooner, valued at the cost of each week before they are productive, and downtime avoided, valued at your cost of an hour of lost production. Every input can be changed, and nothing you enter is stored or sent anywhere.

How would we measure the benefit on our own plant?

Before the build we measure how long it takes today to get someone ready on the machine and how your people score on the assessment the explorer will use. The first build is one machine to a fixed scope, and afterwards your own people judge the result against that baseline, using measures agreed in advance.