Welder Training by the Numbers
Chapter 10 of 10 · 10 min read
Training economics and the training record
Compare training a welder with hiring one over five years, allow for trainees who leave, keep the training record that proves what was done, and review what the fabricator's first year achieved.
The earlier chapters followed the fabricator’s four trainees from their first arc to their qualification tests, putting numbers on how skill grows, what practice costs, how assessment and qualification work and how skill fades. This last chapter asks the question the owner asked at the start: was it worth it? It compares training a welder with hiring one over five years, allows for the chance that a trained welder leaves, sets out the training record that turns a year of training into evidence, and reviews what the first year achieved.
Training against hiring
Chapter 5 costed a competent trained welder’s first year at about A$32,400, including an allowance for trainees who leave or do not reach the target, and compared it with the cost of hiring a qualified welder in a tight market: about A$8,000 in recruitment, a wage premium of about A$9,000 a year, and about A$3,000 of supervision and testing to confirm the new welder’s skills, about A$20,000 in the first year and A$9,000 in each year after. These are illustrations, excluding GST, and every business’s figures will differ, but the structure of the comparison is general.
In the first year, training costs about A$12,400 more than hiring. In each later year, the trained welder costs about A$9,000 less, because there is no premium to pay. If the trained welder stays, the extra first-year cost is recovered in a little under a year and a half after training ends.
Allowing for time and for leaving
Money spent now is worth more than money saved later, and a trained welder may leave. A simple net present value calculation handles both.
- Year 1
- −A$12,400
- Years 2 to 5, if still employed
- +A$9,000 a year
- Chance of staying each year
- 85%
- expected savings: A$7,650, A$6,500, A$5,530, A$4,700 in years 2 to 5
- discounted at 7% and added to the year 1 cost
- NPV ≈ +A$8,000
Training pays on these numbers unless the business loses more than about three in ten of its trained welders each year, and it pays well if it keeps them. The calculation also shows the value of retention: moving the chance of staying from 70% to 85% a year is worth about A$6,600 per welder over five years, which is a useful budget for the things that keep people, such as fair pay as their skill grows, a path to a trade qualification, and interesting work.
The value of capacity
The comparison with hiring assumes a welder can be hired. When none can, as the fabricator found for a year, the real alternative is subcontracting or turning work away. This is an illustration. If a welder enables about 1,500 welding hours a year that the business would otherwise subcontract at about A$20 an hour more than its own cost, the trained welder is worth about A$30,000 a year in margin kept, far more than the hiring premium. In that market, training pays back within the first year after qualification, and the question is not whether to train but how many.
The output during training
Trainees are not idle during training. In their supervised production module (Chapter 4) and in the months after qualification, they produce useful work, more slowly than an experienced welder at first. This is an illustration. If a newly qualified welder produces about 70% of an experienced welder’s output in their second year and 90% in their third, the business gets about three-quarters of a full welder’s output over those two years, on top of the savings above. The fabricator’s records of hours and output by welder, the same records that feed the learning curves, show how close its trainees came to that.
Which numbers matter most
A cost model is most useful when it shows which inputs matter. This is an illustration. Raising the practice hours a trainee needs by a quarter, from 400 to 500, adds about 100 hours at about A$66 to A$77, roughly A$7,000, to the first-year cost, and cuts the five-year NPV at 85% retention from about A$8,000 to about A$1,400. Halving the hiring premium, to A$4,500 a year, which also makes the first year of hiring A$4,500 cheaper, turns the NPV to about −A$6,000, so that hiring would be cheaper, if welders could be hired at all. Changing the chance of staying from 85% to 70% a year cuts the NPV to about A$1,400. And adding the value of capacity in a market with no welders to hire swamps all of these. The lesson for the manager is that the two things the business controls best, the efficiency of its training, through instruction and feedback, and the retention of the people it trains, are also two of the three inputs that matter most.
Improving the plan for the next intake
The first year’s records showed where the plan could improve. The trainees’ learning curves were steepest when the instructor was present most, which confirmed the protection of instructor time in busy periods. Introducing the vertical position early had worked. The module tests had predicted qualification results well, except for Ben, whose test was booked against the evidence. The theory sessions tied to the week’s practice had been remembered; the two sessions held in a block, when the instructor was away, had not. For the next intake, the manager kept the structure, booked qualification tests only on the evidence of mock tests, added a second trained instructor, and staggered the start dates so that the trainees would not all reach the vertical module together.
Costs beyond money
Some costs and benefits of training do not fit a spreadsheet. Experienced welders who instruct and mentor often find the work rewarding, and some become better welders for having to explain what they do. A business known for training good welders attracts applicants, including people who would not have applied to a business that only hires. And a shop with a training culture tends to be a safer and more careful one, because the habits taught to trainees are seen and copied by everyone. On the other side, training takes management attention, and a business that starts training and then neglects it can end with discouraged trainees and wasted money. The manager counted none of these in the numbers but mentioned all of them in the report.
The instructor as the key asset
If one person decides whether training succeeds, it is the instructor. The cost model treats instruction as a cost per hour, but the learning curves showed that it is the largest single influence on how many hours a trainee needs, and the agreement exercise showed that the instructor’s judgement decides whether assessment is fair. The fabricator now treats its lead instructor’s role as a position in its own right, with time allocated, training in instruction and assessment, and a say in the plan, rather than as a duty added to a busy welder’s job when trainees arrive.
The training record
The training record turns a year of practice into evidence. The fabricator keeps, for each trainee:
- Avoid: Training plan and targetmodules, hours and the definition of ready
- Avoid: Coupon log and learning curveshours, coupons, rejects and reasons, by position
- Avoid: Module tests and sign-offsdate, assessor and evidence
- Avoid: Safety induction and supervision stageset-up check, incidents and near misses
- Avoid: Qualification test recordstest pieces, witness, laboratory report, range
- Do: ? Welder register entriespositions used, refreshers, continuity and renewals
The record serves several readers. The trainee sees their progress and what remains. The instructor sees patterns across trainees and can improve the plan. The manager sees cost and progress against the budget. The quality system and customers see how the business’s welders were trained and qualified. And the business keeps the knowledge of how it trains, which would otherwise leave with the instructor.
What the first year achieved
This is an illustration. At the end of the year, the manager reported to the owner.
The cost was slightly over budget, mainly because of Ben’s retest and the extra practice before it. The three qualified trainees were repairing about 4.2% of their weld length in their first production months, above the experienced welders’ 2.5% but below the 6% the manager had feared (Chapter 1), and falling. The two retiring welders’ work could now be covered. And the business had a training plan, a cost model, assessment criteria, a welder register and records that it could use for the next intake. Ana, still on plan, was forecast by her learning curve to be ready for her qualification test about four months into her second year.
What a second intake costs
Much of the first year’s cost was a one-off. The plan, the cost model, the assessment criteria and reference coupons, the record sheets, the booth equipment and the instructor’s training were all built once and can be reused. This is an illustration. The manager estimated that about A$9,000 of the first year’s A$118,600 was set-up cost of this kind, and that the second intake of four trainees would cost about A$4,000 less per trainee than the first, because the instructor would be more efficient, the plan would not be relearned, and fewer coupons would be wasted on criteria still being argued over. Training gets cheaper as the business gets better at it, which is the learning curve of Chapter 3 applied to the business itself.
Telling the owner
The manager’s report to the owner was two pages: what the training had cost against the budget, what it had produced in qualified welders and production hours, the repair rates, the safety record, the five-year comparison with hiring, and the plan for the next intake with its budget. Each number came from the records, and each was labelled as measured or estimated. The owner approved the second intake at the same meeting, which the manager attributed less to the numbers themselves than to the fact that they existed.
Campbell’s standard method
Campbell’s argument in 1919 was that welding schools needed a systematic, standard method of instruction, because classes that each taught in their own way produced welders of uneven quality. The fabricator’s year confirmed his point inside a single business: four trainees, trained to one plan, assessed by agreed criteria, measured by the same records, reached production welding at different rates but to the same standard. The numbers in this book, learning curves, practice costs, sample sizes, pass probabilities, decay rates and payback, are tools for running such a method and for showing that it works.
A failure: no evidence
This is an illustration. Before the structured training began, a major customer had audited the fabricator’s quality system and asked how its welders were trained and how the business knew they were competent beyond their qualification certificates. The manager could show the certificates but nothing else: no training plans, no assessment records, no evidence of refresher practice. The customer accepted the certificates but noted the gap, and the fabricator lost points in a later tender against a competitor that could show a training system.
At the next audit, the manager showed the training plan, the coupon logs and learning curves, the module records, the agreement exercise between assessors and the welder register. The auditor’s report described the training as a strength. The record had cost little to keep, because it came from the work the instructors were doing anyway.
Bringing it together
Each chapter of this book applied one or two relationships to one part of welder training. Repair rate times weld length times repair cost gave the value of skill. A power law described how speed and quality improve with practice. People’s time dominated the cost of practice. The binomial distribution and the rule of three said what a sample of coupons can prove. Every piece must pass, so the chance of passing a test is the product of the pieces’ chances. Skill decays roughly exponentially and recovers faster than it was learned. And discounted cash flows, weighted by retention, decided whether training pays. Together they let a manager plan, run and defend a training programme with numbers rather than hope.
Exercise: a shorter stay
The owner worries that trained welders will leave after their second year. Using the cash flows in this chapter, calculate the five-year NPV of training rather than hiring if the trained welder stays for years 2 and 3 for certain and then leaves, at a 7% discount rate. What does the result suggest?
The cash flows are −A$12,400 in year 1 and +A$9,000 in years 2 and 3, with nothing after. Discounted at 7%: −12,400 ÷ 1.07 ≈ −11,589; 9,000 ÷ 1.07² ≈ 7,861; 9,000 ÷ 1.07³ ≈ 7,347. The NPV is about −11,589 + 7,861 + 7,347 ≈ +A$3,620.
Even if the welder leaves after two productive years, training still beats hiring on these illustrative figures, because the first year’s extra cost is recovered within those two years. The result suggests that the owner’s worry, while understandable, is not a reason to stop training; it is a reason to work on retention, which raises the return considerably, and to keep the training record, so that the next trainee can be trained as efficiently as the last.
