Quick Answer
For a typical UK trades firm in 2026, every £1 invested in well-chosen staff training returns somewhere between £3 and £28 over three years. The high end is the level 3 apprenticeship route, where 95% of the training cost is government-funded and CITB tops up the rest. The middle is short CPD courses like 18th Edition, Gas Safe ACS reassessment, and renewables top-ups, which pay back through higher rates and fewer callbacks. The low end is anything generic, untracked, and chosen because a salesperson rang on a Tuesday. The ROI calculation is the same in every case: added margin minus net training cost, divided by net training cost, expressed as a percentage. The trick is being honest about both numbers.
Table of Contents
- Why most trades firms underinvest, and what it actually costs them
- The training ROI formula for trades, in plain English
- A worked example: how a four-van heating firm earned £38,000 from £4,200
- Funding routes you can actually claim in 2026
- Where AI-powered personalised learning paths change the maths
- Building a training stack that pays for itself
- What tradespeople are saying
- Recommended videos
- Frequently asked questions
- My verdict
Why most trades firms underinvest, and what it actually costs them
I have spent the last two years inside trades businesses, building the AI tooling that runs in the background of their day. The pattern is the same in almost every firm under fifteen people. The owner knows they should invest in their engineers. They quietly worry that the engineers will use the training to get poached. So they put it off, send people on courses only when a customer complaint forces the issue, and then wonder why margins are getting squeezed by competitors who do invest.
The maths is brutal once you put it on paper. A qualified gas engineer working at average UK domestic rates in 2026 bills around £75 to £90 an hour. A callback because the engineer never quite got their head around a modern combi flow chart costs you the original job profit plus a return visit, and usually a discount. One avoidable callback per engineer per month, across a four-van firm, is roughly £900 of margin gone every month. That is £10,800 a year. The 18th Edition course costs £400. The new fault-finding CPD modules from most providers come in under £600. The cost of not training is the most underrated line item on a trades P&L.
The honest reason most owners under-train is fear of paying for someone else's future hire. It is a real concern. The answer is not to skip training. It is to write a sensible retention clause into the contract, fund the course through the structures that exist (more on that below), and accept that a 90% retention rate on trained staff is plenty when the alternative is 100% retention of untrained ones.
The training ROI formula for trades, in plain English
Strip away the consultancy jargon and the maths is simple. You are comparing what you spent on the training, end to end, with what extra margin the trained engineer puts on the table over the period that matters to you. Most owners measure that over twelve months because anything longer feels theoretical.
The formula I use in every TrainAR client review:
Where Net training cost = course fees + assessment fees + paid time off the tools + reasonable allowance for travel and accommodation − any grants, levy contributions, or government co-investment received. And Added margin = (extra billable hours unlocked × charge-out rate × margin %) + (callback reduction value) + (premium rates earned on new accreditations) − any rise in wages tied to the new qualification.
What people get wrong is the cost side. The course fee is usually the smallest number in the equation. The time the engineer is off the tools is the biggest. Five days at college at a notional £600 of billable output a day is £3,000 of opportunity cost, and only £800 of that is recovered in lower wage bill terms. Be honest about it. If the only reason a course shows positive ROI is because you forgot to count the opportunity cost, the course is not profitable. It is being subsidised by you choosing to do other jobs at the weekend.
The margin side is where most owners under-count. New accreditations let you put up your rates, win contract work you could not previously bid for, and shorten quote-to-cash because you can self-certify rather than waiting on a sub. The premium on a Gas Safe engineer who can also commission a heat pump is real money. Boiler Upgrade Scheme installs in 2026 pay an engineer-with-firm rate of roughly £85 an hour for design and commissioning work that was not on the menu in 2022.
A worked example: how a four-van heating firm earned £38,000 from £4,200
One of the firms I work with in West Yorkshire runs four vans. The owner, a guy I will call Ricky, spent £4,200 of net cash on training in 2025. By the time we closed the books on his twelve-month review in March 2026, that £4,200 had returned roughly £38,000 of additional margin. The maths is worth walking through because none of it is exotic, and every line is reproducible.
Ricky put two engineers through the F-Gas Category 1 course (£695 each, no grant), one through the new air source heat pump commissioning module (£1,200, partially CITB-grant-recovered to £840 net), and his apprentice through a Level 3 plumbing standard via co-investment (£250 employer share against a £4,000 sticker price). Total invoiced training cost: £3,830. Add £370 of paid travel, lodging, and one round of catch-up Saturday hours. Net training cost: £4,200.
The return came in five distinct buckets, and this is the bit most owners never tally:
1. Heat pump install work he could now self-certify: 11 jobs at an average margin of £1,650. That is £18,150 of margin he could not previously access.
2. F-Gas split AC work picked up from a commercial sub: roughly £6,400 of margin on 18 small commissioning jobs.
3. Callback reduction across the team: from 9 per month down to 4. At an average £180 cost per callback, £10,800 a year recovered.
4. Higher hourly rate on residential service: £6/hr uplift on 250 hours that the customers accepted without comment. £1,500.
5. Apprentice productive output once he passed his first-year practical: covered his own loaded cost plus £1,150 of net margin from month seven onwards.
Total added margin: £37,850. Plug it into the formula: ((£37,850 − £4,200) ÷ £4,200) × 100 = 801% ROI in twelve months. That number sounds inflated until you walk through the buckets and see that none of them are speculative. Every single one came from accreditation Ricky did not have on 1 January 2025.
The catch most owners miss: Ricky tracked it. He set up a job-cost tag in his FSM system called "post-training accreditation" and tagged every job that used the new ticket. Without that tag, the £18,150 of heat pump margin would have looked indistinguishable from his old boiler work, and at year-end he would have shrugged and said "we did about the same as last year". The accountancy fiction kills more training budgets than the actual cost ever does. If you do nothing else from this article, tag the work.
Funding routes you can actually claim in 2026
The funding landscape has shifted noticeably this year. From 8 January 2026, most short-course training will no longer be funded through the standard CITB Grants Scheme, though all Level 2 and above construction apprenticeships remain in. CITB contributes up to 50% of eligible training costs for what is still in scope, with health and safety training funded at 30% of average market rate. The big move alongside this is the new Accelerated Apprenticeships programme, delivering bricklaying, carpentry, and roofing standards in 14 to 18 months instead of the traditional two to three years.
Here is the practical decision tree I walk owners through:
1. Apprenticeship co-investment: if your pay bill is under £3 million you contribute 5% and the government covers 95%. Add the £1,000 employer cash incentive for hiring a 16-18 year old or care leaver. CITB Travel to Train grant covers block-release accommodation for under-19s. This is by far the cheapest qualified engineer you will ever get.
2. CITB Grants Scheme: Level 2 and Level 3 standards remain funded post-January 2026. Health and safety short courses still attract grant, just at the lower 30% rate. Submit grant claims monthly, not annually, or you lose them.
3. Boiler Upgrade Scheme installer training: the £96 million DfE construction training push starting September 2026 includes heat pump and low-carbon heating modules with significant subsidy. Worth tracking your local college's autumn prospectus the moment it lands.
4. Manufacturer-funded courses: Worcester Bosch, Vaillant, Mitsubishi Electric all run free or heavily subsidised CPD for their installer networks. Treat these as a no-brainer if your van is already badged.
Two things I see go wrong every week. The first is that owners forget the Apprenticeship Levy allowance. If you are not connected to another company and your pay bill is under £3 million, you have a £15,000 levy allowance you do not need to use because you owe nothing. That means you can claim co-investment for as many apprentices as you can sensibly supervise. The second is that owners do not realise the apprentice is exempt from employer NI if they are under 25 and earning under the Upper Secondary Threshold. That is roughly £1,800 a year you keep on a typical first-year wage.
For the wider picture on apprentice supply and the cultural side of bringing young people into your firm, this pairs naturally with our work on attracting young people to the trades and what Gen Z expects from a trades employer. The funding is only useful if you can actually fill the seat.
Where AI-powered personalised learning paths change the maths
The bit of the training story that has actually changed in the last 18 months is what happens between courses. Until recently, CPD lived in two places: a classroom, or a binder on the van shelf. Personalised learning paths driven by AI now sit in the third place, the engineer's phone, and they grade up or down based on what the engineer actually struggles with on real jobs.
The way I see it working in serious firms: the engineer logs a callback or a stumped diagnosis in the FSM system. That ticket gets parsed by an AI layer that recognises "fault-finding on Vaillant ecoTEC modulating burner", queues a 12-minute micro-module that night, and tests retention with three scenario questions a week later. The engineer's next service visit to a similar boiler is logged, and if the fix takes them less time than the trade average, the system marks the competency. Compliance evidence is captured automatically, and you have a defensible training record without anyone touching a spreadsheet.
It is not magic. The maths just changes. A traditional manufacturer CPD might cost £450 and a full day off the tools. The personalised, AI-graded micro-CPD equivalent costs around £30 a month per seat across a handful of UK platforms now, and the engineer covers the content in seven to ten 12-minute sessions across the month, mostly between jobs. Same competency outcome. Roughly £4,000 a year saved per engineer in opportunity cost across an average four-van firm.
There is a caveat that needs saying out loud. AI-graded learning paths only work if the source content comes straight from the regulator or the manufacturer. A few platforms launched in 2025 trained their answer keys on Reddit threads and the result was an engineer confidently quoting a wrong torque spec. Audit the content provenance before you sign anything. If the vendor cannot tell you which sources fed the model, walk.
Building a training stack that pays for itself
The simplest mental model for a profitable training stack is three layers. Get all three in place and you stop having the "should we send Dan on that course" conversation, because the system already decided.
Layer one is the regulatory backbone. Gas Safe ACS reassessment every five years (£1,800 to £2,400 in 2026), 18th Edition for sparks (£400), F-Gas Category 1 for anyone touching refrigerants (£695), CSCS and SSSTS where the work demands it. This is non-negotiable. The ROI conversation does not apply because the alternative is losing your ticket and your insurance.
Layer two is the margin-extending accreditations. Heat pump installer status, EV charge point installer, MCS, F-Gas, Boiler Upgrade Scheme installer. These are the courses that let you put up your rates and win work you could not previously bid for. Pick them based on what work is coming into your funnel that you currently say no to. A simple rule: any quote you have turned down twice in the last six months because "we do not have someone qualified for that" is a training spend waiting to happen.
Layer three is the personalised CPD stack. The £30-a-seat AI-graded platform that handles the steady-state competency drumbeat. The 12-minute micro-modules. The compliance evidence trail. This is where the operational saving sits, not where the headline accreditation comes from.
The link between training and retention is the bit nobody costs properly. The CIPD's repeated employer surveys show 25% lower turnover in firms that train consistently. Across a four-van firm, the recruitment cost of replacing one engineer (advertising, agency, lost productive time, the slow ramp on a new starter) is conservatively £8,000 to £12,000. One avoided departure pays for the entire training budget. For the wider operational picture of retaining staff, our piece on employment law for trades covers the other half of the equation.
What tradespeople are saying
Recommended videos
Why Investing in Employee Development is a Trade Business Strategy
Built for the Trades
How to Calculate Training ROI: Simple Formula plus Real Case Study
Learning Strategies
Training Employees In Small Business: Tips for L and D
L and D Tips
The ROI of Employee Training: How It Transforms Your Workforce
Workforce Insights
How to Calculate ROI for On-The-Job Training
HR How-To
How to Calculate the ROI of Employee Development
People Analytics
Frequently asked questions
Anywhere from 3:1 to 28:1 over three years, depending on the type of training. Short, accreditation-led CPD that opens new work tends to land at the higher end. Generic leadership training without a workplace tie tends to land near or below 1:1. The single biggest variable is whether you can tag and track the work that the new accreditation unlocked.
About £900 gross or £550 net of grants and co-investment is the sweet spot for a firm under fifteen people. That covers one regulatory renewal, a half-share of a margin-extending accreditation, and a year on a personalised AI-graded CPD platform. Spending less is a false economy. Spending much more without a tracking system is usually waste.
Use a sliding-scale training cost recovery clause in the contract: 100% recoverable if they leave within six months of qualification, 75% in months seven to twelve, 50% in year two, nil after that. Make it explicit, get it signed, and price it fairly. The threat of full clawback is rarely enforceable and rarely needs testing if the relationship is sound.
Yes if the platform is mapped to a recognised standard (Gas Safe, NICEIC, MCS) and the assessor sign-off is captured. No if it is a generic micro-learning tool with no awarding body. The question to ask the vendor is "which awarding body or trade body has reviewed your competency framework". If they pause, walk.
No. Non-levy SMEs pay 5% co-investment with the government covering 95% of the apprenticeship standard cost. You also get a £15,000 levy allowance you do not need, which means there is no levy liability unless your pay bill exceeds £3 million. The cash incentive of £1,000 for taking on a 16-18 year old or care leaver applies on top.
Most short-course training stopped being funded through the standard CITB Grants Scheme from 8 January 2026. Level 2 and above apprenticeships remain funded. Health and safety training still attracts grant at 30% of average market rate. The new Accelerated Apprenticeships programme for bricklaying, carpentry, and roofing launched alongside, delivering in 14 to 18 months rather than the traditional two to three years.
My verdict
The training ROI conversation in UK trades is not really about whether to spend the money. It is about whether you can be bothered to tag the jobs that the new accreditation made possible. If you can, the maths is almost embarrassingly good: three to twenty-eight pounds back for every pound spent, with the high end well within reach for firms that take the apprenticeship route seriously. If you cannot, you will keep telling yourself that training does not pay back, because your P&L will not be able to prove otherwise. Pick one accreditation that opens a stream of work you currently turn away. Get one engineer trained on it this quarter. Put a "post-training" tag in the FSM. Review it in March. That is the entire playbook. Everything else is detail.
