Key takeaways
See our guide to CMMS implementation, where these roles usually land.
Every maintenance manager would agree, in principle, that cross-training matters. The team is too dependent on the senior technician who knows the packaging line. The night shift is short on anyone who can handle a PLC issue. The mechanic who retired last month took unwritten knowledge with them. Everyone nods.
The practical question, who should train who, on what, by when, usually does not get answered. The reason is structural: cross-training competes with daily work, and daily work always wins in the moment. Without an explicit structure that allocates time and tracks coverage, the program stays a good intention.
The fix is unromantic. A skill matrix, a coverage rule, a quarterly review. Same shape as every other improvement program that actually executes. The article on manufacturing KPIs covers the team-capability metrics that the matrix feeds.
A table with technicians on one axis and asset classes on the other. Every cell holds a proficiency level for that technician on that asset class. The matrix becomes the single source of truth for "who can do what", replacing the implicit knowledge in the maintenance manager's head, which fragments the moment they leave the role.
Four levels is the right granularity. Three is too coarse to track meaningful progression; five or more is too granular and the boundary judgements become subjective.
This is the part most plants skip. A matrix without a coverage target is just a record-keeping exercise. The rule:
Every critical asset class must have at least two technicians at level 3+ available on every shift.
"Two" is the minimum because one is single-point-of-failure. "Level 3+" is the minimum because level 2 cannot diagnose. "Every shift" is the constraint that catches the night-shift coverage gap that breaks at 2am.
Calculating coverage against this rule produces a clear list of gaps: asset class X has only one level-3 on Tuesday nights; asset class Y has zero level-3s on weekends. Those gaps become the training plan for the next 6-9 months. The piece on work order management systems covers how to surface coverage gaps as a live operational metric rather than a quarterly audit.
Each gap gets a named expert (level 4) and a named trainee. The pair works on the asset together for a defined period, typically six weeks, with the trainee shadowing for week 1, performing under supervision in weeks 2-4, and independently in weeks 5-6 with the expert available for questions.
The six-week cycle is the right length because it covers a typical PM cycle on most asset classes plus one or two real failure events. Compressing to two or three weeks usually misses the failure exposure that turns level 2 into level 3.
Each pair produces a one-page "lessons from this asset" document that goes into the failure mode catalogue. The next trainee on the same asset benefits from the prior pair's documentation, which is what makes the program scalable beyond the first round. The piece on root cause analysis covers the documentation conventions that make these lesson docs reusable.
At the end of the six weeks, the trainee performs one diagnostic and one repair on the asset with a different expert (not the trainer) observing. The audit is binary: the trainee performed at level 3, or the training needs to extend. Without the audit, the matrix gets updated optimistically and the coverage number drifts away from reality.
Training everyone on everything. The temptation is to ship a "cross-training program" that touches all asset classes for all technicians. The result is broad exposure, no depth, and the matrix shows lots of level-1s and almost no level-3s. When a real failure happens, the level-1s cannot handle it, and the program looks like it produced no useful capability.
The coverage rule fixes this. Training resources go to the gaps the rule identifies, not to every cell. Most plants find that hitting the rule across all critical asset classes produces roughly two to three dozen training pairs over a year, a manageable number that produces measurable capability rather than diffuse exposure.
Once the initial coverage is achieved, the program shifts to quarterly maintenance. New technicians join the team, others leave, asset classes get added or retired. Each quarter the matrix is reviewed against the coverage rule, gaps are identified, and the next training pairs are formed.
This quarterly cadence is what prevents the program from being a one-time project that decays. Plants without the quarterly review usually find themselves back at the original "two indispensable people" pattern within a couple of years. The piece on the preventive maintenance schedule covers the parallel ongoing-maintenance pattern for technical programs.
The skill matrix works in any tool, a spreadsheet is enough for plants under 30 technicians.
Where a unified OEE + CMMS platform helps is in two places: the coverage gap can be live-calculated against the work-order assignment data (who actually worked on what last quarter), and the lessons documents from each training pair can attach to the failure mode catalogue rather than living in a separate folder.
Fabrico is built for that workflow. To see what the coverage picture looks like against your live work-order history, book a demo .
Senior technicians are often protective of their expert status. The way most plants resolve this is to make level 4 (expert) the role that gets visible recognition, paired with training responsibilities. The expert is not losing status by training someone; they are gaining a formal recognition. Whether you formalise it through compensation is plant-specific.
The coverage rule is the same; the proficiency-level definitions get stricter. Level 3 on a safety-critical asset usually requires explicit certification rather than peer audit, and the audit cadence is more frequent.
Two numbers: percentage of critical asset classes meeting the coverage rule on all shifts (target: 100%), and rolling-30-day failure-event resolution time on asset classes that newly hit the rule. The second number is the practical confirmation; if resolution time on those assets does not improve, the level-3s were not actually trained to level 3.
That is the single highest-risk gap a plant can have. The fix is usually structural, bringing in external training, rotating the asset's PMs to expose other technicians, or in extreme cases acknowledging that the asset's continued reliability depends on retaining one specific person. None of those is comfortable; all three are better than discovering the gap when the technician calls in sick.
Building the matrix without the coverage rule. The matrix without the rule is a record; with the rule it is an action driver. Plants that build the matrix and then "review it quarterly" without explicit coverage targets usually let the program quietly stall.
Login
Blog
