How to Decide Whether an Extra Credential Earns Its Keep

Why this matters

Credentials accumulate. Somebody suggests one at a supplier counter, a course provider sends a compelling email, a customer asks whether you hold something and you decide never to be asked again. Three years later the shop is carrying six certifications, two of which have never gated a single job, and every one of them costs continuing education hours and renewal administration forever.

The cost of an extra credential is not the course. It is the permanent line it adds to the register and the recurring hours it takes out of a year that was already full. A credential you never use costs you exactly as much as one you use constantly, and it is invisible because it never generates a complaint.

Step 1: Count the demand from your own records

Before anything else, go into your job history for the trailing 12 months and count the jobs you declined, referred out, or subcontracted because you did not hold this credential.

Count jobs, not enquiries, and count only the ones where this specific credential was the reason. A job you turned down because you were booked does not count. A job you referred out because the customer wanted something outside your trade does not count. If your records do not distinguish declined-for-credential from declined-for-anything-else, that is your first finding, and the fix is a decline reason field before it is a credential decision.

Never substitute a provider's market-size claim, an association's estimate, or a supplier's opinion for this count. Those numbers describe a region. Your service area, your customer mix and your reputation are the only market that will actually call you.

Step 2: Name what it gates, and what it does not

Write down, in one line each, what work this credential legally or contractually opens that you cannot do today, and what it does not open despite common belief.

The second line is the one that saves money. Credentials are routinely oversold as opening categories they only partly touch: one that permits you to perform a category of work may still leave the permit, the design responsibility or the sign-off with someone else, and a manufacturer certification that puts you in a warranty network may govern only labour reimbursement rather than access to the work. Read what the issuing body says it confers, not what the course brochure implies.

Step 3: Gate 1, demand

At least 6 jobs in the trailing 12 months, from your own records, that this credential would have unlocked.

The unit of analysis is jobs in a trailing 12-month window. Six is roughly one every other month, which is the floor at which a stream is a pattern rather than a run of coincidences. Below that you are looking at anecdotes, and anecdotes reliably feel more frequent than they are because a declined job is memorable.

Adjust the floor upward, not downward, if your total job volume is high: at a shop running a large book, six declines is noise. The floor is a minimum, never a target to argue your way down to.

Step 4: Gate 2, carry burden

Annual carrying hours must be at or under 10 percent of the delivery hours the credential is expected to unlock in a year.

Carrying hours are everything the credential costs you every year whether you use it or not: continuing education seat time specific to it, renewal application and evidence assembly, any additional insurance or bond administration it drags in, and the register maintenance. Delivery hours are the field hours the unlocked work is expected to consume in a year, taken from the demand count times a realistic hours-per-job figure for that work.

Be precise about what this ratio is and is not. It is a burden measure: how much unbilled administration each hour of unlocked work drags behind it. It is not a return. Do not divide unlocked billable hours by acquisition hours and call the result a payback multiple, because those are two different currencies: acquisition and carrying hours are unbilled time you absorb, while delivery hours are sold work that still costs technician time, travel and materials to deliver. What the shop keeps is the margin on those delivery hours, not the hours.

Ten percent is a starting default to tune to your own overhead tolerance. What it is really testing is whether the credential is load-bearing or decorative. A credential carrying a heavy annual obligation against a thin stream of work is a subscription to a service line you barely run.

Step 5: Gate 3, holder durability

The credential must sit on the owner, or on a person you would confidently expect to still be here at the next renewal.

An individual credential leaves with the individual. If the only holder is inside their first year, or in a role with high turnover, you are not buying a capability, you are renting one, and the day they leave the work stops mid-stream on jobs you have already sold.

The gate passes three ways: the holder is the owner or a long-tenured person, or a second person holds it as well, or you accept the single-holder risk explicitly and in writing, having priced what happens to in-flight work if it walks. That third option is legitimate. Accepting a risk you have named is a decision; not noticing it is not. See related: What Happens to Work in Progress When a Credential Lapses.

Step 6: Apply all three as AND, then price acquisition separately

All three gates must pass. The Boolean is AND, not two of three. Each gate kills the decision for a different reason and none of them substitutes for the others: demand without an affordable carry is a stream you will resent, carry without demand is a subscription, and either without a durable holder is a plan with a single point of failure.

Only after all three pass do you price acquisition: study hours, seat time, exam, application administration. Acquisition is a one-time entry cost, so the question it raises is not payback, it is persistence. Will the demand that cleared gate 1 still be there in three years? A requirement change, a manufacturer exiting your market, or a large competitor entering it can all remove the stream while the carrying obligation stays exactly where it was.

Worked example: one credential, two verdicts

A shop is considering a certification. Its owner would hold it.

Gate 1. The job history shows 9 jobs in the trailing 12 months declined specifically because the shop did not hold it. Nine is at or above the floor of 6, so the demand gate passes.

Gate 2. Carrying hours: 8 hours a year of continuing education tied specifically to this certification, plus 3 hours of renewal application and evidence assembly. Total carrying: 11 hours a year. Delivery hours: those 9 jobs a year at about 5 field hours each is 45 delivery hours a year. The burden ratio is 11 divided by 45, or about 24 percent, against a ceiling of 10 percent. The carry gate fails.

Gate 3. The owner holds it and is not leaving. Passes.

Verdict: 2 of the 3 gates pass, and the rule is AND, so the answer is no. That is not a close call being resolved conservatively. Twenty-four percent burden means that for every four hours of unlocked field work the shop delivers, it absorbs about one hour of unbilled administration purely to keep the certification alive. The nine declined jobs are real, and the right response to them is to keep referring them out or subcontract them, which costs nothing to maintain in a year when they do not appear.

Now the flip. Working through step 2 more carefully surfaces something the first pass missed: the certification also gates the annual maintenance agreements on a category of equipment the shop already services, which it has been handling under a workaround that a customer has started questioning. Recount the delivery hours with that included and the annual figure is about 120 delivery hours, not 45.

Re-run gate 2: 11 divided by 120 is about 9 percent, which is at or under the 10 percent ceiling. The carry gate now passes. Gates 1 and 3 were already passing, so 3 of 3 pass and the answer becomes yes.

Read what actually changed, because it was not the credential and it was not the arithmetic. The carrying hours never moved; they were 11 in both versions. What moved was the shop's own understanding of what the credential gates, which is step 2, and which the first pass ran too quickly. A decision that flips entirely on a step people treat as a formality is a decision worth slowing down on.

The failure mode in the other direction. A shop that skips gate 2 and buys on gate 1 alone ends up with the first verdict's economics and never notices, because a 24 percent burden does not announce itself. It shows up as continuing education hours that always feel heavier than they should, an annual renewal season that runs long, and an owner who cannot say why. Nothing breaks. The shop is just permanently carrying an obligation that does not pay its way, and because it never fails, it never gets reviewed.

What changes the answer

A customer contractually requires it. Then the gates do not apply in the same way, because the alternative is not "refer the work out," it is "lose the account." Run the decision on what that account contributes instead, and check whether the requirement follows the customer or only that contract. See related: The Credential That Opens Work You Do Not Want.

A credential is a prerequisite for one you actually want. Its own gates will fail, correctly, because it gates nothing directly. Evaluate the pair as one decision with combined acquisition and carrying hours against the end credential's demand.

Someone on the team already holds it personally. Acquisition drops to near zero and gate 3 needs a harder look, not a softer one, because the whole capability now rests on one person who did not obtain it for you and can take it with them.

The credential removes an exclusion on your insurance rather than opening new work. Then it is a risk control and this method is the wrong instrument entirely. Price it against exposure, not against unlocked hours.

How to verify you got this right

Re-run gate 1 and gate 2 twelve months after the credential lands, with real numbers rather than projections. Count the jobs it actually unlocked and the hours it actually delivered, and recompute the burden ratio against the same 10 percent ceiling.

A credential that passed on projection and fails on actuals is not a mistake to hide. It is a candidate for retirement, and letting a credential lapse deliberately, with the affected work referred out and a note in the register saying why, is a legitimate decision that almost no shop ever makes. Add the recheck to the annual credential audit so it happens by default rather than when somebody remembers. See related: The Annual Credential Audit SOP.

References

• Trade-standard practice in scoping what a certification confers, which is defined by the issuing body's own documentation rather than by course marketing
• See related: The Annual Credential Audit SOP, The Credential That Opens Work You Do Not Want, Continuing Education That Pays Off, What Happens to Work in Progress When a Credential Lapses