The Tasks Only the Owner Can Actually Do

Why this matters

There is a version of this question that is about willingness, and there is a version that is about structure. This card is the structural one. Some tasks sit with you because a license is in your name, a bank has your signature on file, or a surety underwrote you personally, and no amount of trust in your lead changes that. Others sit with you because the thing that would let them move has never been built. Both feel identical from inside a busy week, which is why owners spend years defending a list where most items were never actually locked. The test here is not whether you would be comfortable letting go. It is: what is the lock, and what is the instrument that opens it.

Name the lock or it is not owner-only

The rule, stated plainly: a task counted as owner-only must have a lock you can name in one sentence (unit: per task, on your current list). If you cannot name the lock, the task is not owner-only. It is undocumented, or it is unbounded, and both of those are things you build your way out of.

There are four locks. Two are real and partially movable with a specific instrument. Two look permanent and dissolve almost entirely once you name them.

The legal lock is real and it has a timeline, not a workaround

Where your trade requires a licensed qualifying individual and that individual is you, the work tied to that qualification does not move until another employee holds the qualification, and that is an examination, experience, and application timeline measured in months to years, not a documentation task. Treat it as a hiring and development plan with a start date, not as an item on a delegation list.

The same shape applies to anything you signed personally. Where a surety, a lender, or a landlord took a personal guarantee, decisions that touch that instrument stay with you for the life of the instrument regardless of who runs the shop day to day, and the only real path is renegotiating or retiring the guarantee.

What does move, and what owners rarely separate out, is the preparation. Assembling the renewal package, tracking the continuing-education hours, gathering the documents, chasing the deadline: none of that requires the license holder. Only the signature does. A legal lock almost always sits on the last five minutes of a task whose other three hours are fully transferable.

The fiduciary lock moves with a band, not with trust

Payroll approval, bank transfers above a threshold, insurance changes. These are held by an institution that has your authorization on file, and they move by adding an authorized person plus a written band that says what they may approve.

Set the band against your own history rather than against your comfort. A workable payroll example: the lead approves any payroll run falling within a stated percentage of the trailing four-run average, and the owner approves anything outside that band (unit: per payroll run, gated on the trailing four-run average, recomputed every run). Pick the percentage by looking at your last dozen runs and finding the variance you have never once questioned. A band drawn tighter than your actual behaviour just recreates the phone call.

The band has to be written where the institution and the person can both see it, and it needs a review date. An approval band with no review date either goes stale as the shop grows or quietly gets ignored, and the second failure is worse because it looks like it is working.

The information lock is the biggest pile and the cheapest to open

This is where most of a stuck owner's list actually lives, and it is the only lock that fully disappears. The task requires you because the pricing history, the customer's odd access arrangement, the reason you stopped using that supplier, and the four exceptions to your own standard exist only in your head.

The instrument is writing, and the useful form is short. One to three pages per task: the trigger, the decisions and their bands, the exceptions with the reason for each, and the two or three things that have gone wrong before. Not a manual. A manual is what owners produce when they are avoiding the handoff.

The diagnostic that proves an item belongs here: ask what you would have to say out loud to a competent person to let them do it once. If the answer is a conversation, the lock is information. If the answer is "I would have to sign it", the lock is legal. Those two answers point at completely different work and owners routinely give the first answer while behaving as though it were the second.

The trust lock is real but it decays on a schedule you set

Some accounts genuinely bought you. The customer calls and asks for you by name, the referral partner sends work because of a fifteen-year relationship, the lender extended terms on your handshake. That is real value and dumping it on a green lead is a way to lose it.

It moves through a sequence, not an announcement. Three contacts is a reasonable shape: on the first, you are present and the other person leads a portion of it; on the second, they lead and you are present; on the third, they go alone and you make one call afterward to the customer. What you are transferring is not competence, it is permission, and permission transfers by being seen to be given.

Not every trust-locked relationship should move. Keep the two or three where losing the relationship would genuinely damage the shop and where your personal involvement is what the other party is actually paying for. That is a deliberate choice with a cost you accept, not a lock.

A worked sort

An owner writes down every task he believes is owner-only. Fourteen items. He assigns a lock to each.

• Legal, 2 items. The annual license renewal signature and signing the bond paperwork.
• Fiduciary, 1 item. Approving the payroll run.
• Information, 6 items. Pricing exceptions on repeat commercial accounts, the callback judgment call, warranty goodwill decisions, supplier selection for anything non-standard, the after-hours dispatch call, and quoting anything involving access constraints.
• Trust, 5 items. Five accounts that ask for him by name.

Now apply the instruments.

The 6 information-locked items are written up over three weeks, between one and three pages each. All 6 move. This is not optimism: the lock is definitionally the absence of a document, so producing the document ends it. What takes time is not the writing, it is the two or three cycles of the person running it and coming back with the case the document did not cover.

The payroll approval moves to a band as described above, so the routine run leaves and the exception run stays. Count it as moved for the weekly load and retained for the exception.

Of the 5 trust-locked accounts, he runs the introduction sequence on 3. Two he keeps deliberately: the general contractor who sends roughly a third of the shop's commercial work, and the property manager whose portfolio he has personally walked. Those two are not a lock he failed to open, they are a choice with a stated reason.

The 2 legal items stay. Their preparation work, which was most of the actual hours, leaves.

The result: 14 down to 4 items genuinely held, the 2 legal signatures plus the 2 trust accounts kept on purpose. Ten of fourteen moved, about 71% of the original list. The weekly hour count fell further than the item count did, because the two legal items were mostly preparation and the payroll item retained only its exceptions.

The decay test on your own list

Re-run the sort every six months. The point is not the total, it is the composition, and one line in it carries the signal.

The information-lock count should fall every cycle. It is the only lock whose fix is entirely inside your control and entirely within one working week. If your information-locked count is the same at two consecutive reviews, twelve months apart, the honest reading is that you are not writing things down, you are re-deciding them, and you will still be re-deciding them next year.

The legal count should be stable and should only fall when someone qualifies. The fiduciary count should fall once, when you set the band, and then stay flat. The trust count should fall as sequences complete and then stop at the number you chose to keep.

A list that grows is a specific signal too, and usually an honest one: growth adds genuinely new commitments, especially fiduciary ones. What should not grow is the information pile, because a growing information pile means new context is accumulating in your head faster than you are writing it down, which is the same condition you started with wearing a newer set of tasks.

Where the sort goes wrong

The most common error is assigning the trust lock to something that carries the information lock. "This customer only wants to deal with me" is sometimes true and is more often "this customer has an unusual arrangement that nobody else knows about, so when someone else picks up the phone the answer is wrong, and now the customer asks for me." The test that separates them: send someone else once, with the context written down first. If the customer is fine, it was information. If the customer asks for you specifically after a competent interaction, it was trust.

The second error is treating the whole task as locked when only the signature is. That one is worth checking on every legal and fiduciary item on your list, because it is where the hours hide.

References

• See related: The Jobs Only You Can Do and the Ones You Just Won't Let Go
• See related: How to Hand Off a Recurring Task Permanently
• See related: The Owner Absence Readiness SOP
• SBA guidance on delegation, authority limits, and management capacity
• Trade-standard practice for licensing and qualifying-individual requirements, which are set by state and local authorities