What an Owner Should Stop Doing First

Why this matters

Owners rarely fail at delegation because they picked the wrong things to keep. They fail because they tried to hand off five things in one month, all five came back within three weeks, and they concluded that nobody around here can be trusted with anything. The list was not the problem. The order and the count were.

Sorting your work into "only I can do this" and "someone else could" is a different job, and a sibling article does it. This one starts after that sort, with a pile of things somebody else could do, and answers the only question that matters next: which single one goes first. Get the first one right and the second is easier, because you have proof the mechanism works and so does the person taking it. Get the first one wrong and you will not attempt a second for a year.

Stopping is a sequencing problem

There is no shortage of candidates. The constraint is that every handoff costs you time before it saves you any - the explaining, the checking, the two or three rounds where it comes back wrong and you fix it rather than argue. That cost is front-loaded and roughly fixed per item, which means running several at once stacks all the cost into the same weeks while none of the savings have arrived yet. That is the month owners quit.

So you run one at a time, and you choose the one whose ratio of recovered hours to handoff cost is best, subject to a hard gate on what a mistake would cost. Four factors do the work.

The four factors

Score each candidate on your list. Two are measured, two are judged.

Frequency, measured. How many times per month does this land on you? Count it from a log, not from memory. Memory is reliable on the monthly ordering run and badly wrong on the seven-minute question you answer thirty times, which is exactly the item that most needs finding.

Duration per instance in minutes, measured, including reset. The four-minute question that pulls you off a quote does not cost four minutes. Include the time to get back to what you were doing. If you have never measured your reset, add a few minutes per interruption and be honest that it is an estimate.

Multiply those two and you have monthly owner hours for the item. That is the recovery number.

Transferability, judged 1 to 5. How cheaply can somebody else be made reliably competent at this? A 5 is a rule you can write in a paragraph and show once - "moves inside 48 hours go to the next open slot in the same zone, anything further out comes to me." A 1 is judgment built over years that you cannot articulate on demand. Be strict here. Most owners score their own work as less transferable than it is, and the tell is that you cannot explain the rule because you have never had to write it down, not because there is no rule.

Error safety, judged 1 to 5. How cheap and how visible is a mistake? A 5 means a wrong call costs a small amount, shows up within days, and is fixable. A 1 means a wrong call is expensive, or does not surface for months, or reaches the customer as a safety issue. Note this is scored on the mistake, not on the task's importance. Ordering the wrong part is important and cheap. Underpricing a large replacement is important and expensive, and the difference is the whole point of the factor.

The eligibility gate

Both scores must be at least 3 out of 5. Transferability at or above 3 AND error safety at or above 3. An item failing either one is not eligible for a first-pass handoff regardless of how many hours it eats, because a low-transferability item burns your teaching time and a low-error-safety item is where a bad first handoff turns into a customer problem and permanently sours you on delegating.

AND, not OR. An item that is trivially teachable but where a mistake is expensive and invisible for a quarter - a filing deadline, a warranty registration - is exactly the trap the gate exists for. It looks easy, so it goes first, and the cost surfaces long after anyone connects it to the handoff.

Items that fail the gate are not permanent keeps. They are later, usually after a written procedure exists or after the person has a track record on something safer.

Choosing among the eligible

Rank the eligible items by monthly owner hours, highest first, and take the top one - with one tie-break, because the hours figure is not precise enough to separate close candidates:

If the top two eligible items are within 1.0 monthly owner hour of each other, take the one with the higher transferability score. A point of transferability is worth more than an hour a month, because it shows up as weeks of your time not spent correcting, and because a handoff that lands cleanly gets you to the second one faster.

One item per cycle. A cycle is four weeks, and you do not start the next item until the current one has gone two consecutive weeks with no instance bouncing back to you. Bouncing back means the question reached you anyway, in any form, including as an FYI you felt obliged to answer.

Eight items on one owner's list

A residential shop, five techs, one part-time office person. The owner ran a two-week interruption log, doubled the counts to a monthly rate, and scored what came out.

Across all eight items that is about 32.1 owner hours a month, which against a roughly 4.3-week month is about 7.4 owner hours a week. Two items fail the gate: final pricing on large quotes (transferability 2, error safety 1) and hiring interviews (transferability 1, error safety 2). Both are genuine keeps for now. The six eligible items account for 26.0 owner hours a month.

Now rank the eligible six by hours: quoting repeat maintenance at 5.8, the ordering run at 5.0, moving booked jobs at 4.8, chasing invoices at 3.6, the part decision at 3.5, discounts at 3.3.

Apply the tie-break. Top is 5.8, second is 5.0, and 5.8 minus 5.0 is 0.8 monthly owner hours, which is inside the 1.0-hour window. So the tie-break governs: compare transferability, 3 for quoting against 4 for the ordering run, and the ordering run wins. First item to stop: the weekly ordering run.

That is not the answer the owner expected, and the reason it is right is worth sitting with. Quoting repeat maintenance is the bigger number, but at transferability 3 it needs a real training arc - shadowing, a few reviewed drafts, corrections. The ordering run is four events a month, each a chunk of time, with a rule that fits on a card: stock list, reorder points, the standing supplier, anything not on the list comes to him. It transfers in one session and a checklist.

He handed it over. Weeks one and two, two items bounced back - a part not on the stock list and a substitution question. That is a bounce, so the two-consecutive-clean-weeks clock reset. Weeks three and four ran clean, so at the end of week four he started the next cycle, and by the ranking that is quoting repeat maintenance at 5.8 monthly hours, now with a person who has done one handoff successfully and an owner who believes it works.

Recovered after one cycle: 5.0 monthly owner hours, about 1.2 hours a week, out of the 26.0 monthly hours the six eligible items represent. Not transformative on its own. But it is the first of six, and the sixth will go faster than the first did.

When the ranking points at the wrong thing

If you have never successfully handed anything off, ignore the hours entirely for cycle one and take the highest-transferability eligible item. In the table above that is moving booked jobs, at transferability 5. The objective of a first cycle is not recovering hours, it is establishing that a handoff can hold. A first attempt that fails costs you far more than the hour-per-week difference between the best and second-best candidate.

If the item only exists because something upstream is broken, fix the cause instead. Chasing overdue invoices scores as eligible and recovers 3.6 monthly hours. If most of those chases trace to invoices that went out late or without a purchase order number, handing the chasing to somebody else staffs the symptom. The item to stop is not the chasing.

If you are covering a vacancy, suspend the whole exercise. You cannot hand work to a team that is already short. Note where you are in the sequence, and pick it up when you are staffed.

If the person taking it is already the bottleneck, moving your bottleneck onto theirs recovers nothing and costs you their goodwill. Check their load before you pick, not after.

Making sure it stays stopped

The failure that nobody watches for is the slow return. It does not come back as a handoff being reversed. It comes back as you being copied on things, then asked to confirm things, then asked first because confirming was faster.

Two mechanics catch it:

Count the touches, not the outcomes. For the first eight weeks after a handoff, keep a tally of every time that item reaches you in any form. Zero for two consecutive weeks is the standard for starting the next cycle, and a rising count in weeks five through eight - after it had been clean - means the rule has hit a case it does not cover. Fix the rule, do not resume the task.

Do not answer the easy ones. The single most reliable way to un-stop something is to answer a question you could have redirected, because it was quicker than redirecting it. It was quicker that once. It reinstalled you as the answer for everything after it.

References

• See related: The Jobs Only You Can Do, and the Ones You Just Won't Let Go
• See related: How to Delegate the Inbox
• See related: The Interruption Log SOP
• See related: Delegating the Decision, Not Just the Task
• Trade-standard practice for owner delegation sequencing in small service businesses