The Owner Who Kept Taking the Task Back
Why this matters
An owner who takes a task back is usually told they have a trust problem, and they usually accept that, because it feels true and it is unfalsifiable. Then they try harder at trusting, the task comes back again, and now they believe something is wrong with them.
Most of the time nothing is wrong with them. Something is wrong with the plumbing of the work, and the work is finding its way to their hands because that is still the path of least resistance. This is one shop's walkthrough, from the first symptom through three fixes that did not hold to the thing that actually did.
The signal
Six techs, an office manager, and an owner who still runs about two calls a day. The owner is doing estimate follow-up on a Sunday afternoon. He handed estimate follow-up to the office manager in March. It is now August.
This is the third time in about fourteen months that a task he formally delegated has ended up back on his own list without a conversation ever happening about it returning. No decision was made. It just migrated.
The right first move here is not introspection. It is measurement, because the story you tell yourself about why you took something back is written after the fact and is close to worthless as evidence.
Hypothesis one: she is not good at it
The comfortable explanation, and the one he half believed. He pulled the numbers instead of trusting the feeling.
Over the eleven weeks she had held it, the shop wrote 88 estimates and she followed up on 74 of them, so about 84% coverage. In the quarter before the handoff, when he held it, the shop wrote 96 estimates and he followed up on 61, so about 64% coverage. Her follow-ups converted at roughly 1 in 5. His had converted at roughly 1 in 4.
Read the wrong way, that last line is damning: she closes 20% of what she chases, he closed 25%. That is the read that would have justified taking it back, and it uses the wrong denominator. The denominator that matters is estimates written, not follow-ups attempted, because an estimate nobody called on is a loss the same as one that was called on and declined.
Run it per hundred estimates. At 84% coverage and 1 in 5, she converts about 17 of every 100 estimates written. At 64% coverage and 1 in 4, he converted about 16 of every 100. She was slightly ahead, on the number that pays the bills, while doing the job part-time alongside everything else she carries.
Hypothesis one is dead. Worth noticing that the conversion rate per follow-up is a real signal about technique, and it says she could be coached. It is not a reason to take the task back, and mistaking the two is how a shop loses a competent person's ownership of something they were quietly doing well.
Hypothesis two: she does not have time
Also plausible. Estimate follow-up is a real block of work, and the office manager already carried scheduling, invoicing, and the phones.
He timed it: the follow-up work ran a little under 3 hours in a typical week. He then looked at her week honestly and found roughly 4 to 5 hours of genuinely uncommitted time in it, mostly in the early afternoon lull.
The fix he applied anyway, because owners under pressure fix the hypothesis they can afford: he moved supplier reconciliation off her plate to a bookkeeper. It bought her about 2 hours a week. Six weeks later he was still doing estimate follow-up himself on weekends. Giving somebody more time does not stop a task from returning to you if the task never actually stayed with them in the first place.
Hypothesis three: the handoff was vague
Better instinct, and it produced a real improvement, which is exactly what made it a trap.
He rewrote the handoff: what a good follow-up sounds like, when to stop chasing, what discount she could offer without asking, when to escalate to him. Concrete, one page, genuinely useful.
Returns dropped for about five weeks. Then they crept back, and by week eight he was answering estimate questions daily again. A fix that works for five weeks and decays is telling you something specific: you have made the correct path easier, but you have not made the incorrect path harder, and the incorrect path is still the shortest route for whatever is actually driving the flow.
The cut that found it
He stopped asking why the task came back and asked a smaller question: where does an estimate conversation physically arrive first?
One week, one line per contact: every call, text, and email about an outstanding estimate, and which channel it landed on. No judgment, no action, just the channel.
Nineteen estimate-related contacts that week. Twelve arrived on his personal cell, four on the office line, three in the shared inbox. That is about 63% of estimate traffic landing directly on the owner's phone for a task the owner had delegated in March.
And of those twelve, he had personally handled nine. Not because he decided to. Because a customer calls, you are already talking to them, the question takes forty seconds, and forwarding it would take longer than answering it.
The cause was never trust, capacity, or specification. The estimate document the shop sends carries a contact block, and the contact block had his cell number on it, because he built the template years ago when he was the whole company. He delegated the task and left the front door pointing at himself. Every reply came to him, and once a reply is in your hand, handling it is always locally cheaper than routing it. Nine of twelve is not a discipline failure. It is what any reasonable person does, forty seconds at a time, until they are doing the whole job again.
The fix
Three parts, and the third is the one that made it hold:
- Move the interface. The estimate template's contact block now carries the office line and the shared inbox. Not an additional option, a replacement.
- Close the old door. His voicemail greeting names the office line for anything about a quote or a scheduled job. His cell stays open for techs and for existing jobs in progress.
- A forwarding rule with no exceptions for 30 days. Any estimate contact that still reaches him gets forwarded within the hour with no answer given, even when the answer would take twenty seconds. Especially then.
Part three is the hard one and it is not optional. For about a month you are deliberately being slower and less helpful than you could be, on purpose, to stop teaching customers and staff that you are the fast lane. Owners skip this part because it feels like poor service. It is a one-month cost against a task that has cost you three weekends a quarter for over a year.
He also told the office manager exactly what he was doing and why, which matters more than it sounds. From her side, the previous fourteen months looked like an owner who kept swooping in, and she had adjusted by not fully committing to anything he handed her. Naming the mechanism gave her a reason to commit that was not just another promise.
How he confirmed it
He re-ran the same one-week channel log four weeks later. Same method, same definitions, so the comparison is honest.
Twenty-two estimate-related contacts. Four on his cell, about 18% of that week's estimate traffic, down from about 63%. He forwarded all four inside the hour. His own time on estimate follow-up went from a little under 3 hours in a typical week to about 20 minutes, which is the weekly check-back and nothing else. That is a reduction of roughly 2.5 to 2.7 hours a week, on the order of 85 to 90% of the time the task had been costing him.
Two secondary confirmations he watched for, because a single log week can flatter you:
- Coverage held. Follow-up coverage stayed in the low-to-mid eighties as a share of estimates written. If routing had broken the work, coverage would have been the first thing to sag.
- The check-back stayed short. Fifteen to twenty minutes, once a week, and she brought numbers rather than questions. A check-back that starts growing again is the earliest sign the task is drifting back.
What would have pointed somewhere else
The channel log is the right first cut for a returning task, but the finding it produces is not always the interface.
If most contacts had arrived at the office and been escalated to him anyway, the cause is authority, not routing. The fix is a written ceiling on what she can decide without asking, plus a review of how he reacted the last time she decided something on her own. People escalate what they have been punished for deciding.
If her conversion per hundred estimates written had been materially below his - not a few points, but something like half - the skill hypothesis would have been live and the correct fix would be sitting in on a set of follow-up calls against a written standard, then coaching the gap. Taking the task back would still have been the wrong move, because it fixes this quarter and guarantees the same conversation next year.
If the returns had been genuinely varied, no repeating shape, no dominant channel, the task itself may be too broad to delegate as one unit. Split it: routine follow-up on standard work goes to the office, anything over a certain scope stays with the owner by design rather than by drift, and that is a decision rather than a failure.
References
- See related: The Owner Who Can't Let Go: A Decision Tree
- See related: How to Write a Handoff That Does Not Come Back
- See related: The Check-Back Cadence for Delegated Work
- See related: The Owner Bottleneck: When Everything Needs My Approval
- Trade-standard practice for estimate follow-up and customer communication routing