How to Stop Being the Bottleneck in Your Business

The short answer: Count every decision that reaches you for ten working days and sort each as strategic, tactical or operational. Most of the volume is small calls you answer the same way every time. Give each of those a named owner and a written limit, keep the strategic few, and repeat the count until no sale or invoice waits on you.

Ask most owners what eats their time and they describe the hard things: the angry client, the big quote, the employee they're worried about. Ask them to count for two weeks and the list comes back different. The hard things are in there, but they're a small slice of the volume. The bulk is a long tail of small questions that take a minute or two each and never feel like a problem individually.

This post is about measuring that tail, and about deciding which of those decisions were ever yours to make. How to write the limits once you've found them is a separate job, and I cover it in the post on delegating decisions.

Why do owners stay the bottleneck after the team grows?

Because the team grew around an owner who still holds every limit. Each new person adds questions, and the answers still live in one head.

Two beliefs keep that in place, and both sound reasonable from the inside.

1. "It's faster if I just answer." For one question, it is. Across sixty questions in ten days, with people waiting on each one, it's the slowest arrangement the business has.
2. "They'd get it wrong without me." Sometimes they would, at first. A written limit caps what a wrong call can cost, and the waiting has a cost of its own.

How do you sort decisions by level?

Before you count anything it helps to have a way to sort what you collect. I use three levels.

A strategic decision changes what the business is or commits it for a long stretch: your pricing structure, which customers you take, hiring a manager, signing a lease, taking on debt. These are expensive to reverse, they come up a few times a year, and they belong to you.

A tactical decision bends the plan for one customer, one job or one month, inside the strategy you've already set: a discount, a priority call between two projects, which vendor to use on this particular order. Money is bounded, the damage of a wrong call is bounded, and the person closest to the customer usually has the best information.

An operational decision is just running the plan: swapping two crews on a Thursday, reordering stock, rescheduling a visit. Whoever is doing the work should make these, almost always without asking.

When a decision is hard to place, I ask three questions. What does it cost if it's wrong, and how awkward is it to undo? How often does it come up? Would the answer become a precedent that other people will cite next month? The third question is the one owners forget, and it's usually why a decision that looks small feels like it needs them.

Should the owner decide a $5,000 discount?

Here's a decision that looks strategic and mostly isn't. A client on a $46,000 job asks for $5,000 off because the project ran three weeks late, and the delay was partly the supplier's fault.

Five thousand dollars is real money for a company this size, so the instinct is to send it to the owner. But look at what kind of decision it is. It's one client, one project, and a question about margin on that project. That's tactical. What makes it feel strategic is the precedent risk: if discounts get handed out casually, clients talk, and the next one asks for ten percent. So I'd call it tactical with precedent risk, and the answer to "who owns it" is whoever owns the client relationship and the project margin, usually an account lead or project manager, up to a limit.

The limit is where the decision turns into a pricing question, and you have to pick a number. A reasonable starting rule might be: the account lead can give up to 5% of the contract value, which is $2,300 here, as long as the job's gross margin stays at or above 28% and the reason is a failure on your side rather than a client simply asking. This request is $5,000, about 10.9%, so it's outside the limit.

What happens next matters more than the limit itself. If the account lead's only move is "I'll ask the owner," you've built the bottleneck back in. The rule I'd attach is that anything over the limit goes up with a short brief: the contract value, the current margin, what caused the problem, what the account lead would offer (say $2,300 now plus a priority slot on the next phase), and what the client has said they'll do if they don't get it.

The owner answers within a day, and the answer goes into a shared log so that the next version of this question has a precedent. Also, if the client is the owner's old friend or the largest account in the book, that's a legitimate reason for the owner to keep the decision. Keeping it should be something the owner chooses on purpose each time.

Here is the same sorting applied to seven decisions at a made-up 22-person electrical contractor:

• Swap two crews on Thursday. Level: Operational. Why: Undone in an hour, comes up weekly, no precedent. Who should own it: Dispatcher. When it goes up: If any site falls below minimum staffing, to the service manager.
• Substitute a $90 part when the supplier is out of the specified one. Level: Operational. Why: Small, same spec, reversible on the next visit. Who should own it: Lead technician. When it goes up: Over $300 or if it touches a warranty, to the service manager.
• Credit a customer $200 for a missed appointment. Level: Tactical. Why: Bounded cost, frequent, but repeat credits set expectations. Who should own it: Office manager, up to $250. When it goes up: Over $250, or a second credit for the same customer within 90 days.
• Give $5,000 off a $46,000 job. Level: Tactical, with precedent risk. Why: One project, but the answer gets quoted by other clients. Who should own it: Account lead, up to 5% of contract if margin stays at or above 28%. When it goes up: Above that, to the owner with a one-page brief.
• Replace a technician who left. Level: Tactical. Why: Role and pay band already approved. Who should own it: Service manager. When it goes up: New role, or pay above the band, to the owner.
• Raise labor rates 8%. Level: Strategic. Why: Touches every job, hard to walk back. Who should own it: Owner, with the service manager's input. When it goes up: Stays with the owner.
• Accept a commercial client who wants net-90 terms. Level: Strategic. Why: Cash flow risk, and every other client will ask next. Who should own it: Owner. When it goes up: Stays with the owner.

How do you count the decisions that reach you?

Run it for ten working days, not five. A single week can be skewed by a holiday, a crisis or your being out of town, and ten days gives you a pattern you can trust.

A note on your phone is enough. Each time someone asks you for a yes, a call, a fix or an opinion, write one line: who asked (a role is fine), what they asked in their words, what you said, and how long it took you. If the question came by text or Slack, also note how long they waited, which you can read straight off the timestamps. Don't change any rules during the ten days, because the moment you do, the sample stops being honest.

At the end, sort each line into strategic, tactical or operational, and work out five numbers. These thresholds are my own rules of thumb from looking at a lot of these tallies, not figures from a study, so treat them as a way to read your own data rather than a benchmark.

1. Asks per working day. Under about eight and you probably don't have a volume problem; look at wait times instead. Between roughly eight and twenty is normal for an owner with a team of ten to thirty. Above twenty-five and you're operating as a switchboard.
2. Share by level. Strategic questions should be well under a tenth of the total. If more than half are operational, you have a rules problem and a habit problem. If a third or more are strategic, you're either in a genuine period of change or you're making tactical calls that should belong to someone else.
3. Same-answer rate. Of the operational and tactical asks, how many did you answer the same way you answered last time? Above 70% means there's a rule in your head that you haven't written down. That is where your first handoff comes from.
4. Your minutes. Add them up. It will undercount, because it ignores the time it takes you to get back into what you were doing, but even the raw figure tends to be larger than people expect.
5. Median wait. If people typically wait more than about four business hours for your answer, the bigger cost is their time and the work sitting still. This number is often what finally persuades an owner, because it's the one that shows up on customers.

Here is a ten-day tally for the same made-up contractor:

• Credits, discounts, waived trip charges. Asks: 24. Answered "yes" or same as last time: 20. Median wait: 2 h 40 min.
• Scheduling and crew exceptions. Asks: 14. Answered "yes" or same as last time: 11. Median wait: 1 h 15 min.
• Parts and vendor choices. Asks: 9. Answered "yes" or same as last time: 6. Median wait: 3 h 10 min.
• Hiring and pay questions. Asks: 4. Answered "yes" or same as last time: 1. Median wait: 1 day.
• Estimates over $15,000. Asks: 6. Answered "yes" or same as last time: 2. Median wait: 5 h.
• Reassurance, no decision needed. Asks: 6. Answered "yes" or same as last time: n/a. Median wait: n/a.
• Total. Asks: 63. Answered "yes" or same as last time: 40 of 57 decisions (70%).

How did one contractor cut his asks from 63 to 27?

That list comes from a composite, not a real client, but the shape is one I've seen several times. Picture a 22-person electrical contractor run by a former electrician. He was sure his problem was big estimates, because those were the ones that kept him up. The count said otherwise: 63 asks in ten days, 38 of them some version of "can I give them something." He said yes to 31 and changed the number on four. Median wait on those was a couple of hours, and one sat for a day and a half because he was on a roof.

When we traced it back, there were two causes. The first was that the limits lived entirely in his head. He had a clear sense of what he'd approve and had never said it. The second was the penalty for guessing wrong. Some years earlier he'd overruled a coordinator over a $120 credit, in front of the office, and since then everyone had quietly concluded it was safer to ask about everything.

He wrote three limits: the office manager could credit up to $250, the dispatcher could rearrange crews as long as every site met minimum staffing, and an account lead could discount up to 3% of a job. He told the team he would live with any decision inside those limits even when he'd have chosen differently. The first test came within a fortnight, when the office manager gave a $240 credit he privately thought was generous. He said nothing, which he told me was harder than any of the writing.

Three weeks later he counted again. Asks had dropped from 63 to 27 in ten days and the median wait was 35 minutes. It hadn't gone away cleanly, though. Nine of the 27 were new: people checking whether something was inside the limit. That told him which edges the rules hadn't covered, and he added two rows. One miss showed up too, when the same customer received two credits in a month, so he added a once-per-90-days condition. Neither problem was visible until the rules existed and people started using them.

When will the decision count mislead you?

A count only captures what reaches you. If people have learned that asking is painful, the most expensive decisions may be the ones that never arrive, and the sign is work sitting for no obvious reason. Pull up ten open jobs and see how many are waiting on a decision nobody feels allowed to make. The tally also misleads in your busiest month or your quietest, with a brand-new employee whose questions are training and not a bottleneck, and when you answer by silent approval, such as being copied on an email and saying nothing. Count those too.

There's one more thing worth admitting before you start. Several owners I've worked with have noticed that a good share of what they answer is stuff they like answering. One described fixing small problems as the part of her day that felt like finishing something, in contrast with the slow strategic work that never feels done. The count will show you which questions you've been keeping for yourself, and you may have to decide whether you want to let them go.

How does your week change once decisions stop waiting on you?

The test that matters most is money. Quotes go out, discounts inside the limit get given, and invoices get sent on days nobody can reach you. Decisions are also only one kind of dependency. The logins, clients and vendor relationships that never interrupt you need their own plan, which I cover in how to make your business less dependent on you.

When you have your numbers, find the row with the most asks and the highest same-answer rate, write one limit for it, and count again in three weeks. If you'd like a harsher version of the same diagnostic, what breaks first when the owner steps away shows what waits once you're out of reach.

Frequently asked questions

Should I hire a manager to take these decisions off me?

A manager helps once the limits exist. Hire one into a business where the rules still live in your head and the questions simply move to a new desk, or come back to yours when the manager isn't sure. I go into this in you can't hire your way out of owner dependency.

What if someone makes a call inside the limit that I would have made differently?

Let it stand, and bring it to your next review. If the call was within the limit and still felt wrong, the limit needs adjusting, which is a change to the rule for next time. Overturning the decision in the moment teaches everyone that the limit only applies when you agree with the result.

Where should the limits and rulings be kept?

Somewhere the team already opens every day, in one place. Each entry needs the decision, who owns it, the line, and what goes up with a brief. Rulings on cases over the line go beside it, so precedent builds up. That record is a ruling log, and why your best employee keeps bringing you problems shows how one works.

Ready to write your first limits?

If you'd like a second pair of eyes on your tally, a Planning Call is where we go through your rows together and pick the first limit to write. If outside help isn't what you need yet, we'll say so plainly.

Book a Free Call