Decision speed vs decision quality: what a $35M business taught me

Most operators I meet are slow on purpose. They call it being careful. I call it being expensive. Running Roofed Right America to $35 million taught me the cost of a slow decision is almost always higher than the cost of an imperfect one.

TL;DR: Most operators optimize for decision quality and pay for it in speed. After running a $35M roofing business, I learned the cost of slowness is almost always higher than the cost of an imperfect call. The exceptions are decisions that are hard to reverse and decisions involving people. For everything else, ship it and learn faster than the people still talking about it.

The cost of a slow decision is rarely zero

When I started running operations, I thought a good decision was one with all the information. I would gather quotes, run scenarios, talk to three more people, then talk to two of them again. By the time I made the call, the situation had moved.

One morning a foreman called me about a tear-off in Brookfield. The membrane spec we ordered was wrong. The crew was standing on the roof. I had about ninety seconds to decide whether to send them home, switch to in-stock material, or push the schedule. The right answer was not in a spreadsheet. The crew needed an answer.

I made the call. We switched material. It cost us about $1,800 in margin. If I had waited two hours to talk to my supplier rep, we would have lost the day. A lost crew day on that job was closer to $9,000.

That is the math nobody runs. Slowness has a price. It is usually higher than the price of being wrong. The people who deliberate never see the bill because the bill arrives later, as a missed window, a competitor who moved, a customer who went elsewhere, a team that stopped bringing problems to the boss because the boss takes a week to respond.

Most decisions are reversible. Treat them that way.

Jeff Bezos called these Type 2 decisions. I call them doors you can walk back through. If the call goes bad, you can fix it. Reorder the material. Refund the customer. Pull the marketing spend. Restructure the route. The downside is bounded.

For Type 2 decisions, speed is the strategy. Make the call with what you have, watch the result, adjust. The information you collect by acting is better than the information you collect by asking. Markets and crews and customers tell you the truth. Meeting rooms tell you opinions.

I run a quick test in my head. If I make this call and it turns out wrong, what does the recovery look like? If the recovery is a phone call, an email, or a check, I move now. If the recovery is a lawyer, a layoff, or a year of damage, I slow down.

That second category is where speed gets you killed. Which brings me to the exceptions.

Exception one: decisions you cannot unmake

Some doors only swing one way. Selling the company. Signing a five-year lease. Taking on a partner with equity. Firing your top customer to chase a bigger one. Capital expenditure on equipment you cannot resell. Bonding capacity commitments. These are Type 1 decisions, and they deserve the slow treatment.

The mistake I see operators make is treating Type 2 like Type 1. They run a $40,000 truck purchase through the same committee process as a $4 million acquisition. The truck is reversible. You can sell it next year and lose a few thousand. The acquisition can sink you.

When I face a Type 1 decision, I do the opposite of my normal process. I slow down on purpose. I sleep on it. I bring in people who will tell me I am wrong. I write out the worst case and ask if I can live with it. I look for the version of the deal that gives me an exit if the assumptions break.

If a decision will compound for years, an extra week of thinking is cheap. If a decision will play out in days, an extra week of thinking is the loss.

Exception two: decisions involving people

I have moved fast on hiring and fired the person inside six months. I have moved fast on firing and watched a team lose someone who was about to turn a corner. People decisions punish speed in a way that operational decisions do not.

The reason is that you cannot run the experiment. You cannot rehire someone you fired badly. You cannot un-promote someone you elevated past their range. Reputations move with people, into the market, onto Glassdoor, into the next foreman’s group chat. Your team watches how you treat the person leaving more than they watch how you treat the person staying.

So I slow down on people. Not on whether to act, but on how. If I know someone is not going to make it, I still owe them a clean process. Documented expectations. Real coaching. A conversation that respects the work they did even when the fit is wrong. That takes time. It is worth it. The team is doing the math on whether you are someone they want to work for, and they grade you on the people decisions more than any other.

The same goes for hiring. Move fast on the screening, the interviews, the offer. Move slow on the gut check. If something is off, do not paper over it with the pressure of an open seat. Open seats heal. Bad hires bleed.

How to get your team to ship without you

The single biggest constraint on a growing company is the founder’s inbox. If every decision needs your signature, the company runs at your speed, and your speed is bounded by your sleep.

The way out is to give your team a frame, not a rulebook. I tell mine three things.

First, name the door. Is this Type 1 or Type 2? If it is Type 2, you do not need me. Make the call, send me a one-line note, move on. If it is Type 1, bring it to me with your recommendation, not your question.

Second, set a financial threshold. I picked numbers that scaled with the role. A project manager could spend up to a certain figure without asking. A division head had a higher number. Above the number, escalate. Below it, decide. It tells people I trust them inside a defined range, and it stops the small stuff from clogging my calendar.

Third, post-mortem the bad ones together. If a team member makes a Type 2 call that goes wrong, we do not punish the speed. We pull the decision apart, find the missing piece of judgment, and the next ten decisions get sharper. If I punished bad outcomes, my team would stop deciding. They would route everything back to me, and I would become the bottleneck I was trying to remove.

What a good post-mortem actually looks like

Most post-mortems I see are blame ceremonies dressed up in a meeting invite. That is not useful. The point of a post-mortem is to separate the decision from the outcome. A good decision can produce a bad result. A bad decision can produce a good result. If you only grade the result, you teach your team to be lucky.

I use four questions.

What did we know when we made the call? List the actual information available at the moment of decision. Not what we know now. What was on the table then.

What did we assume? Every decision rides on assumptions. Some are about the market, some about the customer, some about our own capacity. Name them.

Which assumption broke? This is the one that matters. If the decision was sound and an assumption broke that we could not have known, that is a different lesson than if the assumption was lazy. The first means the world surprised us. The second means we surprised ourselves.

What would I do again with the same information? If the answer is the same decision, the process was right and the outcome was just variance. If the answer is different, write down the new rule and put it where the next decision will find it.

I keep a running document of these. After ten years, it becomes a private playbook. It is more valuable than any book on management I have read.

Practical takeaway

If you want to move faster without getting reckless, run this checklist on the next decision in front of you.

Ask if the decision is reversible. If yes, decide today. If the recovery from a wrong call is a refund, a reorder, or a redo, you are not betting the company. You are running an experiment.

Ask if the decision involves a person’s livelihood, reputation, or trust. If yes, slow down on the how. Do not slow down on the whether. Avoiding a hard conversation is a decision too, and it usually compounds badly.

Ask what you will know in two weeks that you do not know now. If the answer is nothing, decide now. Waiting will not improve the information. It will only cost you two weeks.

Ask what the cost of waiting actually is. Put a number on it. Lost margin, lost momentum, lost employee patience. The number is rarely zero.

Then make the call. Write down what you decided and why. When the result comes in, grade the decision, not the outcome. That is how operators get better.

Most of the leaders I know who got stuck around $10 million stayed stuck because they would not delegate decisions. Most of the ones who broke through to $35 million and beyond did so because they trained their teams to make calls at speed and they kept their own bandwidth for the doors that only swing one way.

If you want help thinking through how your team handles decisions under pressure, or if you are looking for a keynote that gives your leaders a real framework instead of a poster, book me to speak. I bring the math, the receipts, and the stories from running a real operating company.