How does a property management company go from a 12% margin to 53% without hiring anyone?

Transcript, lightly cleaned for reading

0:00 Sam's story

A property manager named Sam came to me and said he needed to hire, but he couldn't, because his profit margins were too low. At the same time he was turning people away, because he was at capacity. He was also the one answering the phone for every little thing: weekends, nights, any given time.

The strange part is that two years ago they had 600 doors. Since then they've doubled their doors, and their profit stayed the same. They were barely breaking even. Last year they hired more people to take on more doors, but the margin stayed where it was, because the math was not mathing. It didn't fix anything. No matter how much he grew, he wasn't growing in revenue, in profit margin, in capacity, or in anything that matters when you grow a business.

So he felt stuck. He wanted to sell the business, because it was no longer serving him and he wasn't really making that much money from it. Even though, on paper, his monthly gross revenue was $112,000. They had 25 people, 1,120 doors, $112,000 a month in revenue, only a 12% margin, and a team running at 132% capacity.

In this video I'm going to show you how I approached auditing this company, how we fixed the operations, used data, and then implemented AI. This is how I do a complete audit, how I map out a business, how I find what's really capping you and why you're stuck where you're stuck, and then how we turn all of it into systems. At the end, I'll show you what happened 90 days after they went live.

1:11 His numbers

Before we go into anything, the first thing to look at is the numbers. Revenue was $112,000 a month. They had 1,120 doors, priced at about $100 a door. With 25 team members, that's 45 doors per person. The industry average is 65 to 100 doors per person.

Given those numbers, their capacity was at about 132%. They were always staying late, there was a long backlog, and a lot of what was going on wasn't documented, processed or written down anywhere. Realistically, the business owner had no idea what the team was doing.

The last thing to look at is where the money comes from. About 92% of this company's revenue came from the base management fee, and almost nothing from leases, renewals or maintenance. The profit margin was about 12%, which means they were taking home about $13,500 a month.

2:07 The audit

The first thing in the audit is to interview the whole team and understand what their day to day looks like. What are they doing that the business owner doesn't realize they're doing? I talked to every single role: the owner, Sam, the operations manager, and each individual person in the company. I had them walk me through their day out loud, literally from when they log in to when they log out.

I also had them keep a log of everything they were doing, every 20 to 30 minutes. That's annoying, but it's really helpful, as a business owner and as someone from the outside, to see where the time is actually going. For two weeks they did that log alongside the interviews.

While they were doing that, I got access to all of their tools: the property management software, project management, task lists, the CRM. Anything they had, I had access to. We extracted all the data and looked for patterns. Are there constant emails that look alike? Are there frequently asked questions from the team, or from outside? Are there patterns we can turn into rules, decision frameworks or judgment frameworks, so they stop being questions your team has to ask you?

3:10 The map

Then I take everything we just talked about and turn it into a map. The map helps you understand and visualize the process, and it helps me see where things go wrong. Where are the bottlenecks? Where are the handoffs? Where are the operational issues? Where do you not have data, have too much data, or have the wrong data? There are so many questions you don't know the answer to.

To map out a process you need a few things. First, a trigger: when is this process supposed to happen? Is it frequency based, by time of day or week, or is it conditional, when something else happens? Then the input: what information does the person doing this process need coming in, so they can execute without a back and forth? Then the steps themselves. On top of the steps you can add decision frameworks and judgment frameworks: conditional logic, if this then that statements, that let your team and your AI make judgment calls as an extension of you. You don't need to be in every conversation for something to operate as an extension of you.

Then the output. What information is this process producing, where does it go, and in what format? An email notification, a PDF report? Then the owner: who is responsible for this task? Even if AI is doing it, you want a human being responsible for it. And then the completion condition. A completion condition should be a binary statement, like "the contract is signed." Nobody can argue with it, and it marks the process as complete. Wherever there are gaps in those fields, that's what we're looking to solve.

For this company, after mapping all of it out, we found 214 recurring tasks across 25 people. Here's one. The trigger: a tenant reports a maintenance issue, so that's a conditional trigger. The task: create a work order and assign someone to it. It's done when a vendor has confirmed they'll take the job and confirmed an appointment time, and the tenant knows they're coming. There are several parts to that completion condition. The inputs are the vendor list, the type of problem, maybe past experience with each vendor. Based on their volume, and the analysis of their emails and tenant requests, this one process was taking 96 hours a month, and it was owned by the maintenance coordinator.

Then I put all of these in order, because the completion condition of one system often becomes the trigger for the next, and the output of one system becomes the input of the next. Map it in the right order and you stop creating unnecessary back and forth.

Here's an example of a bad workflow in this company. A tenant texts the property manager. She forwards it to the shared inbox. Someone calls a vendor, and we don't know who, because it wasn't assigned to anyone. The vendor texts back a time. Someone types it into the software, if they remember. Then someone calls the tenant to let them know. That's realistic. All of these workflows look like this at first. A handoff is the same information getting passed from person to person or tool to tool, and when different people handle the same process you get a lot of them. That causes miscommunication, and nobody knows what's going on.

6:20 The diagnosis

Now that everything is mapped, we have to diagnose the problem. We can already see the biggest ones: bottlenecks and handoffs. How we solve those with AI comes later in the video. First we have to understand where the capacity number from the beginning actually comes from.

From the audit, the interviews, the mapping and the data analysis, I added up every hour each door actually takes. For this company it was 5.1 hours per door. Where do those hours go? About an hour and a half goes into actual property work: inspections, conversations with owners and tenants, managing vendors, leasing. The main job. The other three and a half hours was around the job. 1.8 hours went into retyping the same information from one system into another, because the tools don't talk to each other. 1.2 hours went into handoffs, chasing and follow ups that nobody wants to do, and that could have been resolved with basic systems in the right order.

At 5 hours per door, each person could realistically carry about 34 doors. They were carrying 45, and that was still below the industry average. That's where the 132% comes from, and that's why they felt buried. Sam was getting frustrated: the industry average says you should be able to handle more, so why can't you? But it wasn't that they were incapable. There were literally no systems in place.

It also explains why the profit margin kept dropping while the doors went up. Marketing was working. This is how operations can cost you far more than you think when the systems aren't there.

One more thing: not every door is equal. About 200 of their doors are in older buildings, which means way more maintenance requests, three times more to be exact, for the same fee. Those doors cost more to manage. And remember, 92% of their revenue was the base management fee. The best firms make 20 to 40% of their revenue on everything else: leasing fees, renewal fees, maintenance fees. So either this company wasn't charging them or wasn't collecting them consistently. When I talked to Sam about it, they were supposed to be collecting them. They just weren't, because it slipped through the cracks. Nobody owned it. And with 12% margins, it's not like you can hire. Hiring was never going to be the answer here.

8:45 The fix: operations, then technology

The fix happens in two moves. The first is operations. How do we turn every process we just identified into something that would work well even manually? We redesign the entire structure. Trigger, input, the steps, the decision frameworks, the output, the owner, and the completion condition. All seven need to be present for a system to work, and they need to run in the right order so dependencies don't cause delays. This is also where the org chart changes. Some roles merge, some split, so you don't have five people inside the same process.

Then you layer in technology. For this company, we fixed all of their systems first, and then built the technology in layers. The first layer is data. From the 11 tools they were using, which nobody was updating, we went to six, with one source of truth underneath and one tool per function. You don't need multiple CRMs or multiple project management tools. One of each function, all connected to an AI tool such as Claude Code, and your tools are talking to each other. You don't go in and open them. You talk to Claude Code, it tells you what you need to know, and it updates things itself.

Then you add skills. The skills are the decision rules we came up with in the systems in the first place. The way you delegate to AI is very similar to how you delegate to a human being. Once this works and your team is trained on the company brain, you go from 55 open tabs and nothing working to one open tool, everything connected, everybody on the same page. Then you layer in agents, which actively take tasks off your plate.

10:18 The demo

This is what it looks like when it's running. For this team we created a team dashboard, because nobody knew what was going on. Every door, every open request, every renewal coming up, every fee to follow up on, and how full each person's plate is. Sam can see capacity before he hands more work to people who are already overloaded.

Here's a maintenance request coming in. A tenant texts that the kitchen sink is leaking. The agent reads it, marks it urgent but not an emergency according to the rules we were given, picks the plumber who covers that area and has a good track record with them, books the first open slot, and texts the tenant the time. The property manager gets one line: handled. Nothing needed from a human. The six handoffs from the process we started with became one.

Here's renewals. 90 days before every lease ends, the renewal agent checks the market rate for that unit, drafts the offer with the new rent, and puts it in front of the property manager. Once they approve it, it goes to the tenant.

And here's what Sam liked most. Someone calls with 40 doors they need managed. Before, he would have had to say no, because he didn't have the people or the capacity. Now he opens the dashboard, sees exactly who can take it on and how much room the team has, and says yes. That owner gets set up incredibly quickly.

Here's the first of the month. The owner statements are already drafted from the bank feed, from the data itself. Accounting's job is to review and approve them, not to build them. You can imagine how much faster that gets.

11:44 The results

The goal was to get from five hours a door to two. At two hours a door, each person can handle 85 doors, and the same 25 people can handle 2,125 doors. At their old price, that takes them from $112,000 to $212,000 a month. That's $100,000 a month from that change alone, without a single new hire.

And because the service got faster, they could charge more. Maintenance requests used to get a response in maybe two days. Now everything is done within four hours. Owner statements went from day eight to day two. New owners used to take three weeks to set up, now it's five days. When you're that fast and the owners can see everything, you stop being the cheapest option and start being the reliable one. So they raised their price from $100 a door to $150, and they started charging the maintenance fees, renewal fees, and the other fees that normally make up 20 to 40% of a firm's revenue.

12:36 90 days after go-live

Here are the numbers after everything was done and live for 90 days. Team size: 25 people became 17. These were people he had already wanted to let go, but couldn't, because then who would do the little bit they were doing. Capacity: 132% became 83%, with eight fewer people and 80 more doors. They stopped turning owners away. Monthly revenue: $112,000 became $180,000. Profit margin: 12% became 53%. That's about a million dollars a year in profit.

13:10 Why he never needed to hire

He came to me wanting to hire, wanting to sell, wanting to get rid of the whole thing. This was his last chance to make the numbers work. Between operations and AI, he was able to do a lot more than he thought was possible. His problem was never really his people. It was the systems. That's how you go from a people based company to a process, workflow and AI based company.

If you need help doing this in your company and you're making over $100,000 a month, you can apply through the link in the description. If you're on YouTube, subscribe. This is a new channel and it's my first video. I finally have enough stories to take you behind the scenes of how I've produced results for these clients. If you have any questions, leave me a comment.