Something I’ve noticed talking to property managers at different portfolio sizes — the breaking point is almost always the same. It’s not leasing, it’s not tenant communication. It’s maintenance reconciliation.
Here’s what the workflow looks like manually at scale:
Maintenance request comes into Property Meld
Vendor completes the work and sends an invoice
Someone on the team manually cross-checks the invoice against the original quote
Then manually verifies a completion photo was actually submitted
Then manually pushes it into the accounting system
At 50 doors that’s manageable. At 300 doors, that’s a part-time job that nobody budgeted for.
The firms that scale cleanly are the ones where that entire chain request → vendor → invoice validation → photo proof → ledger entry runs without anyone touching it.
Curious if anyone else has hit this wall. What did you do to fix it hire another coordinator, switch PM software, or find another way?
Sorry, if you don't have a decent system by the time you hit 50, you are not likely to see 300. Also, in your scenario, who pays for the mistakes, Client or PMC when you are managing for dozens of owners? You have no control over how invoices are prepared, they must be reviewed by a human for both accuracy and to code single or split payments. Who approves the vendor dispatch to begin with, and based on what? What if additional work is determined to be needed by vendor when on site?
Versus, as PM, I respond to all tenant calls, email request, and, if I am in the office, walk ins, triaging, prioritizing, and dispatching or holding based on initial info provided 99% of the time. Vendors contact me direct from site for approval or additional items depending on their findings upon arrival. All incoming invoices for properties I am responsible for are routed to my desk upon receipt by admin or through regular mail. I review, code, approve, and hand off to payables accounting.
For large projects, I meet with our well vetted contractor on site to review scope. They are familiar with our standards, and produce an estimate with sufficient detail for the owner to understand. Depending on the owner, I may meet him/her at the property either with the contractor or separately, to walk through the scope and options. More often than not however, the owner is OOS so they have received extensive photo series and my written notes of the issues. Once the owner signs and accepts the proposal, they provide payment in full for us to deposit in Client Trust Account. We then, depending on scheduled start date and ordering details, issue payment for materials deposit. Generally, we do not have multiple draws, but if we do, we still hold back 10% until after final acceptance either by myself or by owner if they choose. I make at least a weekly walkthrough taking pics and reviewing progress and status with the contractor. Pics may or may not be sent to owner, depending on their wishes, but all are filed permanently with clear reference to the specific project and date.
In addition, I create and post all ads, show all vacancies, screen all applicants, draft, review docs with applicants, sign and perform move in, move out, and interim inspections. In my spare time I do nighttime drivebys of multi-fam properties checking on security lighting and related issues. Also for multi-fams I meet with the Fire Marshall annually upon request to inspect the property for Life Safety issues. I remotely manage access control and video systems on several properties. I meet with Owners at properties for a walk through when they come to town once a year, making arrangements with the tenants to do so.
All of that for an average of 125 - 150 rental units for the past 20 years. Plus, for the last 10, as Managing Agent for numerous COA/HOA's totaling an additional 200 units, made regular visits to the properties for a wide variety of issues, including meeting with and educating the Boards to accomplish the mandate of the By-laws. Prepared annual budget and reserve funding plans, met with contractors, handled rule violations including pulling screen grabs from security video systems.
I don't see AI taking over any single process effectively...yet. There also is no Killer PM app that I have seen. They are primarily financial/accounting focused, and generally poor at collecting and managing non financial data at the level of detail needed while being user friendly. There are far too many variables depending on type of property, type of rental, local laws and restrictions, type and number of ownership interests, desired reporting details, tax issues, etc. for one app.
Did I mention, for the last 10 years I only worked a 4 day week? While handling a few sales on the side? Not that difficult if you have the systems and stick to the systems, and have a solid understanding of the issues that really matter for operations. These are the usual breakdown points, regardless of number of doors.
Richard, appreciate the detailed breakdown that’s a genuinely well-run operation. The 4-day week after 20 years of refining your systems is the goal for any serious PM.
The problem I’m describing shows up at a different point when a firm is trying to replicate you across 10 property managers instead of one. The judgment calls you describe (coding invoices, approving dispatch, reviewing scope on site) work perfectly when one experienced person owns the whole chain. They start breaking down when you’re trying to get 10 different people to execute that chain consistently across 300+ doors.
That’s where the manual process becomes the bottleneck not because the process is wrong, but because it doesn’t scale without the person who built it.
Curious whether you’ve seen firms try to grow past that single-PM model and what broke first.
Not to be argumentative, but as I said, a system is a system. McD's has dozens of employees all trained to, and in large part, executing to, the same standards. PM is no different, except accounting or reception do not require additional people at the same ratio as PM positions. You need to either hire, or effectively train the additional PM's, and in both cases you need the standards, policies, strategies, and achievable goals pre-determined. OJT is not an effective solution in this business. When you do have those systems in place, you keep your clients for decades, and most of your tenants 5 years or longer. Taking on an additional unit is just another Tuesday. Taking on a client with 50 units at once is a little more work for a couple weeks, but still well within the system. No real difference aside from actual initial data entry, which accounting has responsibility for most of, based on the management agreement and existing rental agreements.
That’s a well-built operation and you’re right that the McDonald’s model works when the standards are set and enforced consistently. The firms that have that discipline scale well.
The ones I’m thinking about are earlier in that journey they know they need the system but they’re still building it, often while managing growth at the same time. That’s where the cracks show up, because the training and the volume are happening simultaneously.
Your point about data entry being mostly accounting’s responsibility is interesting in your experience, where does that hand-off typically break down when a firm is still figuring out its standards?
Accounting really just handles accounting, which includes usually short, cryptic descriptions of invoices and of course owner and tenant primary info. Accounting would not pay invoices that did not have corresponding COI and License in their file. PM handles everything else and has all responsibility. I always had an Owner property file for each street address (or HOA project), with the property tax info, management agreement, insurance info and claims, owner correspondence, copies of contracts (accounting maintained invoice copies) and copies of reports sent to owner, whether condition reports or financial. Secondly I had tenant files by address and unit number where applicable, with current application package, rental agreement and correspondence copies, and all prior complete rental agreement packages; copies of tenant notices and repair invoices, insurance dec page when required, copy of new appliance manuals, and copy of HOA docs. All emails and pics were filed by property address, unit, and purpose. I would send emails to myself with details for specific properties, and file them in either owner or tenant file depending on what the detail was. Sometimes both.
The problem with one man shows is that they rarely have the requisite knowledge, so they fail a little, or a lot, on every topic, and often don't even have rudimentary office skills with regard to basic organization and establishing policy. I have yet to see even a midsize office that was not using spreadsheets for a wide variety of non-financial lists, resulting in numerous, out of sync lists used throughout the office. Quite simply, there are no shortcuts in this as an actual business.
@Paul Olusola This feels less like a staffing problem and more like a process-design problem. Hiring another coordinator may relieve the pressure temporarily, but it also adds another manual handoff. Before switching platforms, it may help to standardize vendor requirements, approval limits, invoice matching, photo proof, and exception handling, then automate everything that follows the normal path. Ideally, staff only touch mismatched invoices, missing documentation, or repairs above a set threshold. If the current software cannot support that workflow through native features or integrations, then a platform change may be justified—but only after the desired process is clearly mapped.
exactly right. The platform switch conversation almost always happens before the process is mapped, which is why firms end up with the same problems on new software.
The automation piece is where it gets interesting once you define what "normal path" looks like, the exception handling becomes much smaller than people expect. Most of the manual work is chasing things that should have been automatic.
Paul, I think you’re onto something with the maintenance reconciliation piece. At that scale, it’s not really the individual repair that becomes the problem, it’s the number of times the same process has to happen across hundreds of properties.
The request -> vendor -> invoice -> completion -> ledger chain is especially important because the accounting entry needs to match what actually happened at the property. I’d also want the system to preserve the documentation behind each expense rather than just pushing a number into the ledger. That becomes really valuable when you’re trying to distinguish routine repairs from larger capital improvements and when you need clean property-level records at tax time.
At 300 doors, I’d much rather automate the repetitive validation and leave people handling exceptions than add more people whose primary job is checking information that the system could reconcile. The real win isn't just fewer hours, it’s fewer errors and better financial visibility by property.
Feel free to DM me, I’d be happy to send over a few resources that might be helpful.
@Paul Olusola We're managing several hundred doors in Rochester, NY, and I agree that these handoffs become increasingly important as you scale.
Where I'd push back a little is on making the entire process touchless.
We operate property management and general contracting businesses, and one of the biggest issues is making sure responsibility is clearly assigned when something falls outside the normal workflow.
A vendor changes the scope, an invoice exceeds authorization, documentation is missing, or an expense gets assigned to the wrong property.
Those exceptions need an accountable person, not just another automated step.
I think the opportunity for AI is handling the routine work while escalating exceptions to the right people, with a clear audit trail.
Are you finding the biggest breakdowns happen before the work is completed or during invoice reconciliation?
@Mark Updegraff that's exactly the model. The system handles the routine 80% touchlessly matched invoice, confirmed photo, correct property, within authorization. Everything that falls outside that scope change, amount discrepancy, wrong property code gets routed to a named person with the specific exception flagged, not just = something went wrong.
The audit trail piece is built in by default. Every step logs who touched it, what the system did automatically, and what got escalated and why. That's actually what makes the exceptions easier to manage the routine work isn't cluttering the queue anymore, so when something real needs a human it's obvious.
To your question from what I've seen, the breakdown happens in both places but for different reasons. Pre-completion it's vendor communication requests that sit open because nobody followed up. Post-completion it's reconciliation invoices that sit unmatched because cross-referencing is manual and nobody gets to it until end of month.
The first creates tenant complaints. The second creates accounting headaches. Both are solvable at the workflow level.
There is another handoff after the completion photo: does the person living there know the repair is finished, and did it actually resolve their problem? An invoice can match the quote while the same issue keeps coming back.
I would keep the original request linked to any return visits and give someone ownership of confirming the outcome. For a guest-facing operation, closing a work order and closing the loop with the guest are two separate jobs.
The part I'd add is the owner side. When invoice matching gets loose, the mistake doesn't stay in your office, it shows up on an owner statement. A $400 plumbing bill charged to the wrong unit or the wrong owner is a lot harder to unwind after the owner draw goes out than before.
What I've seen work is pretty boring. Every bill gets entered against the property and the work order before it's paid, nothing gets paid without the invoice attached, and someone looks at each owner's statement for odd repair spikes before distributions go out. Whatever software you're on, that last review is what catches most of it.
Richard's question about who pays for the mistakes is the real one. If the answer is "the owner, until they notice," that's what breaks a firm at 200 doors, not the workload.
Different seat on this one. I am not a third party manager. We own the doors, 454 of them, majority Section 8, and we have run our own crews and our own management in house since 2003. So take this as a report from the other side of the fence rather than a disagreement.
The invoice-versus-quote reconciliation you are describing is a third party vendor problem, not a scale problem. It exists because the person doing the work and the person paying for it have different incentives and the scope was written loosely enough to argue about. We mostly do not have it, because the recurring work is done by people on our payroll. The invoice is payroll. There is no variance to reconcile.
That is not advice for everybody. It only pencils above a certain door count and above a certain geographic density, and it converts a variable cost into a fixed one, which is exactly the wrong trade in a vacancy spike or a soft leasing quarter. But it is why the wall you are describing is not our wall.
Here is what replaced it, because something always does. Two things.
Crew utilization. A crew you own gets paid whether or not there is a work order. The metric stops being is this invoice correct and becomes how many hours did the crew actually put on the ground this week versus in the truck. Scattered doors across a metro will eat a third of a day in drive time and nothing in your software will flag it, because drive time does not generate a line item. Density is the whole ballgame once you own the labor.
The date calendar. This is the one I would put ahead of maintenance reconciliation for anyone running assisted units. Every voucher unit carries dates that are not yours and that nobody invoices you for. The annual inspection. The window to cure after a failed item. Your HAP contract anniversary, which is what your rent adjustment request runs against. The authority's own payment standard publication, which runs on a different clock entirely. Miss a re-inspection window and the housing assistance payment abates, and the abated period generally does not come back to you. That is cash straight off the top with no bill attached, and no bill attached is precisely why nobody builds a process for it. At fifty doors you remember. At three hundred you do not, and no single miss is big enough to force a system until you add them up at year end.
On your actual question, the three options you offered. Not another coordinator. Coordinators are the symptom.
And a caution on automating the chain you laid out. The manual cross-check exists because the quote and the invoice disagree, and they disagree because the scope was never written tightly. If you automate the chain before you standardize the scope, you have not removed the judgment call, you have just made the wrong answer arrive faster and with less friction to catch it. Write fixed-price line items for the twenty jobs that are eighty percent of your volume first. Once the quote and the invoice are supposed to be identical, reconciliation becomes a match instead of a judgment, and then it automates almost trivially. Do it in the other order and you have bought software to rubber stamp variance.
Simon's point about the owner statement is the right one to pair with this. The review that catches things is a human looking at repair spikes before the draw goes out, and that review does not get cheaper with better software. It gets cheaper with tighter scopes.
@Mark Updegraff Agree on the exceptions. The routine path is easy to automate, it's the scope change at 4 p.m. on a Friday that needs a name next to it. What's worked for teams I've talked to is deciding the owner of each exception type up front: scope change goes to the PM who dispatched, over-authorization goes to whoever holds the owner relationship, missing COI or photos goes back to the vendor coordinator, and wrong-property coding goes to accounting. Each one gets a due date, and nothing gets paid until it's cleared. Running PM and GC under one roof probably helps, since your own crews' scope changes are easier to catch than a third-party vendor's. When an exception does slip, where does it usually show up first for you, the owner statement or the vendor calling about payment?
You can NOT automate maintenance!
So, many things that can go wrong:
1) Was everything actually done?
2) Quality of work acceptable?
3) Were cheaper solutions found?
4) Cost over-runs on materials? Labor?
You tech guys are clueless about reality:(
You're right the work itself can't be automated. Someone still has to show up, fix the pipe, and do it right.
What get automated is everything around the work. Dispatching the right vendor, capturing their confirmation, updating the tenant, matching the invoice when it comes in.
The coordinator job isn't fixing things. It's chasing people who fix things. That part can run automatically.
From the bookkeeping side, the thing that breaks first is the invoice-to-ledger handoff, and it breaks quietly. At 300 doors the invoices still get paid, so nothing looks wrong. What's actually happening is the coding drifts: repairs vs. improvements gets miscoded, expenses land on the wrong property, and descriptions are too cryptic to reconstruct at tax time.
My company sees this constantly in PM firms that grew fast. The fix isn't a new platform first, it's invoice data hygiene. Standardize what a complete invoice looks like before it reaches accounting: property address on every invoice, vendor name matching the W-9, and enough description to code it without guessing. Richard's point about cryptic descriptions is exactly right. Accounting can't code what it can't read.
And watch the repairs vs. improvements split. At scale, a $3,500 job coded wrong doesn't just move a number, it distorts depreciation for years. My company has seen PMs expense capital improvements across dozens of properties because the person coding the invoice didn't know the difference. That's not an automation problem. It's a knowledge problem, and it has to be solved before the automation is worth anything.
Actually, my comments regarding cryptic descriptions is a reflection on the often limited field size for the data entry OR display and lack of customization ability in reports, as well as accounting personnel's attempt at condensing wordy invoice details.
Paul, you are right that maintenance reconciliation is where scale shows cracks. From the operator seat, the failure mode is usually not “we forgot to answer the tenant.” It is request → vendor → invoice → completion photos → ledger coding that still needs a human to match every time. When that chain is manual, mismatches land on owner statements after the draw goes out, and unwinding a mis-coded plumbing bill is harder than catching it before close.
What we insist on even at smaller portfolio size: authorization limits before work, close-out photos in the unit file, invoice matched to the ticket before it hits the owner statement, and a clear owner of exceptions (scope change, over-auth, missing docs). Leasing and chat can look fine while the back office is quietly drifting.
Owners feel scale problems as “my statement does not match what happened at the property.” Fix the reconciliation loop and a lot of trust issues go with it. Happy to compare notes by message.
The part that helped me most was separating the routine check from the judgment call. If an invoice matches the approved amount, has a photo, and is tied to the right unit, it should move through without someone re-reading it three times. The time saved is not the point by itself. It leaves a person free to look closely at the odd ones, where scope changed or the price jumped. I also found a short weekly review of open, unmatched jobs catches more than a big clean-up at month end, when everyone has forgotten what actually happened on site.