The first month — what we got right, what we got wrong.
The first month is over. Twelve landlords' portfolios are being managed from a small office on Bathurst north of Bloor. Three hundred and twelve units are on the rent roll. The first statements went out on the first; the second statements go out on Friday. Most of what surprised me about the month is in this letter, including the parts I'd rather not have to write down.
We accepted twelve of the thirteen portfolios we spoke seriously to. The one we declined was a large portfolio of student rentals near the university — a different operating problem than the rest, and one we didn't have the bench to do well. The landlord took it well; we suggested two PMs who specialize in that segment and forwarded the financials with the landlord's permission. Telling someone you can't do the work is the first version of the work being done.
A few things went better than I'd expected.
The trades dispatch took half the time we'd budgeted. I'd planned for forty hours a week, across the team, on dispatch and follow-up — calls to the bench, scheduling, invoice reconciliation. It took twenty. The reason isn't that the volume was low (the volume was about what we'd modelled). The reason is that the bench, when it's the same five or six people across the whole book, starts answering on the first ring within two weeks. A scheduling call that should take six minutes takes ninety seconds. By week three, Marco at Aurora was calling us before we called him, on a unit he'd been to twice. The compounding is faster than I'd modelled.
Renewal letters went out on time. We sent twenty-two renewal proposals in the month, ninety days ahead of each lease end. Eleven came back accepting at the asked-for number in the first two weeks. The other eleven are in various stages of back-and-forth or wait-and-see. The renewal cadence is the one operating detail that, more than anything else, separates a real PM from a PM-themed listing agency. I'll write more about this in next month's letter, which will be about the cost of the placement-fee that didn't happen.
Notices went out the day they were eligible. We filed four N4s in the month, two of which resolved within seventy-two hours (one with a payment, one with a payment plan), one of which is at the Board, and one of which the tenant cured before the hearing date. The N4 is the part of this job that operators care about and software companies don't understand. Filing it the day it's eligible — not the week after, not the day before the hearing — changes the outcome more often than the people who haven't run a board file would expect.
Two things went worse.
The bookkeeping took twice as long as budgeted. I'd planned for sixty hours a week across the team on accounting work — daily rent posting, trust-account reconciliation, vendor payments, monthly statement preparation. It took one-twenty. The reason isn't that the work is harder than I'd thought. It's that the work is spikier than I'd modelled: 80% of it falls in the first five and last five business days of the month. We staffed for the average and got crushed by the peak. Hiring a second bookkeeper next month, part-time at first, full-time by the end of the quarter if the second month confirms the pattern.
One trade invoice went unpaid for eleven days. Eli at Cedar Electric did a hallway repair on March 22 and invoiced the same day. We didn't cut his cheque until April 2. That's a problem because the entire trade bench works because we pay on the day they invoice. The reason is small — our accounting platform's vendor approval workflow had a step we didn't know about — but the consequence is large: I called Eli to apologize, took him to lunch the following week, and rebuilt the workflow so it can't happen again. The bench is built on three things: the call answered, the work done well, and the cheque on the day the invoice arrives. We failed at the third one once, and once was a problem.
A few things I didn't expect either way.
Landlords used the inbox more than the phone. I'd planned for the bulk of landlord communication to be a Tuesday-afternoon phone call. The pattern that emerged was different: a stream of short questions, sent by email, expected to be answered within a day. The questions were mostly about specific units (a tenant's request, an invoice line item, a market question). Few were urgent, but the answering was structural — we built a queue and a same-day SLA in the first week. The portal we hadn't planned to build — a simple landlord-facing view of every email thread on every unit — is now on the road map for July. The version of it we have today is an inbox with rules.
Tenants almost never used the inbox. Most contact from tenants came through SMS or the building intercom. We have a single phone number that rolls to whichever operator is on duty. The thing we'd assumed would be a tenant-portal feature was just being a phone number that answers. The portal that handles tenant requests will exist eventually; today, the SMS-to-operator routing works better than a portal would.
Two of the twelve portfolios are not yet at the state I'd want them to be by month one. One is a multi-unit on the Beaches with a maintenance backlog we underestimated; we are catching up but it's going to take through May. The other is a five-unit on Spadina with a tenant on a long-running rent arrears file that the previous PM hadn't been actively managing. We are now actively managing it. The landlord knows. The tenant knows. Neither of those situations is going to be a surprise on the May statement.
The rest of the month is the work. The May letter — out on the first of June — will lead with the math of the eleven-of-twelve renewal rate, and the cost of the placement-fee that didn't happen. The June letter, depending on the month, will probably be about the maintenance backlog on the Beaches portfolio and what we learned from underestimating it.
If you'd like these by email on the first, write to hello@jaraygroup.com.