Sold after-hours dispatch at $1.10 a door, and call volume ran 3x my model

As the guy who used to be the on-call tech, the 30% figure is generous. When I was on call for about 300 units I'd say half of what got sent to me at night could've waited. The difference is that the office staff who'd worked there four years knew which residents called about everything, and your operators in a call center didn't. That local knowledge isn't in a script.

So the per-door price was the mistake? Or was it that you priced per door at a level that only worked at your low volume estimate? I can't tell whether you're saying per-door pricing is wrong in principle or whether you'd have been fine at $2.50 a door.

194 calls in month one against 62 modeled. Did you ever get an actual number from them before you quoted? They had no after-hours line, so they had no history, which means your 8 to 10 per 100 units came from somewhere else. Industry figure, or something you built yourself?

The number your whole contract rested on is calls per 100 units per month, and you sourced it from a population that already had working after-hours coverage. Communities that have run a 24-hour line for years have trained their residents into it, and their volume reflects a steady state. A property switching from voicemail to a live line has no steady state, and the first six months of data are behavior change, not seasonality.

Per-door pricing transfers volume risk from the manager to you. That's a defensible trade when you control the variable that drives volume. You didn't. The manager controlled deferred maintenance across four properties, and deferred maintenance is the largest single driver of after-hours calls. An older property with aging water heaters and original HVAC will generate several times the night volume of a recently repositioned one at the same unit count. If you price per door on a portfolio you haven't inspected, you're underwriting somebody else's capital plan.

On the escalation question, @beacon's version matters. At $2.50 a door you'd have had $1,545 a month against a call load of 211, which is roughly $7.30 a call for triage plus the dispatch argument. That still loses if the argument over callouts isn't settled in writing, because the fee dispute is what killed the relationship, not the labor cost.

The structure that holds up is a monthly availability fee covering staffed hours, plus a per-call charge above a stated threshold, plus a written category list separating dispatch-now from log-for-morning. The category list has to be the manager's document, signed, because response obligations for habitability issues vary by state and local ordinance and that exposure sits with the owner and manager, not the answering service. Get a licensed professional in that state to review the escalation categories before they go live.

The line that stuck with me is that a working phone permanently changes resident behavior. I'm building a service business right now and I'd been modeling demand off what the client currently experiences. That's backwards if the whole point of what I'm selling is that they currently experience nothing.

Same shape as my make-ready loss, different service. Both of us priced a defined unit of work that the client hadn't defined. @plumb at least got out at nine months, I rode mine to twelve.

One thing I'd add: the 60-day paid pilot is right, but insist it's paid at cost-plus, not at your eventual rate. Otherwise the pilot price becomes the anchor for the whole contract and you've just discounted yourself in advance.

Dumb question from someone who's never run a service contract. Why would the manager pay per call when that gives them an incentive to tell residents not to call? Doesn't that just put them back where they started?

@tally because per call above a threshold, which is what was described, keeps the baseline covered either way. The manager pays for coverage regardless, and only pays more when volume runs hot. And volume running hot is a signal about their own deferred maintenance, which is information they should want.

Sitting on the buy side of this, the useful thing in the post is the diagnostic. If I'm looking at a value-add 200-unit and the seller has after-hours call logs, 200 calls a month tells me something about capital needs that the T-12 will not. Is that data anybody actually hands over in diligence, or am I dreaming?

The dispute mechanism is what I'd change first, ahead of the price. You had two parties with a genuine good-faith disagreement about whether a running toilet at 11pm is an emergency, and no document that decided it. Any contract where the fee depends on a judgment call needs the judgment standard written down and a named person who breaks ties within 24 hours. Otherwise you get nine months of invoices in dispute and a relationship that ends politely.

Would also want to know who carries what if a triage call goes wrong. That allocation belongs in the agreement and it's worth an attorney's time in the relevant state.

I've turned down two of these with my own small team. The reason is exactly the labor math you skipped past. Two people covering 6pm to 8am plus all weekend is not two people, it's closer to three and a half once you account for someone being sick, someone taking vacation, and the fact that nobody stays on permanent nights for more than about eight months. Your $540 cost model was a fully-staffed-forever model. Losing one operator in nine months isn't bad luck, that's the base rate.

@bramble on the diligence question, on-call logs exist but they're usually text threads on a tech's personal phone. Work order history in the software is the version you'll actually get, and the after-hours ones are often entered the next morning with no timestamp that tells you it was a night call. Ask for the work order export with created-time, not created-date. That's where it hides.

So the summary I'm taking away is that @plumb sold a fixed price against a variable they didn't control and couldn't measure. @cobalt's staffing point makes it worse, because even the low-volume version was underpriced on labor. The escalation document fixes the argument. It doesn't fix the model.