The Signals a Property Account Is Going Bad

Why this matters

Property accounts almost never end with a phone call. They end with a slow reassignment of the work you cared about to somebody else, while the work you did not care about keeps arriving on schedule. By the time the volume drops far enough to notice, the decision was made two or three quarters earlier by a person you may not have met.

Every signal that fires early is operational. Every signal a shop actually watches is financial. That gap is the whole problem, and the case below is what it looks like from the inside.

The signal nobody was watching

A shop had held a 400-unit portfolio for four years. Nothing looked wrong. Monthly revenue from the account was within a few percent of the previous year, the manager was friendly, and payment arrived inside one cycle every time.

What was actually happening showed up only when the owner compared quotes issued against quotes awarded, split into two half-year windows:

The award rate fell by roughly 42 percentage points across the two windows. Meanwhile routine dispatch held: 7 percent fewer work orders at about 10 percent more hours each, so the top line barely moved. The mix underneath it had changed completely. The high-margin quoted work had gone from eight jobs to five while the low-margin dispatch work stayed put, and the revenue line reported none of it.

Hypothesis 1: we got slower

The first assumption was response time, because that is what managers complain about. Median time from dispatch to arrival was 6.5 hours in the first window and 6.9 hours in the second, a rise of about 6 percent.

That is a real increase and it should not be dismissed as nothing. It is also nowhere near large enough to move an award rate by 42 points, and it does not touch the mechanism at all: quoted work is awarded at the estimate stage, before response time on that job exists. Eliminated as the cause, kept as a separate item to watch.

Hypothesis 2: we got expensive

The hourly rate had not changed in fourteen months. But average hours per work order rose from 2.1 to 2.3, about 10 percent, so the typical bill did in fact go up even with a flat rate.

This one did not eliminate cleanly, and pretending it did would have been the mistake. Bills genuinely rose. What it could not explain is the quote pattern, for the same structural reason as the first hypothesis: a quote is priced from a scope, not from accumulated hours, so a drift in hours per dispatched order cannot lose a quote. The drift stayed on the list as a real issue with its own cause, and came off the list as an explanation for this one.

Hypothesis 3: our quality slipped

Callbacks, meaning a return visit on the shop's own work within 30 days, ran 6 of 68 in the first window and 4 of 63 in the second. Both the count and the rate fell, from about 9 percent to about 6 percent.

Quality had improved, not slipped. Eliminated, and worth noting that the shop's instinct had been to blame its own techs first, which would have produced a demoralising team meeting about a problem that did not exist.

What was actually true

The manager's supervisor had changed five months earlier, and the new one had instructed that anything above a routine threshold go out for competitive quotes. The shop was still being asked to quote, more often than before in fact, but its numbers were now serving as the comparison bid on a list of three.

Nobody hid this. Nobody mentioned it either, because from the manager's side nothing had changed in the relationship. The vendor was still the vendor. The procurement rule above her had changed, and vendors are not told about procurement rules.

How they confirmed it

One question, asked directly: "Has the approval process for larger items changed on your side in the last six months?" The answer came back in a sentence and cost nothing.

Two pieces of physical evidence had also been available all along, unread. A second vendor of the same trade had appeared in the property's contractor sign-in log on days the shop was on site. And two tenants had mentioned, in passing, someone else looking at the same equipment. Both were noticed by techs and neither reached the office, because there was nowhere for that kind of observation to go.

The signals, ranked by how early they fire

Every one of these is measured per account, and the window matters as much as the number:

Run the case above through the first three rows. The award rate fell about 42 points against a 20-point trigger, with 11 and 16 quotes in the two windows, so both the minimum sample and the threshold are satisfied and the signal fires. The vendor-sighting row fires on a single observation and had fired twice, unrecorded.

The third row did not fire, and it is worth saying so rather than letting it pass as supporting evidence. Quote requests rose from 11 to 16, about 45 percent, against a trigger of more than 50 percent, and dispatch fell 7 percent, which is inside the 10 percent band. Under the rule as written, that signal was quiet. One of three signals firing was enough, because the rows are independent triggers rather than a scoring system, and the award-rate row is the strongest of the set.

What to do at each stage, and what not to do

On the first firing signal, ask one question. Not a relationship review, not a lunch, not a proposal. A single factual question about their process, sent in writing to the manager. The most common mistake is to respond to a procurement change with a charm offensive, which cannot affect a rule set two levels above the person receiving it.

If the answer is that the category is out to bid, decide whether to compete on that category specifically. Sometimes yes, sometimes no. What matters is that you now know which part of the account is contested and which is not, and you can price accordingly instead of discounting everything defensively.

Do not cut your rate across the board. The routine dispatch work was never at risk in this case, and a blanket reduction would have given away the safe revenue to defend the contested revenue, which is the exact inversion of the right move.

Fix the unrelated thing you found. The 10 percent drift in hours per work order had nothing to do with the quote losses and was a genuine problem. Investigations that turn up a second issue should close both, separately.

Building the observation channel the shop was missing

Both physical signals in this case were seen by techs and lost. The fix is small and it is the highest-return thing on this page: one field on the work order, filled in every visit, asking whether any other contractor was on the property and what they appeared to be doing.

Techs will fill it in if it takes five seconds and if somebody visibly reads it. Review the field weekly. A competitor's presence on your account is not gossip, it is the earliest signal available to you, and it lives entirely in the heads of people who assume the office already knows.

How to run this on your own accounts

Once a quarter, per account, pull four numbers and one list: quotes issued, quotes awarded, work orders dispatched, median days to pay, and the log of other-contractor sightings. Compute the award rate against the prior two-quarter window rather than against the prior quarter, because quote volumes are small enough that a single quarter swings on two jobs.

Then check the thing the numbers cannot show. Ask yourself who on your side has spoken to the manager's supervisor in the last year. On an account where the answer is nobody, every procurement change above the manager will reach you the way this one did: as a slow drift in a ratio you were not measuring, discovered a year late.

References

• See related: The Concentration Risk in One Large Property Account
• See related: How to Handle a Manager Who Changes Every Two Years
• See related: How to Avoid Becoming the Cheapest Vendor on a Portfolio
• See related: Why Property Work Goes Out to Bid and What Holds It