A rate increase that buys a move-out loses money

Almost every existing-customer rate increase program starts the same way: find the tenants paying well under street, raise them toward it, and bank the difference. The logic is clean, the spreadsheet is easy, and it is wrong often enough to matter.

It is wrong because the gap to street measures opportunity, not outcome. The outcome depends on something the gap says nothing about: whether the tenant stays.

The arithmetic nobody runs

Take a tenant paying $118 in a unit that streets at $149. A 12% increase puts them at $132 — about $14 a month, or $168 over a year if they stay.

Now suppose that increase raises their chance of moving out in the next 90 days by eight percentage points. What does a move-out cost? The unit sits empty for some period, you pay to re-rent it, and the replacement tenant likely comes in on a promotion. Put a conservative number on that — say one month of lost rent plus a discounted first month — and a single move-out costs well over $150.

Multiply: an 8% chance of a $150-plus loss is roughly $12 of expected cost against $168 of expected gain. Still positive, but nothing like the $168 the spreadsheet promised. Push the increase to 25% and the churn probability does not rise politely in a straight line — it accelerates. At some point along that curve the expected value crosses zero, and every raise past it is a decision to lose money with confidence.

The question is never "how far below street is this tenant?" It is "what is this increase worth, net of the move-out it might cause?"

Why the gap survives as a rule of thumb

Because it usually works. Most tenants in most units at most reasonable increments do stay, and a portfolio-wide program built on gap-to-street will produce a revenue lift. That lift is real, which is what makes the method so durable.

What it hides is the distribution. The program nets positive overall while a specific slice of it — long-tenured customers in soft unit types at stores with real competition nearby — loses money on every raise. Averaged into the portfolio, those losses disappear. Reported per unit, they are obvious.

What a defensible recommendation needs

To price the churn side of that equation you need to know who the tenant is, not just what the store did last month. Monthly aggregates cannot tell you which tenant to raise. The specific inputs that matter:

• A unit-level rent roll — in-place rent, tenant tenure, and the date of the last increase, per unit.
• Your own current asking rates by unit type, which is what "street" actually means here.
• Time since the last increase, because two raises in six months behave nothing like one.
• Occupancy by unit type, since a raise in a type running at 96% is a different bet than one at 78%.
• Recorded outcomes from past increases, so churn sensitivity is fitted to your customers rather than assumed.

The uncomfortable part

An engine built this way recommends fewer increases than a gap-based one. That feels like the software is underperforming — until the first quarter where move-outs do not spike after the batch goes out.

The honest framing is that the held-back raises were never revenue. They were revenue on the spreadsheet and a cost in the rent roll, and the only difference between the two programs is which one told you before you sent the letters.