Vacancy Prediction in Commercial Real Estate: What Turnover and Lease Data Already Show

Vacancy prediction in commercial real estate is the practice of reading turnover, payment, and lease renewal data a property management system already stores as one combined early signal of tenant departure, months before formal notice arrives.

A vacant unit in a shopping center rarely arrives without warning. Turnover reports decline for two quarters. Payment amounts shrink and arrive later than required. A renewal deadline passes without a word from the tenant. Each signal already sits inside the property management system a portfolio runs, recorded for a different purpose and read by a different team. Together, they produce a warning six months ahead of a formal notice that otherwise arrives with six weeks left. This article maps what a vacancy costs, the order these signals appear, and what six months of lead time buys a leasing team.

Key Takeaways

• A vacant unit costs more than lost rent: common area costs shift to the landlord, occupancy metrics move, and the next lease gets negotiated under time pressure.
• Turnover decline, payment softening, and renewal silence typically appear in that order, months before a tenant files formal notice.
• Six months of lead time replaces a reactive scramble with active choices: a rent holiday for the next tenant, planned refurbishment, or an early leasing campaign.
• Combining these signals is a data aggregation exercise inside the property management and sensor systems a portfolio already runs.

What Does a Vacant Unit Actually Cost a Landlord?

Lost rent is the visible cost. It is rarely the largest one.

A shopping center allocates common area maintenance, or CAM, across every leased unit based on floor area or revenue share. Once a unit sits empty, its share of cleaning, security, and landscaping cost does not disappear. It transfers to the landlord, and in some lease structures, to the remaining tenants. A portfolio with a handful of long-term vacancies absorbs that reallocation every month the units stay dark.

Occupancy rate is also a portfolio-level metric that lenders, investors, and asset management committees track on a fixed schedule. A vacancy that appears suddenly moves that number in a single reporting cycle, which draws attention at the exact moment a leasing team has the least room to negotiate calmly.

The last item on that table explains why timing matters as much as the total cost. A leasing team with six months of runway can plan a campaign, adjust terms, or line up a replacement tenant before the unit ever appears empty on a report. A team that learns about the departure six weeks out negotiates from a weaker position and often accepts worse terms just to close the gap quickly.

Which Signals Warn of Tenant Departure Before Formal Notice?

A commercial tenant rarely decides to leave overnight. The decision builds over months, and each stage leaves a trace in data the property management system already collects.

• Turnover decline. Retail leases with a percentage rent component require tenants to report sales regularly, which means turnover-based rent reports exist for almost every retail unit in a shopping center. A tenant whose reported turnover drops for two consecutive quarters is showing the earliest visible sign of trouble, well before any conversation about renewal takes place.
• Payment pattern shift. Late payment is a well-known warning sign, but by the time it shows up, the tenant has usually been under pressure for months already. The earlier version of this signal is subtler: payments that stay on time but shrink in amount, or a tenant that starts requesting short extensions instead of missing a due date outright.
• Renewal silence. Most commercial leases carry a defined window before expiry during which either party can raise renewal terms. A tenant planning to stay typically initiates that conversation, or at minimum responds to an inquiry. Silence through that window, with no counteroffer and no request for an extension, is itself a signal, not an absence of one.
• Footfall decline. Shopping centers commonly run people-counting sensors at entrances and sometimes inside individual units or wings. A sustained drop in traffic near a specific unit, isolated from seasonal or center-wide patterns, often lines up with the same period as declining turnover.

None of these signals is decisive on its own. A single quarter of soft turnover can reflect a slow season rather than an exit. The signal becomes reliable once two or three of these categories move together over a sustained period, which is exactly the pattern a combined view is built to catch.

What Does a Tenant’s Six-Month Path to Departure Look Like in the Data?

A mid-size apparel retailer in a regional shopping center offers a useful illustration. None of the figures below are drawn from a real case. They represent the kind of pattern property managers describe when they walk back through what a system already recorded after a tenant leaves.

Viewed at any single point on that timeline, each entry looks like routine noise. A retailer’s turnover dips for a quarter regularly, for reasons that have nothing to do with an exit. A late payment happens once and gets resolved.

The pattern reads differently once someone lines the entries up against each other for the same unit. By month three, three separate systems, turnover reporting, payment processing, and sensor data, are pointing in the same direction. That alignment is the signal a decision layer built on top of the property management system is designed to surface. Most teams currently treat the formal notice as the first indication of a problem.

What Can You Do With Six Months of Lead Time Instead of Six Weeks?

An early signal is valuable mainly for the choices it keeps open while the unit is still occupied.

• Offer a rent holiday as a negotiating tool. A landlord who knows a unit is likely to come open in six months can pre-negotiate incentive terms with a prospective replacement tenant. That beats offering concessions reactively after the space has already sat empty for a quarter.
• Time refurbishment to the vacancy. Fit-out or renovation work planned around a known departure date runs during the void period that was going to happen anyway. That avoids adding extra downtime after the unit is vacant and a new tenant is waiting to move in.
• Run an active leasing campaign instead of a reactive one. Marketing a unit before it is empty, with a defined handover date, attracts a different pool of prospective tenants than marketing a unit that has already been dark for two months.
• Reconsider the unit’s category. A departing tenant is also an opportunity to evaluate whether the same retail category still fits the center’s current tenant mix, rather than defaulting to a like-for-like replacement under time pressure.

Every option on that list depends on lead time. None of them is available once a formal notice arrives with six weeks left on the lease.

How Do You Combine Turnover, Payment, and Footfall Data Into a Vacancy Risk Score?

Turnover reports, payment records, and lease renewal dates already live inside the property management system. Footfall data usually lives in a separate sensor platform tied to the building’s security or facilities infrastructure. The technical work is joining those existing sources for the same unit over the same time window.

Three steps structure that work in most portfolios.

1. Inventory what already exists per unit. Confirm that turnover reporting, payment history, lease renewal dates, and, where available, footfall data can all be pulled for the same unit identifier and aligned on a common timeline.
2. Start with a rule before a model. Flag a unit when turnover declines for two consecutive quarters and the renewal window passes without contact from the tenant. That simple rule surfaces at-risk units before anyone builds a statistical model.
3. Validate against tenants who already left. Run the rule or score against units that became vacant in the past two or three years, and check whether it would have flagged them early enough to matter, before it drives any live leasing decision.

This is an aggregation problem across data the portfolio already owns. A shopping center that already runs percentage rent, tracks lease dates, and operates people-counting sensors has every input this approach requires. The remaining work is connecting three systems that currently answer separate questions into one that answers the question a leasing team actually needs answered: which units are at risk, and how much time is left to act.

Software that handles lease administration and CAM reconciliation well is widely available today. Choosing which platform to run that layer on is a separate decision from reading the signals the current system already produces.