What is predictive vacancy analytics in commercial real estate?
Predictive vacancy analytics is the practice of modelling the probability that a specific tenant will vacate a specific unit, derived from observed signals in that tenant's business rather than extrapolated from market vacancy rates. Scayled is a predictive vacancy analytics layer purpose-built for industrial and logistics property: it ingests occupier-level signals such as contract events, financial deterioration, M&A and footprint announcements, then outputs a departure probability and an action window for each tenancy that aggregates to a portfolio-wide outlook. The hard part is not the model but the occupier signal data underneath it, which is exactly what most generic AI-for-real-estate tools lack and what Scayled is built to supply.
- Predicting tenant departures, not extrapolating market rates
- The inputs that actually predict a departure
- The output: a probability and an action window per unit
- Why generic AI for real estate fails without the right data
Predicting tenant departures, not extrapolating market rates
There are two very different things people call vacancy prediction, and conflating them is the most common mistake in this space. The first is forecasting market vacancy: taking a submarket's current availability rate and trend and projecting where it goes. That is useful for acquisitions, development appraisals and strategy, but it says nothing about whether the tenant in your unit, on your rent roll, is about to leave.
Predictive vacancy analytics in the sense that matters to an asset manager is the second thing: estimating the probability that a named tenancy ends in a void, and when. It is a bottom-up, tenant-by-tenant calculation, not a top-down market average. A submarket can be at 3 percent vacancy while your anchor tenant is quietly preparing to consolidate out of your building, and the market figure will never warn you.
Scayled operates entirely at the tenancy level. Each unit's vacancy probability is built from what its current occupier is actually doing, then rolled up to give a portfolio outlook. The market rate becomes context for re-leasing, not a proxy for your own income risk.
The inputs that actually predict a departure
A move out of an industrial unit is the end of a chain of operational decisions, and that chain leaves a trail. The inputs that genuinely predict departure are the ones tied to how the tenant runs its business. Contract events are among the strongest: a 3PL winning a major new retail mandate will need more space, while losing one removes the reason it held the cross-dock at all. Both are leading indicators of a move, in opposite directions.
Financial deterioration matters because distress forces footprint decisions: profit warnings, restructuring, covenant breaches and divestments frequently precede a site rationalisation. M&A is a particularly reliable trigger in logistics, where an acquirer consolidating two overlapping distribution networks will hand back the redundant facility, often within a year of the deal closing. Network and footprint announcements, a new automated mega-DC, a stated move to fewer, larger nodes, a regional exit, point directly at which existing sites become surplus.
Two subtler inputs round out the picture. Hiring direction reveals intent before bricks move: a surge of senior supply-chain and operations hires in a region signals expansion there, often at the expense of legacy sites elsewhere. And lease activity by the same legal entity at other locations, a sister company touring or signing space across town, is a strong tell that the occupier is on the move. Scayled monitors all of these per tenant, because no single signal is decisive but together they are predictive.
The output: a probability and an action window per unit
Analytics is only useful if its output drives a decision. For each tenancy, Scayled produces two things an asset manager can act on: a departure probability, and an estimated action window, the rough period during which the fund can still influence the outcome before the unit empties. A high probability with a twelve-month window is a re-gearing or re-leasing project to plan now; a high probability with a three-month window is an urgent backfill.
Those per-unit outputs aggregate into a portfolio-level vacancy outlook that is grounded in tenant behaviour rather than assumption. Instead of an ARGUS model that applies a uniform renewal probability across the rent roll, the fund gets a differentiated view: these three tenancies carry most of the near-term vacancy risk, here is the income exposed, and here is when each window closes. That is a far more honest input to cash-flow planning, hold-or-sell decisions and capital-partner reporting.
Because the analysis refreshes fortnightly, the outlook is a living view, not a point-in-time report that is stale the week after it is produced. Probabilities move as new signals arrive, and the ranking re-sorts so the team's attention follows the risk.
Why generic AI for real estate fails without the right data
The modelling here is not the hard part. Estimating a probability from a set of features is well-understood, and any competent data team can build a scorecard. The reason most AI-for-real-estate tools fail at predictive vacancy is that they are trained on the data the property industry already has: rent rolls, lease terms, payment history, market comps. None of that contains the leading signal, because the event that empties a unit happens in the tenant's business, not in the lease record.
An AI model fed only historical lease and arrears data will, at best, rediscover that tenants near expiry sometimes leave and that tenants in arrears sometimes default. It cannot see the contract loss, the acquisition or the new mega-DC announcement that is the actual cause, because that information was never in its inputs. Garbage in is the polite version: more accurately, the predictive signal was simply absent from the dataset.
Scayled's work is in the data layer, not the dashboard. The defensible part is assembling, per tenant and per surrounding submarket, the occupier-level operational signals that precede a move, and keeping them current. The score sits on top of that. Access is by request. Request access and Scayled works your first at-risk unit free, so you can see the signal evidence behind a real departure probability on your own portfolio rather than take the analytics on trust.
Fill your first vacancy free
Request access and Scayled monitors every tenant in your submarket for movement signals, then identifies verified replacement tenants for your first vacancy at no cost. See the value on your own portfolio before you pay anything.
Fill Your First Vacancy Free →