The industry has spent the last decade investing in data: more platforms, more market reports, more analytics tools. The thinking made sense: better data would mean better forecasting, and better forecasting would mean fewer surprises.
It hasn’t worked out that way.
The volume of available data has never been higher. But the ability to turn that data into accurate forecasts hasn’t kept pace. The market has changed in ways most forecasting tools weren’t built to handle, and the gap between what the data says and what’s actually happening on the ground is widening.
The Signals Are Noisier, Not Clearer
The problem isn’t a lack of data; it’s that more data hasn’t meant better data. The same forces that expanded access to market information also expanded the volume of low-quality, outdated, and inconsistently sourced inputs sitting alongside the useful stuff. Most teams don’t have a clean way to separate the two. The forecasts built on that data inherit the noise without knowing it.
The market itself has compounded the problem. Demand is shifting across tenant types, locations, and deal timelines in ways that are hard to track. AI-driven investment is reshaping where leasing demand concentrates, pulling activity toward certain markets while pulling it away from others. By the time that shows up in a quarterly report, it’s already priced in.
MetLife Investment Management’s 2026 outlook makes a similar point: leasing demand is now less tied to national trends and more tied to where high-paying jobs are growing. A market can look healthy nationally while specific submarkets are quietly losing demand, and vice versa. The aggregates hide as much as they reveal.
Why Traditional Models Are Breaking Down
Most forecasting models were built on lagging indicators: past absorption rates, old comps, historical trend lines that describe where the market was, not where it’s heading. That worked well enough in a stable environment. It works less well when demand is shifting mid-cycle and the data hasn’t yet caught up.
The input teams aren’t necessarily wrong. They’re just stale. A comp from 18 months ago, a trend line built on pre-2024 behavior, and an absorption model trained on a different rate environment. These describe a market that no longer looks the same. Acting on them with full confidence is where the exposure comes in.
What the Better Approach Looks Like
The answer isn’t to stop forecasting. It’s to rely on it differently. The teams handling this well are moving away from single-point projections and toward scenario-based thinking. Rather than asking “what will happen,” they’re asking “what happens in each of these scenarios, and what do we do in each case?” It’s a harder way to operate, but it’s more honest about what the data actually supports.
The other shift is moving closer to the signal. Leasing teams that replace lagging market reports with real-time demand intelligence, actual prospect engagement, broker activity, and submarket interest as it’s happening are working with a fundamentally different picture than those waiting for quarterly data to confirm what the market has already moved on. The advantage isn’t more data. It’s data that hasn’t gone stale by the time it reaches the forecast.
That’s as much an organizational question as an analytical one. The teams building that capacity now, with tighter feedback loops, faster signal-to-action cycles, and a clear process for adjusting when the forecast is wrong, will be better positioned than those waiting for the market to settle before they adapt.
