Help us direct you to the right place to sign up
Accurately estimating market rent is critical to commercial real estate decisions, from underwriting acquisitions to evaluating lease renewals. Yet traditional rent forecasting methods often rely on historical comparables and manual adjustments that can make it difficult to account for differences between properties, spaces, and market conditions.
To address these challenges, CompStak developed Rent Predictor, an AI-driven solution that estimates current starting rent and effective rents for specific commercial real estate spaces using CompStak's extensive lease transaction data.
In our latest whitepaper, we explore the methodology behind Rent Predictor, the data and modeling techniques that inform its estimates, and how predictive analytics can support more informed commercial real estate decisions.
Two Methods Behind Every Market Rent Estimate
Every Estimated Starting Rent Today on a CompStak lease comes from one of two models. Which one runs depends on how much lease data CompStak has for that market and space type.
- Space Insight is a comparable adjustment model. It picks similar leases, adjusts each one for how it differs from your space, and blends the results. It covers 44 market and space type combinations: 24 office markets and 20 industrial markets.
- Stratified Rent Index (SRI) carries a lease's signed rent forward using a quarterly rent index for its cohort. It covers office, industrial and retail nationally.
Space Insight runs wherever its data coverage supports it. SRI covers everything else, including retail and markets outside Space Insight's 44 combinations. So a Class A office lease in Seattle and a retail lease in Manhattan, built a different way.
The next question is how Space Insight decides which leases count as comparable.
How Space Insight Adjusts Comparable Leases
Space Insight works the way an appraiser does. It just does it at scale.
- Find comparables. Property CompSet searches millions of properties using two signals: location, and building characteristics such as size, year built and building class. A model trained on compsets that CompStak users built themselves then ranks the candidates.
- Adjust each one. For every comparable lease, a market-specific model estimates the dollar impact of each difference: floor level, building size, transaction size, building class, year built, and sublease or renewal status. Industrial models also adjust for ceiling height and loading docks.
- Weight and blend. Each comparable gets a weight based on how similar the property is, how small an adjustment it needed, and how recent the lease is. The final estimate is the weighted average.
Each market is trained separately. In markets with deep recent coverage, the model leans almost entirely on fresh comparables. Where data is thinner, it allows older leases to contribute so the estimate stays stable.
How the Stratified Rent Index Carries Rents Forward
SRI answers a simpler question: how much has rent moved for leases like this one since it was signed?
For each market and space type, SRI builds a family of quarterly indexes. They range from market-wide down to submarket, refined by building class and property subtype. A granular index is only used if it tracked real renewals more accurately than the market-level baseline. Otherwise SRI falls back to the market index.
Here's a real example from the platform. A retail lease executed in April 2025 at $82.50 per square foot saw index growth of +3.48% between Q2 2025 and Q4 2026, or 2.22% annualized. Its Estimated Starting Rent Today is $85.37.
Leases signed within the most recent quarter get no adjustment at all. Which raises the obvious question for anyone underwriting off these numbers: how close do they land?
How Accurate Is the Market Rent Estimate?
We test Space Insight on a rolling basis against comparables that were available at the time, using only comparables that were available at the time. Accuracy is reported as median absolute percentage error (MdAPE), which keeps a few extreme leases from skewing the result.
- The latest evaluation covered 3,659 leases executed since July 2026, across all 44 market and space type combinations.
- Space Insight's overall MdAPE was 6.0%.
- On renewals, where tenant and property stay the same, Space Insight's MdAPE was 5.8% against 9.3% for SRI. It had the lower error in 21 of 31 market and space type combinations.
Accuracy varies by market. It's strongest where CompStak has the most recent, closely matched leases. Phoenix industrial came in at 3.9%, for example, while Houston office was 11.5%. The whitepaper lists all 44 results.
Data from CompStak's proprietary database. Learn more at compstak.com.
Where You'll See It in CompStak One Connect
Every covered lease in CompStak One carries an Estimated Starting Rent Today field. For a Space Insight estimate, the Rent Prediction tab shows up to ten comparable leases behind the number. Each one comes with its similarity score, its attributes and its own estimate, plus a map of the subject and comparable properties.
Take the 2022 office lease from the top of this post. It signed at $55.00. Its Estimated Starting Rent Today is $68.12, and you can check that figure one comparable at a time, the same way you'd review an appraisal.
For an SRI estimate, you get the index trend from the execution quarter to today and a breakdown of the calculation.
Rent Predictor also runs on any office or industrial space, including vacant space and new construction. The same estimates are available through the Space Insights API, so the number in your model matches the number on the platform.
Frequently Asked Questions
What is a market rent estimate in commercial real estate?
A market rent estimate is the rent a space would likely achieve if leased today. It's built from comparable lease transactions and adjusted for differences in the space, the building and the timing. Asset managers use it to mark leases to market, size rollover risk and underwrite renewals.
How does CompStak calculate Estimated Starting Rent Today?
CompStak uses one of two models. Space Insight adjusts and blends a weighted set of comparable leases in 44 office and industrial market combinations. The Stratified Rent Index carries the original signed rent forward using a quarterly index for its cohort, and covers office, industrial and retail elsewhere.
How accurate is Rent Predictor?
In the latest evaluation, Space Insight had a median absolute percentage error of 6.0% across 3,659 leases executed since July 2026. Accuracy is strongest in markets with deep recent lease coverage and varies by market and space type.
Can Rent Predictor estimate rent for vacant space or new construction?
Yes. Rent Predictor runs the same estimate on any office or industrial space, including vacant space and new construction, using comparable leases rather than an existing signed rent.
See the Comparables Behind Your Rent Estimate
The point of a market rent estimate is to make a decision with it: hold or sell, renew at what number, or what rent to underwrite on rollover. Rent Predictor gives you that number along with the comparables and adjustments behind it, so you can defend it in an investment committee.
Download the full whitepaper for the complete methodology and market-by-market accuracy results, or see Rent Predictor on your own portfolio.
Related Posts
Data Science: CompStak's Space Insights Rent Prediction