Most regional brokerages in the Dallas–Fort Worth market have an active CoStar subscription. The data is there — lease comps, sale comps, submarket vacancy, asking rents, historical absorption. The problem is not access. The problem is process: what happens between pulling data out of CoStar and delivering a defensible comp set to a client or counterparty.
That gap — between raw CoStar data and a structured, repeatable comp package — is where regional brokerages lose ground. Not on relationships. Not on market knowledge. On the quality and speed of the analytical work product that supports the deal.
What "comp data quality" actually means
When we talk about comp data quality in the context of CRE brokerage, we're talking about three distinct things that often get conflated: coverage completeness, attribute accuracy, and comparability precision.
Coverage completeness is whether the comp pool you're drawing from includes the transactions that actually happened in your target submarket over the relevant time window. CoStar's coverage is strong for large metro markets like Dallas, but not every lease gets reported to CoStar — particularly smaller, privately-negotiated deals in suburban submarkets like Las Colinas or Frisco. CompStak fills some of those gaps via crowdsourced broker submissions, but it requires a separate workflow to integrate both sources.
Attribute accuracy is whether the deal-level data fields — asking rent PSF, effective rent PSF, TI allowance, free rent abatement, lease term, GLA, lease commencement date — are populated correctly and consistently. In practice, CoStar transaction records often have incomplete concession data. A comp might show an asking rent of $32.50/SF NNN without disclosing the 6-month free rent period or the $55/SF TI allowance that were part of that deal. The contract rent looks better than the economics actually are.
Comparability precision is the hardest of the three. It requires human judgment to determine which transactions from the pool are genuinely comparable to the subject lease — matching on building class, submarket micro-location, floor plate size, load factor, and lease structure. A Class A tower in Uptown Dallas and a Class B low-rise in the Stemmons Corridor may both appear in the same CoStar submarket pull, but presenting them side-by-side in a comp package is analytically misleading.
How the gap shows up in real deals
Consider a leasing scenario that plays out regularly in the Dallas office market. A tenant rep is working a 12,000 SF requirement in the Preston Center micro-market. The asking rent is $38/SF full-service gross. The tenant's counsel wants to see market comps before approving the LOI. The broker pulls a CoStar comp set filtered by the Preston Center submarket, 10,000–15,000 SF range, Class A, last 18 months.
The raw pull returns eight transactions. Two of them are renewal comps, which have different economics than new deals — no TI package, potentially below-market rates driven by retention rather than market clearing. Two more are sublease transactions, which carry different structural risk and pricing. One has incomplete concession data. That leaves three genuinely comparable direct new-lease comps, and one of those is in a building with a substantially different common area factor — the load factor difference alone affects the usable area calculation enough to change the effective $/USF comparison.
A broker who processes that eight-comp pull carefully gets to a defensible position. A broker who presents the raw output — or who works under deadline pressure and doesn't filter it — delivers a comp package that will draw scrutiny from a sophisticated tenant counsel. At best it creates a back-and-forth that slows the deal. At worst it undermines credibility at a critical negotiation moment.
The spreadsheet problem is a process problem
To be clear: we're not saying that CoStar's data is unreliable. CoStar Group has built the most comprehensive commercial real estate transaction database in the US, and for most major markets, the coverage depth is genuinely strong. The data quality problem we're describing lives downstream of CoStar — in the manual workflow that most regional brokerages use to transform raw transaction data into structured analytical work product.
The typical process looks like this: export data from CoStar into Excel, manually review each row to flag anomalies, apply filters in the spreadsheet, format the table for a client presentation, update rent figures to adjust for concessions when data is available, and repeat this for every new comp pull. Each step introduces opportunities for error and each repeated comp pull requires starting over from scratch.
This isn't a skills problem — experienced CRE analysts know how to do this work. It's a time and systematization problem. A process that takes 2–3 hours to execute correctly on every comp pull becomes a bottleneck when a Managing Broker has 15 active requirements in the pipeline simultaneously.
Why concession data is the biggest blind spot
Of all the attribute accuracy issues in CRE comp data, TI allowance and free rent abatement are the most consequential and the most commonly missing. Landlords and tenants both have incentives not to disclose full concession packages — the landlord wants to maintain face rents for future negotiations, and the tenant's counsel may treat concession details as confidential.
The result is that published asking rents in submarkets like the Dallas CBD and Uptown consistently overstate effective rents during soft market conditions. Comparing a subject lease asking rent against a CoStar comp pool that doesn't capture full concession packages produces a distorted picture of where the market is actually clearing.
CompStak's crowdsourced model attempts to fill this gap — the platform relies on broker contributions of completed lease terms in exchange for comp credits, and because the contributing brokers have firsthand knowledge of the deals, concession data is often more complete than what CoStar captures. But CompStak requires a separate subscription and a different pull workflow, and integrating both sources into a unified comp analysis adds time and complexity.
What a structured comp process looks like
The brokerages that handle this well share a common trait: they've built repeatable filters and documentation practices rather than treating every comp pull as a one-off analytical exercise. They know in advance which CoStar data fields they always check, which transaction types they always exclude from specific analyses, and how they document the exclusion rationale. They track which comps they've used in previous deals so they can identify when a comp has moved from "recent" to "stale" as market conditions shift.
The other structural advantage these brokerages have is internal comp history. Over years of deal-making in a specific submarket, a brokerage accumulates first-hand knowledge of deal terms that never made it fully into CoStar — TI packages, free rent periods, landlord work letters, lease commencement timing. That institutional knowledge, if systematically captured, makes their comp packages more accurate than what the public data sources alone can produce.
The challenge is that this institutional comp knowledge lives in email threads, PDF packages, and the memory of individual brokers. When a broker leaves the firm, the knowledge leaves with them. When a Managing Broker wants to review comp accuracy across the team, there's no systematic way to do it.
The cost of getting it wrong
Inaccurate or poorly structured comp packages have downstream consequences that extend beyond the individual deal. A tenant rep who delivers a comp set that opposing counsel picks apart loses credibility in the negotiation. A landlord rep who overstates effective market rents based on incomplete concession data sets pricing expectations that fall apart during due diligence. A Managing Broker who can't audit comp accuracy across the team has no way to identify systematic errors before they compound across the pipeline.
Regional brokerages compete on local market knowledge. That knowledge is only as credible as the work product that conveys it. Comp data quality is the foundation of that work product — and it deserves a more systematic approach than most firms currently give it.
If you're evaluating whether your brokerage's comp process is exposing you to the risks described here, we're happy to walk through what that looks like in practice. Request a demo and we'll connect to your CoStar account to show you what a structured comp workflow surfaces from your own deal history.
