How to Pull Accurate Comps for Vacant Land

Residential real estate has a well-built comping ecosystem. MLS data is comprehensive. Zillow and Redfin produce automated valuations calibrated against millions of transactions. Appraisers pull three comps and make adjustments for bedrooms, bathrooms, square footage, and condition — variables that are well-defined and consistently measured across properties. The entire system assumes a world of improved properties with standardized features.

Vacant land exists outside that system. And the investor who tries to use residential comping infrastructure for land will either get no data at all or get data that is dangerously misleading.

Why the Standard Tools Fail

MLS coverage for vacant land is structurally incomplete. A significant share of vacant land transactions never appear on MLS. Most land deals — particularly in the sub-$50K segment that drives the land flipping business — are cash transactions. U.S. News & World Report reported in 2024 that most land deals close in cash. When there is no lender requiring an appraisal and no agent collecting a commission, there is no mechanism pushing the transaction through MLS. Agent commissions on land sales run 5% to 15% (compared to 5% to 6% for homes), which creates strong economic incentives for both buyers and sellers to bypass the traditional brokerage process. The result is that MLS captures a fraction of the land transaction market, and the fraction it misses skews heavily toward the lower-priced deals where land flippers operate.

Zillow explicitly excludes vacant land from its Zestimate product. No widely available consumer-facing AVM exists for vacant land.

Automated valuation models do not cover vacant land. The company's FAQ page states that vacant land is not eligible for a Zestimate. This is not a case of poor performance; it is a complete absence of coverage. Redfin's automated estimate similarly appears to exclude unimproved parcels, referencing property characteristics like year built, bedrooms, and bathrooms that do not exist for raw land. No widely available consumer-facing AVM exists for vacant land. The European Banking Authority has noted that AVM application is largely confined to residential improved property. The technology gap is real: the data inputs that make residential AVMs work — standardized features, dense transaction histories, comparable improvements — simply do not exist in the land market.

The comparability problem is fundamentally harder. Two five-acre parcels in the same county can differ in value by 300% based on access quality, utility availability, zoning designation, flood zone status, wetland presence, topography, and soil conditions. Adjusting for bedrooms and square footage is straightforward because those variables are well-defined and their market impact is well-studied. Adjusting for "this parcel has a deeded easement across BLM land while that parcel has paved county road frontage" requires market-specific judgment that no algorithm has reliably automated.

Where the Data Actually Lives

The absence of MLS and AVM coverage does not mean the data does not exist. It means the data lives in different places, in different formats, and requires more work to extract.

County property appraiser and tax assessor records are the most reliable source for actual transaction prices. County records capture all deed transfers — not just those involving agents or lenders. In most jurisdictions, the recorded sale price reflects the actual consideration paid. The limitations are data freshness (some counties update monthly, others quarterly or less frequently) and the presence of non-arm's-length transactions (family transfers, tax sales, foreclosures) that show nominal consideration and must be filtered out.

Deed transfer records and documentary stamp tax offer an alternative method for calculating sale prices. In states that levy transfer taxes — Florida charges $0.70 per $100 of consideration, for example — the tax amount recorded on the deed can be used to back-calculate the sale price even when the price is not stated explicitly. This technique is standard among appraisers and data aggregators. However, 16 states do not levy transfer taxes, including Texas and Arizona — two of the most active markets for land investors. In those states, deed records may show only nominal consideration, making county appraiser records or direct transaction research the necessary alternative.

Land-specific listing platforms fill part of the gap left by MLS. LandWatch, Lands of America, and Land.com are the dominant platforms, all now owned by CoStar Group following a 2017 acquisition. Listings syndicate across the network and to 45+ partner sites. These platforms are useful for understanding current asking prices and days on market, but asking price is not sale price — and since all three share the same CoStar-owned listing pool, they are not independent data sources. Treat them as a single source showing seller expectations, not market values. For a broader view of where to find deals, see Sourcing Channels Compared.

Commercial data providers offer cleaner, more comprehensive transaction data at a cost. DataTree (owned by First American Financial) provides property data, ownership history, and recorded documents but is primarily enterprise-grade with no public pricing. ATTOM covers 155+ million properties through yearly license agreements at enterprise pricing; individual investors typically access ATTOM data through third-party tools like PropStream. Regrid is the most accessible option for individual investors, offering a self-serve Data Store at $200 per county for spreadsheet data or $400 per county for shapefiles, covering 159 million parcels across 3,229+ counties.

Facebook Marketplace and Craigslist are increasingly significant channels for land sales in the sub-$30K segment. Craigslist maintains a dedicated land category with active listings in most metros. Facebook groups focused on land investing and FSBO land sales have grown to 75,000+ members in some cases. These platforms capture real market activity but produce unstructured data that vanishes when listings are removed — useful for gauging the market in real time, but not for historical analysis.

The Step-by-Step Comping Process

Comping vacant land is a five-step process. Each step matters, and the most common errors come from skipping steps three and four.

Define the comp criteria before searching. Useful comps share the same county (preferably the same township or section), similar acreage (within roughly 50% of the subject parcel's size), similar characteristics (access type, utility availability, zoning classification), and a recent transaction date — ideally within 6 months for active markets, or 12 months as a standard window. Setting these parameters before searching prevents the unconscious tendency to select comps that support a desired value rather than comps that represent the market.

Pull five to ten potential comps. Use county records and listing platforms together. County records capture transactions that listing platforms miss; listing platforms show active inventory and recently sold parcels with photos and descriptions that county records lack. Cast a wider net than seems necessary — several candidates will be discarded after the verification and adjustment steps.

Verify each comp individually. This is the step that separates reliable comps from noise. Confirm each transaction was arm's-length — not a family transfer, tax sale, foreclosure, or other distressed disposition (unless distressed sales are the relevant market). Verify the comp parcel's key characteristics: access type and quality, utility availability, flood zone status, zoning designation, topography, and any restrictions. These are the variables that drive adjustments, and getting them wrong produces garbage outputs regardless of how carefully the math is done.

Adjust for differences. This is where experience and market knowledge matter most. Common adjustment factors and approximate ranges for vacant land:

Road access quality accounts for significant value differences. The spread between paved public road frontage and private dirt-road access typically runs 20% to 40%. Landlocked parcels — those without any form of legal access — face far steeper discounts, often 50% to 80% below comparable parcels with clear access. The article on legal access verification covers this distinction in detail.

Utility availability (municipal water and sewer versus well and septic versus no utilities) typically drives a 15% to 30% adjustment. Parcels where all utilities are available at the lot line command premiums; parcels requiring well drilling and septic installation require corresponding discounts.

A parcel in a flood zone faces a 20% to 50% value discount for vacant land, with the high end applying to parcels entirely within high-risk zones.

Flood zone status creates material value gaps. Academic research on improved properties finds SFHA designation discounts of 4% to 12%, but for vacant land the impact is substantially larger because flood zone status constrains what can be built, triggers mandatory insurance requirements for any financed buyer, and shrinks the buyer pool.

Zoning differences produce adjustments of 10% to 25% for within-category comparisons (agricultural with residential permitted versus agricultural with conditional use only, for example). When comparing across major category boundaries — agricultural versus fully entitled residential — the delta can reach 50% or more, though at that point the parcels may not be truly comparable.

Per-acre pricing decreases as parcel size increases. This inverse relationship is well-established in appraisal literature: a one-acre parcel might sell for $15,000 per acre while a ten-acre parcel in the same area sells for $5,000 per acre. Appraisers sometimes value "excess" land at one-third to two-thirds of the core site's per-acre rate. Failing to account for the size-to-value relationship is one of the most common errors in land comping.

Topography and usability adjustments of 10% to 20% apply for moderate differences (gently rolling versus flat, partially wooded versus cleared). Extreme terrain — slopes exceeding 25%, heavy rock, or significant portions of the parcel that are physically unusable — can push adjustments well beyond that range.

Calculate an adjusted value range, not a single number. After applying adjustments to each comp, the results should produce a range — low, mid, and high. The width of that range is itself useful information. A tight range (all adjusted comps falling within 15% of each other) indicates high confidence in the valuation. A wide range (adjusted comps spanning 50%+) signals that the market for this type of parcel is thin, the adjustment factors are uncertain, or the comps are not truly comparable. Report the range honestly rather than cherry-picking the midpoint.

When Comps Are Scarce

The Fannie Mae Selling Guide (B4-1.3-08) requires a minimum of three comparable sales for residential appraisals. Vacant land in rural markets frequently cannot meet that threshold within a 12-month window. This is a real and common problem, not a rare edge case.

Expand the time window to 18 to 24 months when recent comps are insufficient. Fannie Mae guidance and practicing appraisers both support this approach when justified and when market-trend adjustments are applied. A comp that sold 18 months ago in a market that has appreciated 10% needs an upward time adjustment; ignoring the adjustment makes the comp stale. For broader market context, see Land Investing Market Trends 2026.

Use listed but unsold parcels as ceiling indicators. If comparable parcels are listed at $25,000 and have been sitting for six months, the market value is below $25,000. Active listings cannot substitute for closed sales in a formal appraisal, but they provide useful boundary information for investment decisions.

Use tax-assessed value as a rough sanity check, not a valuation. Assessed values can flag extreme outliers — a parcel priced at $40,000 with an assessed value of $3,000 warrants investigation — but they are unreliable as standalone indicators. Assessment ratios vary dramatically by jurisdiction: some counties assess at 100% of market value, others at 40% or less. In California, Proposition 13 caps assessed value increases at 2% annually, meaning long-held properties can be assessed at a fraction of market value. Treat assessed values as one data point among many, and understand the local assessment methodology before drawing conclusions.

A market with no recent comparable sales within a reasonable radius and time frame may be illiquid. Illiquidity means that regardless of what the parcel might theoretically be worth, finding a buyer at that price may take significantly longer than projected.

Recognize when the absence of comps is itself the answer. Time on market directly erodes investment returns.

Common Comping Mistakes

Comparing raw acreage without adjusting for usability. A five-acre parcel that is 80% wetland is not comparable to a five-acre buildable parcel. Usable acreage — the portion that can actually support construction after accounting for wetlands, floodplain, steep slopes, and required setbacks — is the relevant unit of comparison.

Ignoring the access premium. Paved-road-frontage parcels and dirt-road-access parcels serve different buyer segments at different price points. Per-acre pricing can differ by 50% or more based on access alone. Using a paved-road comp to value a dirt-road parcel without adjustment produces systematically inflated valuations.

Using asking prices as comps. What a parcel is listed for reflects what the seller wants, not what the market will pay. In land markets with limited liquidity, asking prices can sit 30% to 50% above eventual sale prices for months or years. Only closed transaction prices constitute reliable comp data.

Failing to account for market trajectory. A comp that sold 10 months ago in a market that has since declined needs a downward time adjustment. Land markets in specific micro-areas can shift meaningfully over 12 to 18 months based on infrastructure development, zoning changes, or macroeconomic conditions. Flat comparisons to stale comps in a moving market produce systematic errors.

Survivorship bias in comp selection. The natural human tendency is to unconsciously select comps that support a desired value rather than comps that represent the market. The discipline: pull all potential comps meeting the objective criteria, then filter and adjust — rather than searching for comps that confirm a price already in mind.

The Missing Variables

Several factors that materially affect land value are absent from the outline above and often absent from investor comp analyses.

Water and mineral rights can be the most valuable component of a western-state land parcel. In states where water rights are separable from surface rights, a parcel may have been sold with water rights reserved by a prior owner — meaning the current parcel lacks the water access needed for development. Mineral rights create similar complexity: if mineral rights have been severed, the surface owner may be unable to prevent mineral extraction activity. Comps should ideally compare parcels with similar rights bundles, though this information is often difficult to obtain without a title search.

Soil quality and percolation test results determine whether a parcel can support a septic system. In areas without municipal sewer access — which describes most rural land — a failed perc test makes the parcel unbuildable for residential use. This is a binary constraint that dwarfs most adjustment factors: a parcel that passes is developable; a parcel that fails is worth a fraction of comparable buildable land.

Highest and best use analysis is the foundational concept underlying all land valuation. The value of a parcel is determined not by what it currently is, but by the most profitable legally permissible, physically possible, and financially feasible use. An investor comping a parcel as a single residential lot may be undervaluing a parcel whose highest and best use is subdivision into multiple lots — or overvaluing a parcel whose constraints limit it to recreational use only. For a deeper understanding of key valuation concepts, see the Land Investing Glossary.

Comps at Scale

Manual comping at a pace of a few parcels per hour is one of the primary bottlenecks for investors screening deals at volume. The emergence of land-specific comp automation tools — Price Boss, PRYCD, Land Insights, The Land Portal, Pebble, and LotHunt among others — reflects the market's recognition of this constraint. These platforms pull candidate comps, apply basic adjustments for acreage, county, and time, and produce estimated values faster than manual research. For a detailed comparison, see Land Investing Tools Compared.

The limitation is that the adjustment step still requires human judgment for land in ways it does not for residential. Access quality, usability percentage, micro-market dynamics, and the specific rights bundle attached to a parcel are not yet reliably assessed by any automated system. The practical model is hybrid: automated systems handle the initial comp pull and first-pass adjustments, human review applies the nuanced market knowledge that determines whether a deal's margins are real. This is one area where the manual screening ceiling becomes most apparent.

Land Prowler's approach to comp verification fits this hybrid model — pulling transaction data and candidate comparables programmatically, then flagging deals where adjustment uncertainty exceeds acceptable thresholds for manual review. The full verification workflow, including comping as one step in a broader due diligence process, is outlined in The Complete Land Due Diligence Checklist.