8.2 Units of Comparison
Key Takeaways
- A unit of comparison is the price metric used to compare sales—total price, price per square foot of GLA, per room, per unit, per acre, per front foot, or other market-recognized units.
- Choose the unit that market participants use and that produces the least unexplained variance among similar sales after proper analysis.
- Residential improved property often compares on price per square foot of gross living area (GLA); land may use price per square foot, per acre, or per front foot.
- Multifamily and some commercial properties frequently use price per unit, per room, or per square foot of building or rentable area depending on how buyers underwrite deals.
- Total price is appropriate when properties are highly similar overall; unit prices help when size differs and buyers think in unit metrics—but unit rates still require adjustment for remaining differences.
Why Units of Comparison Matter
AQB ECO IV.b Units of comparison sits between selecting comps and adjusting them. Two houses can sell for $400,000 and $480,000 yet be consistent with the same market if one has 1,600 sq ft and the other 1,920 sq ft at roughly the same price per square foot. Conversely, identical total prices can hide very different unit rates when size differs.
A unit of comparison is the common denominator that lets you compare sale prices across properties of different size or configuration. It is not a fourth approach to value. It is a scaling tool inside sales comparison (and often inside land valuation and income multipliers).
Common Units of Comparison
| Unit | Typical use | Notes for exam |
|---|---|---|
| Total price | Highly similar properties; overall reconciliation | Simplest; weak when size varies a lot |
| Price per sq ft of GLA | Single-family, many residential 2–4 units (living area) | GLA rules matter (below grade often excluded from GLA) |
| Price per sq ft of gross building area (GBA) | Some multifamily, industrial, retail shells | Includes different floor types than GLA |
| Price per sq ft of net rentable / usable area | Office and some retail | Matches how leases are written |
| Price per room | Older residential analysis; some hotels | Define “room” consistently |
| Price per bedroom | Small residential / rental markets | Less common as primary unit than GLA |
| Price per unit | Multifamily, condo warehouses of units | Unit mix (studio vs 3BR) still matters |
| Price per acre / per sq ft of land | Sites, acreage, some commercial land | Frontage or buildable area may dominate |
| Price per front foot | Retail strips, waterfront, some lots | Depth and shape still affect utility |
| Price per animal unit / per hole / other | Special-purpose | Only if market participants use it |
Key idea: Use the unit the market uses. If apartment buyers quote $/unit and $/sq ft together, analyze both and explain which controls.
Gross Living Area and Residential Practice
For site-built single-family homes, the dominant improved-property unit is usually sale price ÷ GLA.
GLA (gross living area) is generally finished, above-grade, residential living space measured with a consistent standard (often ANSI-type rules in modern residential practice). Common exam-sensitive points:
- Above-grade finished area typically counts toward GLA
- Below-grade area (basement), even if finished, is often valued separately—not simply dumped into the same $/SF as above-grade GLA without analysis
- Garages, porches, unfinished attic are usually not GLA
- Measure and apply the same standard to subject and comps
| Component | Usually in GLA? | Appraisal handling |
|---|---|---|
| First-floor living | Yes | Primary $/SF base |
| Second-floor living | Yes | Primary $/SF base |
| Finished basement | Often no (not above-grade GLA) | Separate contributory value |
| Garage | No | Separate line-item adjustment |
| Enclosed porch (finished, heated, quality equal) | Sometimes disputed | Consistency + market evidence |
If Comp A is $450,000 with 1,800 sq ft GLA → $250/SF. Comp B is $420,000 with 1,680 sq ft → $250/SF. Total prices differ; unit rates agree. That is the power of the unit—before you still adjust for bath count, garage, condition, and location.
Land and Site Units
Vacant land and site valuation (also Content Area III) commonly use:
- Price per square foot — small urban lots, commercial pads
- Price per acre — suburban/rural tracts
- Price per front foot — commercial street frontage, waterfront
- Price per buildable unit — multifamily or condo land where zoning density drives value
- Price per lot — finished subdivision lots of similar utility
Choose the unit that equalizes utility. Two 1-acre parcels can have very different values if one has 50 feet of frontage and the other 200 feet on a commercial arterial. Price per acre alone misleads; front foot or buildable SF may fit better.
| Land situation | Often better unit | Why |
|---|---|---|
| Interior residential lots, similar size | Price per lot or per SF | Buyers buy a homesite |
| Deep vs shallow retail lots on same street | Price per front foot (with depth adjustment) | Frontage drives exposure |
| Multifamily-zoned acreage | Price per buildable unit or per SF of allowable GFA | Density rights create value |
| Large rural tract | Price per acre | Bulk land market |
Multifamily and Commercial Units
Multifamily: Price per unit and price per room appear frequently; price per SF of GBA is also common. A 20-unit building of mostly two-bedroom units is not the same product as 20 studios—even if unit count matches. Unit mix, rent potential, and condition still require adjustment.
Hotels: Price per room (key) is classic.
Office: Price per SF of rentable area; sometimes per SF of usable area.
Retail: Price per SF of gross leasable area; sometimes per front foot for small shops.
Industrial: Price per SF of GBA; clear height and dock count often adjusted separately because pure $/SF hides functional differences.
Choosing the Unit That Reduces Variance
A practical test: compute several candidate units across a set of similar sales and see which produces the tightest cluster (lowest dispersion) after accounting for known differences—or which matches buyer underwriting language.
Simplified illustration (residential):
| Sale | Price | GLA (SF) | Bedrooms | $/SF GLA | $/bedroom |
|---|---|---|---|---|---|
| 1 | $360,000 | 1,500 | 3 | $240 | $120,000 |
| 2 | $400,000 | 1,670 | 3 | $240 | $133,333 |
| 3 | $432,000 | 1,800 | 4 | $240 | $108,000 |
| 4 | $384,000 | 1,600 | 4 | $240 | $96,000 |
Here $/SF GLA is identical at $240 while $/bedroom scatters. The market in this toy set is speaking in living-area units, not bedrooms. Choosing $/bedroom would invent false variance and lead to noisy adjustments.
Counter-example (small multifamily):
| Sale | Price | Units | GBA SF | $/unit | $/SF |
|---|---|---|---|---|---|
| A | $1,200,000 | 8 | 6,000 | $150,000 | $200 |
| B | $1,500,000 | 10 | 9,000 | $150,000 | $167 |
| C | $900,000 | 6 | 4,200 | $150,000 | $214 |
If these buildings are similar in quality and unit mix, $/unit is stable while $/SF jumps because floor plans and efficiency differ. Buyers quoting “about $150,000 a door” point you to per-unit comparison—with remaining adjustments for condition, location, and income.
Analytical Steps to Select a Unit
- Ask how brokers and buyers quote the product.
- Calculate candidate units for the competitive sale set.
- Prefer the unit with lower unexplained variance among truly similar sales.
- Confirm the unit does not mask a critical difference (for example, $/SF that ignores whether basements are finished).
- Apply adjustments on a consistent base (adjust the unit price, or adjust total price, but do not double-count size).
Total Price vs Unit Price
| Prefer total price when… | Prefer a unit price when… |
|---|---|
| Comps are nearly identical in size and utility | Size differs materially and buyers scale price with size |
| You are reconciling final indications after size is already adjusted | You need a common metric across a size range |
| Special-purpose property with no meaningful unit market | Land or building markets are explicitly unit-based |
| Very small sample where unit math over-precision misleads | Adequate sample shows a stable unit rate |
Important: Using $/SF does not eliminate the need to adjust for quality, baths, garages, or location. It primarily addresses size so you are not comparing raw totals of a 1,400 sq ft house to a 2,200 sq ft house as if size were irrelevant.
Double-counting trap: If you convert all sales to $/SF of GLA and then also apply a large “size adjustment” for GLA differences, you may adjust for size twice. Typical residential grids either:
- Adjust total prices for GLA differences using a supported $/SF contribution rate, or
- Analyze on a unit basis carefully—not both without a coherent method.
Worked Residential Example
Subject: 1,720 sq ft GLA, 3 bed / 2 bath, 2-car garage, average condition.
| Comp | Sale price | GLA | Unadjusted $/SF |
|---|---|---|---|
| 1 | $430,000 | 1,720 | $250.00 |
| 2 | $451,000 | 1,820 | $247.80 |
| 3 | $410,000 | 1,650 | $248.48 |
Unadjusted unit rates cluster near $248–$250/SF. Differences remaining might be garage, bath, or condition—not gross size chaos. If Comp 2 had sold for $520,000 ($285/SF) with no quality or location superiority, you would verify the sale (concessions, personal property, non-arm’s-length) before treating $285 as market.
Indicated value sketch using subject GLA × reconciled unit rate: 1,720 × $249 ≈ $428,000, then refine with line-item adjustments rather than stopping at the multiplication alone.
Worked Land Example
Three commercial pad sales on the same arterial:
| Sale | Price | Acres | Front feet | $/acre | $/FF |
|---|---|---|---|---|---|
| 1 | $600,000 | 1.00 | 150 | $600,000 | $4,000 |
| 2 | $800,000 | 1.60 | 200 | $500,000 | $4,000 |
| 3 | $480,000 | 0.90 | 120 | $533,333 | $4,000 |
$/FF is constant; $/acre is not, because depth/acreage varies. For a subject with 180 front feet and deep residual land, starting from $4,000 per front foot (then adjusting depth/utility) is more defensible than averaging inconsistent per-acre figures.
Exam Traps for Units of Comparison
- Mixing GLA definitions across subject and comps (one includes basement, one does not).
- Assuming $/SF always beats total price even when all comps are twins.
- Using per-acre for frontage-driven retail land (or the reverse).
- Treating unit price as final value without property adjustments.
- Double-counting size through both unit conversion and a full size line item.
- Applying residential GLA logic unchanged to industrial GBA markets.
- Ignoring that personal property or franchise value inflated price before you compute $/SF of real estate.
Link Forward
Units of comparison put sales on a common scale. They do not explain why one sale differs from another after size is normalized. That is the job of elements of comparison—transactional and property characteristics—covered next. After elements are identified, adjustment sequence converts each comparable’s price into an indication for the subject.
Four similar single-family sales show tightly clustered prices per square foot of GLA but widely scattered prices per bedroom. What is the best conclusion about the primary unit of comparison?
An appraiser converts all comparable sales to price per square foot of GLA and also applies a large separate GLA size adjustment on the grid for the same size differences. What error is most likely?