How Valuation Actually Works When You Get Past the Textbook
I've been cranking out valuations for longer than I care to count, across everything from suburban single-family homes to rural agricultural parcels and a handful of mid-rise multifamily properties. The standard materials tell you about three approaches: sales comparison, income capitalization, and cost. What they don't tell you is how much time you'll actually spend second-guessing each one and why the final number rarely matches what anyone wants to hear. Here's how I start. I find comps first, then immediately ask myself which ones are actually useful and which ones I'll have to discard after spending twenty minutes on adjustments. The sales comparison approach sounds straightforward until you're looking at a property that sold six months ago in a different school district and you're trying to figure out what kind of adjustment dollar-per-square-foot makes sense. Nobody writes a rulebook for that. The income approach is where most residential valuers get uncomfortable. It works perfectly fine for multifamily, small commercial, and even some single-family rentals when you have solid rent rolls and vacancy data. Cap rates are the number everyone argues about. Get them wrong by a quarter point and you shift value by tens of thousands of dollars on a typical income property. I learned this the hard way on a small fourplex in Kansas City where the cap rate in my comp set was 7.2 percent but the actual market was transacting at 8.4 percent because lender appetite had dried up in that zip code. The difference between those two rates changed the property's value by about forty thousand dollars.
The cost approach is useful for new construction and unique properties with no comparable sales, but it tends to overvalue older buildings. Depreciation estimates are rough at best. I use it sparingly and almost never as the primary value indicator unless I'm valuing something like a church or a custom-built warehouse where there genuinely are no comps anywhere nearby.
A Problem That Broke My Standard Process
About three years ago I was valuing a small strip center in Ohio — eight tenants, mostly local businesses, built in 1978. The sales comparison approach was useless because there had been exactly one comparable sale of a small retail center in the entire county in the previous three years. The cost approach inflated the value by roughly twenty percent because the land was worth more than the building ever could justify. The income approach was the only one that came close to making sense, but the lease terms were messy. Some tenants were on gross leases, some on net, and one had a renegotiation clause that was about to trigger. My workaround was to build a custom database from scratch. I pulled all small retail center sales across three surrounding states over a five-year period, segmented by occupancy type and tenant quality, and then applied income-based valuations to each comparable transaction to derive implied cap rates. I ended up with twelve data points instead of the usual three or four, and the resulting cap rate range was tighter than I expected. The final value landed about five percent below the owner's expectation, which was frustrating for them but more defensible in underwriting than any number I could have gotten from a single comp set.
Get the Full Details
Things Nobody Tells You About AVMs and Automated Tools
Automated valuation models exist now and they're decent for quick estimates. Zillow's Zestimate, Redfin's estimate, and various proprietary tools used by lenders can give you a ballpark in seconds. The problem is that AVMs rely heavily on last sale price, square footage, and basic location data. They don't understand interior condition, renovation quality, or whether the HVAC system was replaced last year. I've seen AVMs come in fifteen percent low on a fully gut-renovated 1920s bungalow in Portland because the algorithm couldn't differentiate between a preserved original and a completely rebuilt interior. The correction involved pulling recent sales of other renovated homes in the same neighborhood and adjusting the value per square foot manually. For most professionals, the practical use of AVMs is as a starting point, not an answer. Run it first to see where the algorithm lands, then spend the actual time going through public records, recent closings, and direct market contact to validate or correct the number.
When the Data Doesn't Exist
This happens more often than you'd think. Rural land parcels, specialty properties, newly zoned areas, and distressed sales all suffer from thin data. When I encounter these situations, I stop trying to force a traditional comps analysis and shift to a benchmark-based approach instead. I look at what similar properties transacted for in adjacent markets, adjust for location differentials using price-per-acre or price-per-unit metrics, and then apply a market-condition adjustment based on local absorption rates and days on market. I did this for a ninety-acre agricultural parcel in central Illinois where there had been no comparable sales in five years. I pulled twelve transactions from neighboring counties, adjusted for soil quality differences using USDA soil survey data, and factored in a location discount based on distance to major grain elevators and highways. The resulting valuation was defensible even though no one could point to a directly comparable sale.
Common Mistakes I See Regularly
People cherry-pick comps that support a preconceived value. It's tempting when you already know what you want the number to be. Another mistake is using list prices instead of closed prices. List prices tell you what sellers hope for. Closed prices tell you what actually happened. The gap between those two numbers in a shifting market can be significant. Cap rate selection is another frequent error. People grab a cap rate from an online article or a national report and apply it directly to a local property without adjusting for local conditions, property condition, lease structure, or tenant credit quality. A Class A multifamily property in a growing Sun Belt market and a Class C property in a rust belt city might both show a cap rate of 6.5 percent in a national summary, but the risk profile is entirely different. The actual transacted cap rates for those properties could differ by a full percentage point or more.
What I Actually Do Before Writing a Final Number
I cross-check my value against at least two independent data sources. If I'm using recent comp sales, I'll also run an income-based check. If I'm leaning on the income approach, I'll verify that the derived value aligns with what similar properties sold for on a per-unit or per-square-foot basis. If all three approaches converge within a narrow range, I'm confident. If they diverge widely, I dig deeper into why before finalizing anything. I also verify the comparables myself whenever possible. Public records are useful but frequently contain errors — incorrect square footage, wrong year built, missing additions or renovations. A quick phone call to a local agent or a site visit can surface discrepancies that would otherwise slip through unnoticed. This adds time to the process but prevents embarrassing corrections later when a buyer's appraiser finds the same issues.
The Bottom Line on Limitations
Valuation is not a precise science. It's an estimate based on available market evidence, and the quality of that estimate depends entirely on the quality and relevance of the data you feed into it. When the market is stable and transactions are frequent, valuations are relatively reliable. When markets are volatile or data is scarce, any number you produce should come with a wider range of uncertainty attached to it. The tools and methods described here — sales comparison, income capitalization, cost approach, AVMs as a starting point, custom databases for thin markets, and cross-checking across methods — are the standard framework. They work well within their limits and fail noticeably outside them. Knowing where those limits are is more important than memorizing the formulas.
