Market Analysis
How to Build a Comparative Market Analysis
Search windows, acceptable distance, data availability and what constitutes an arm's-length sale all depend on the market you are working in and on the rules governing the data source you use. Nothing below prescribes a threshold; set yours from local evidence and record why you chose it.
The method behind a comparative market analysis is the sales comparison approach, scaled down. Find properties a buyer would have considered interchangeable with this one, note what each has that this one does not, and correct their prices for those differences until they are answering the same question.
Done carefully the result is a defensible range with reasoning attached to every step. Done carelessly it is a number pulled from an average, and averages of unlike properties describe nothing.
The Procedure, In Order
Run the steps in sequence. Each one constrains the next, and skipping forward is where bad conclusions get introduced.
Describe the subject first. Measure or verify living area, confirm bedroom and bath count, note lot characteristics, record condition and any updates. An analysis built on a public record that is wrong about the square footage is wrong everywhere.
Define the competitive market. Draw the area a buyer shopping for this property would actually search: the neighborhood, the attendance zone, the side of the arterial, the subdivision with its own amenity package. Political boundaries matter only where buyers act on them.
Pull closed sales from that area, recent enough that conditions have not shifted underneath them. Prefer a short window with genuinely similar property over a long window with convenient property.
Screen out sales that were not arm's length. Transfers between relatives, estate liquidations under time pressure, foreclosure sales and transfers among affiliated entities reflect the circumstances of the parties, not the demand of the market.
Adjust each remaining comparable to the subject, feature by feature, and write down the reasoning for every adjustment you make.
Layer in current competition and recent failures to sell. These do not change the adjusted figures; they tell you where inside the supported range the property should be positioned.
Reconcile to a range and a recommendation. Give the most weight to the comparables that needed the fewest adjustments, and say plainly which ones those were.
Choosing Comparables That Actually Compare
Similarity is judged by what drove the buyer's decision, not by what is easy to tabulate. Location leads, because a buyer who wants this street cannot substitute another one. Then physical type and size, then age and construction, then condition, then the amenities that carry weight in this specific market.
Proximity is a proxy for similarity, not a substitute for it. A sale across the street is the wrong comparable if it sits in a different attendance zone, a different flood designation, or a different association with different dues and rules. A sale farther out in the same development is usually the better choice.
Recency works the same way. A sale is useful while the conditions that produced it still hold. In a market moving quickly, older sales need a market-conditions correction or they should be dropped; in a stable market with thin inventory, reaching back further is defensible as long as the analysis says so.
When the property is unusual, acreage, a converted structure, a waterfront parcel, genuine comparables may not exist inside any reasonable window. The professional answer is to widen the search deliberately, state the limitation, and lean harder on the competing inventory, rather than to force weak comparables into a table and present the result as precise.
Adjusting Without Losing the Thread
The direction rule is the part candidates invert under time pressure, so memorize the logic rather than the phrasing. You are asking what the comparable would have sold for if it had been the subject. If the comparable is superior in some respect, it sold for more than the subject would have, so you subtract. If it is inferior, you add. Every adjustment lands on the comparable. The subject is never adjusted, because the subject is what you are solving for.
Adjustment amounts come from market evidence, not from a fee schedule. The defensible sources are paired sales, where two otherwise similar properties differ in one feature and the price gap suggests what that feature contributes, and the observable behavior of buyers in this market segment. Contribution is also not cost: an improvement can cost far more than it returns, and over-improvement relative to the neighborhood routinely returns very little.
Restraint improves the result. A comparable requiring a long list of corrections was never very comparable, and each correction adds error. If the adjusted figures scatter widely, the problem is upstream in the selection, not downstream in the arithmetic.
Then reconcile rather than average. Weight the comparables by how little they needed to be moved, look at where the adjusted figures cluster, and state a range. A single number implies a precision the method does not have, and it invites an argument the evidence cannot win.
Errors That Show Up Repeatedly
Most weak analyses fail for one of these reasons rather than from a calculation mistake.
Averaging the comparables. An average treats a well-matched sale and a poorly matched one as equally informative, which is exactly what the reconciliation step exists to prevent.
Adjusting the subject instead of the comparable. This reverses every conclusion and is the single most common way an otherwise competent analysis goes wrong.
Using asking prices as evidence of value. An unsold listing records what a seller wanted. Only a closed sale records what a buyer agreed to.
Ignoring condition because the data feed does not carry it. Two properties identical on paper can differ substantially in value based on systems, finishes and deferred maintenance.
Letting distressed sales in unscreened. Where they dominate a segment they may in fact be the market, but that is a conclusion to state explicitly, not an assumption to make silently.
Pricing to the seller's expectation. Producing a supportable range and then recommending above it turns the analysis into decoration and generally ends in a price reduction later.
Describing the neighborhood by who lives there. Confine commentary to market characteristics. Any characterization touching a protected class is a fair housing violation regardless of intent.
Working Questions About Building a CMA
How many comparables are enough?
Enough that the adjusted figures corroborate one another, and few enough that each one genuinely resembles the subject. One close match plus two supporting sales is stronger than a long table of distant properties. If the adjusted results refuse to cluster, treat that as a signal to reselect rather than to add more.
Can the analysis be done without visiting the property?
It can be started. A preliminary version built from records and listing data is a reasonable first pass for an initial conversation. The final version needs the walk-through, because condition, updates, layout problems and functional obsolescence are the factors least likely to be recorded anywhere and most likely to move the number.
What if the market has moved since the comparable sold?
Apply a market-conditions adjustment before any physical adjustments, and support it with evidence of direction and pace rather than with a feeling. If the evidence for the trend is thin, it is more honest to prefer recent sales and say the older data was set aside.
How is a homeowner's renovation handled?
By what it contributes in this market, which is frequently less than it cost. Improvements beyond the standard of the surrounding properties tend to return a fraction of their expense, and some, such as a pool or a highly specific layout, appeal to a narrower buyer pool. Support the figure with paired sales where you can find them.
What does the seller need to see?
The comparables, the adjustments, the reasoning and the range, in that order, plus the competing inventory and the recent failures. A seller who understands how the range was produced will discuss positioning inside it. A seller handed only a conclusion will argue about the conclusion.
How does this get tested?
Usually as a direction-of-adjustment item or a comparable-selection item, occasionally as a terminology item about what a licensee may call the result. RE License Prep drills all three patterns in the national bank, tracks which one is costing you points through your PassPulse Score, and orders the remaining work in your PassMap™. The diagnostic is free; full access is $79 one time with a 30-day money-back guarantee.