How to Run a CMA That Actually Holds Up at the Kitchen Table

The Kitchen Table Test

A CMA lives or dies at the kitchen table. It is the moment when a seller looks at your number, looks at the Zestimate on their phone, and decides whether to trust your expertise or the algorithm.

Most CMAs fail this test. Not because the number is wrong, but because the methodology is invisible. The seller sees a list of comps and a suggested price. They do not see how you got there. They do not see why you chose those comps over others. They do not see the adjustments, the reasoning, or the confidence level.

A CMA that holds up is one the seller can follow, question, and ultimately trust because the logic is transparent.

Step 1: Start with the Subject, Not the Comps

The most common CMA mistake is starting with comparables. Agents pull comps first, then back into a price. This inverts the process and invites anchoring bias.

Start with the subject property. Assess its condition. Rate it honestly: excellent, good, average, fair, or below. Note specific features that add or subtract value: updated systems, deferred maintenance, lot characteristics, noise exposure, view quality.

This assessment becomes the lens through which you evaluate every comp. Without it, you are comparing properties on paper characteristics alone, which is exactly what AVMs do and exactly why they hit the 7% error wall.

Step 2: Select Comps with Discipline

Three to six comparables is the right range. More than six dilutes the analysis. Fewer than three lacks statistical support.

Selection criteria, in order of priority:

Proximity. Same neighborhood first. Same school zone second. Same city is a last resort and should be flagged as such.

Recency. Sales within 90 days carry the most weight. Sales within 180 days are acceptable with market adjustment. Anything older requires explicit justification.

Similarity. Match on the characteristics that drive value in your market. In some areas, lot size dominates. In others, bedroom count or school proximity matters more. Know which variables your market prices most heavily.

Arms-length transactions. Exclude foreclosures, estate sales, and family transfers unless the market is dominated by distressed sales, in which case they are the market.

Step 3: Adjust with Precision

Raw comp prices are not the answer. Adjustments are. And adjustments are where most CMAs fall apart.

Common adjustment categories:

Condition. The biggest single adjustment in most CMAs. A comp that sold in excellent condition versus your subject in good condition requires a downward adjustment to the comp's price. Many agents skip this or eyeball it. The best agents quantify it.

Size. Price per square foot is a useful reference but not a direct adjustment factor. A 3,000 square foot home does not sell for exactly 50% more than a 2,000 square foot home. Marginal value per square foot declines as homes get larger.

Lot characteristics. Corner lots, cul-de-sacs, busy street adjacency, and grade all affect value. These adjustments are hyperlocal and require market knowledge that no algorithm can replicate.

Market conditions. If the market has appreciated 4% since a comp closed six months ago, that comp needs a time adjustment. This is where many CMAs go wrong: they use stale comps without accounting for price movement.

Step 4: Show Your Confidence, Honestly

Every CMA should include a confidence assessment. Not just a single number, but a range and an explanation.

"Based on six recent comparable sales with adjustments, I estimate this home's market value between $865,000 and $895,000. My recommended list price is $885,000. My confidence in this range is high because three of the six comps are within a quarter mile and closed within 60 days."

This does three things. It demonstrates analytical rigor. It sets realistic expectations. And it gives the seller a framework for understanding the inevitable gap between your estimate and their Zestimate.

Step 5: Address the Algorithm in the Room

Every seller has checked their Zestimate before you arrive. Do not ignore it. Address it directly.

"The Zestimate on your home is $905,000. That estimate is based on public data without any assessment of your home's condition, the quality of your renovations, or the specific characteristics of your lot. Automated valuations carry a median error of around 7% for off-market homes. On your home, that means the algorithm could be off by roughly $63,000 in either direction. My CMA is based on comparable sales I have personally reviewed, with adjustments for the specific characteristics of your home."

This is not a sales pitch. It is an honest explanation of the difference between algorithmic estimates and professional analysis. The sellers who value accuracy will trust your number. The ones who want to hear the highest number will list with someone else, and you are better off without that listing.

The Koqi Approach

Koqi's CMA tool automates the mechanical parts of this process (comp selection, appreciation calculations, market trends) while requiring the human inputs that matter (condition rating, market positioning). The result is a CMA that is both data-rich and judgment-informed.

Every CMA you run feeds your ACCS score. When the property eventually sells, Koqi compares your estimate to the actual closing price. Over time, you build a verified accuracy record that turns your next kitchen table presentation from "trust me" into "here is my track record."