How to Read a Comp, and Why Price Per Square Foot Lies

Every pricing conversation eventually arrives at the same sentence: the house down the street sold for this, so ours is worth that. It is a reasonable instinct and it is how appraisers work too. The difference is that an appraiser does not stop at the sale price — the sale price is where the work starts.

A comparable sale is evidence about your house only to the extent that the two properties are genuinely alike, and where they are not alike, the difference has to be priced. That pricing is the entire discipline, and it is what separates a valuation from an anecdote.

What actually makes a sale comparable

Four things, roughly in order of how much they matter:

Location. Not the town — the location. Same school district, same neighbourhood, ideally the same side of the same busy road. A sale half a mile away across a district boundary or a highway is a different market, not a close one.
Recency. The market moves. A sale from fourteen months ago tells you about a market that no longer exists, and the further back you go the more of the difference is time rather than property.
Style and size. A colonial compares to a colonial. Gross living area within roughly ten or fifteen per cent is workable; well outside that and the adjustment starts doing more work than the comparison.
Condition and age. A renovated 1958 ranch and an original 1958 ranch are different products at the same address.

What does not make a sale comparable: that it is the nearest recent sale, that it is the highest recent sale, or that it is the one that supports the number somebody already had in mind.

The adjustment grid

An appraiser lines up three to six closed sales beside the subject property and adjusts each one toward it, line by line. If the comp has a garage and yours does not, the comp's price comes down by what a garage contributes on that street. If yours has a second full bathroom and the comp does not, the comp's price goes up.

The direction confuses everyone the first time, so it is worth stating plainly: you adjust the comp, never the subject. The comp is the known quantity; you are asking what it would have sold for if it had been your house.

Two things follow from that, and both matter to a seller:

The adjustment is what a feature contributes to value on that street, not what it costs to build. A $90,000 kitchen does not generate a $90,000 adjustment. See which renovations pay back for why that gap is normal rather than unfair.
A grid full of large adjustments is a grid telling you these are not good comps. When every line needs a big correction, the honest conclusion is to find better sales, not to trust the arithmetic.

Why price per square foot lies

Price per square foot is popular because it is a single number you can compute in your head, and it is misleading for exactly that reason. It assumes every square foot in a house is worth the same, and none of them are.

Consider two houses that both sell for $900,000 and both measure 1,900 square feet. One is a colonial with four bedrooms upstairs and two and a half baths. The other is a raised ranch where 600 of those feet are a finished lower level with low ceilings and small windows. Same $474 per square foot on paper, two completely different products, and the next buyer will not treat them as interchangeable.

Three specific failures to watch for:

Smaller houses carry a higher price per square foot than larger ones, almost everywhere, because the kitchen, the heating system, the roof and the land come with the house regardless of size. Applying the small house's rate to a big one overshoots badly, and applying the big house's rate to a small one undershoots.
Finished basements and converted attics are counted inconsistently. Some listings include below-grade space in the square footage and some do not. If one of your two numbers counts it and the other does not, the ratio is meaningless.
Land is not in the number at all. A half-acre and a quarter-acre house of identical size produce identical price per square foot and are not identically valuable. What a lot is actually worth goes through that separately.

Use price per square foot as a sanity check on a tight set of genuinely similar houses. Do not use it to value a house.

The square footage is probably wrong anyway

This one surprises people. The square footage in a listing usually comes from the town's assessor record, and assessor records are built by mass appraisal — a system designed to value thousands of properties consistently for taxation, not to measure one of them precisely.

An addition never permitted, a three-season porch counted as living area or not counted, a finished basement recorded decades ago, a field card last updated by someone standing in the street: all common. If the number is load-bearing in your pricing, measure the house or have it measured. The same field card is worth reading for other reasons — see how Massachusetts property tax works for what is on it and how to correct it.

Closed, pending, active, expired — four kinds of evidence

They are not interchangeable and each answers a different question.

1.Closed sales are the only ones where money actually changed hands at an agreed number, which is why they are the foundation and the only thing an appraiser can put in the grid.
2.Pending sales are the most current signal available. The price is not public until recording and the terms are unknown, but a cluster of quick pendings tells you the direction of travel before the closed data does.
3.Active listings are your competition, not your evidence. An asking price is an opinion held by one seller. A buyer choosing between your house and that one is comparing them directly, which is why they belong in the analysis even though they prove nothing.
4.Expired and withdrawn listings are the most useful data nobody looks at. They are the market's record of prices that were tested and refused, and they draw the ceiling more reliably than the sales do.

How far back, and how far away

The honest answer is: as tight as the data allows, and then you widen only as far as you must, with your eyes open about what widening costs. Six months and a mile is a common starting frame in a suburban Massachusetts market with normal turnover. On an unusual house, an unusual street, or in a town where twelve houses sell a year, you will be reaching further and the answer gets wider with it.

A valuation with a range attached is not a weaker valuation. It is a more honest one, and the range itself is information — a wide one means the market has not spoken clearly about houses like yours, which changes both how you price and how you read the first two weeks.

What no comp can tell you

That the buyers who paid that number are still in the market. That the sale was arm's length rather than between family. That the seller was not in a hurry, or in a divorce, or holding a second offer. That nothing was credited back at closing that never showed up in the recorded price.

Comps describe the market that existed when those houses sold. Your house sells into the market that exists now, and the gap between the two is exactly what the first two weeks of showings is for measuring.

Questions to ask about any comp you are shown

1.When did it actually close, and when did it go under agreement? The second date is when the price was agreed.
2.Is it the same school district, the same side of the main road, the same walk to the station?
3.What condition was it in — and is there photographic evidence, or an assumption?
4.Does the square footage include below-grade space, and does mine?
5.What did it ask, and how long did it take? A quick sale at asking and a long grind to the same number are different facts.
6.Which sales were excluded from this analysis, and why?

That last one is the question that gets asked least and matters most. A comp set is a set of choices, and the choices are where a valuation is made or lost.