What should I price my Bay Area home at?
Start with recent sales of homes a buyer would compare with yours in the same local market. Account for differences in condition, layout, size and location to establish a likely selling range, then choose an asking price based on current competition and how buyers search. Your estimated market value is a supported range; your asking price is the number you use to bring the home to market.
- Find the right sales. Compare the same property type in your neighbourhood and check sale dates, finished area and property details.
- Work out the differences. Review what makes each sale a better or worse match, then use market evidence to support adjustments.
- Choose a listing strategy. Use the supported range, today's competing listings and your selling timeline to decide the asking price.
Key takeaways
- A comp is an argument, not a price. Three honest agents can read the same six sales and land three different numbers, and all three can be defensible.
- Price per square foot is the most quoted and least reliable number in residential real estate. It flattens lot, condition, layout and light into one figure that hides all four.
- Buyers do not browse by your price. They browse in bands, so $1,525,000 is invisible to everyone whose filter stops at $1.5M — the most expensive twenty-five thousand dollars you can ask for.
- Your lender's appraiser will re-run this exercise after you are already in contract. Pricing that ignores what an appraiser can defend is how a sale falls apart in week four.
Every seller is handed a number early. It arrives in a listing presentation, or from a website that has never seen the inside of the house, and it is presented as though it were a measurement — as if the home has a price the way it has a square footage, and a professional simply reads it off.
It does not work that way. A price is a prediction about what a specific buyer will do on a specific weekend, supported by evidence about what similar buyers already did. The evidence is the comparable sales — the comps — and learning to read them is the difference between setting a price and being handed one.
This is the long version. If you want the short one: the comps tell you the range, your house's condition tells you where in the range you sit, and the way buyers search tells you what number to actually type into the listing. Everything below is the detail underneath those three sentences.
What a comp actually is
A comparable sale is a home similar enough to yours, close enough to you, and sold recently enough that a reasonable buyer would have considered it an alternative to your house. That last clause is the whole test. If a buyer touring your home would not have also toured that one, it is not a comp, however similar the spreadsheet says it is.
Three categories get confused constantly, and they answer different questions:
- Closed sales are the only real evidence. Money changed hands, an appraiser signed off, a lender funded it. This is what your price has to be defensible against.
- Pending sales tell you where the market moved after the closed sales were negotiated. They are the most current signal you have, and the least verifiable — you can see the list price, rarely the contract price.
- Active listings are not evidence of anything except what other sellers hope. They are your competition, not your comps. A neighbour asking a fantasy number does not raise your value; it just gives buyers something to compare you favourably against.
An appraiser will typically build a valuation from three to six closed sales. Not thirty. The exercise is not averaging a large sample — it is finding the few sales that genuinely bracket your home, then reasoning about the differences.
The four filters that make a sale comparable
Every comp gets tested against four things. A sale that fails one badly is usually not rescuable by being strong on the other three.
| Filter | The usual standard | Where it breaks in the Bay Area |
|---|---|---|
| Proximity | Same neighbourhood, ideally within half a mile | A ridge, an arterial road or a school boundary can make half a mile meaningless |
| Recency | Sold within three to six months | In a fast-moving quarter, a five-month-old sale is describing a different market |
| Similarity | Comparable size, beds, baths, lot and age | Bedroom count matters more than square footage to most families searching |
| Condition | Similar level of finish and system age | The single largest source of disagreement, and the hardest to see in listing photos |
A real San Francisco comps example: Parkside, early 2026
These three sales in San Francisco's broader Parkside market show why matching bedrooms and square footage is only the start. All were marketed as three-bedroom, two-bathroom single-family homes, yet their reported sale prices spanned $350,000.
| Home | Reported sale date | Sale price | Listed size |
|---|---|---|---|
| 2032 Wawona Street | Jan 6, 2026 | $1,850,000 | 1,475 sq. ft. |
| 176 Escolta Way | Feb 9, 2026 | $2,050,000 | 1,585 sq. ft. |
| 2191 33rd Avenue | Feb 23, 2026 | $1,700,000 | 1,500 sq. ft. |
The unadjusted historical range is $1.70 million–$2.05 million. Before using it to price another home, investigate the differences. Escolta's listing describes a renovated kitchen, updated systems and a position across from Pine Lake Park. The 33rd Avenue listing places its third bedroom and second bathroom on the lower level. Those details deserve review before treating the sales as equally comparable.
For your home, verify condition, layout, exact location, finished area and market timing before narrowing the range. These are candidates for comparison, not interchangeable homes. This historical example is not a current valuation of your property; today's listing decision needs evidence from today's market.
Why price per square foot misleads
Price per square foot is quoted more than any other number in residential real estate, and it is close to useless as a pricing tool. It is an output of a sale, not an input to one.
The arithmetic hides the reasoning. Divide a sale price by finished area and you have compressed lot size, condition, layout, natural light, garage, outdoor space, street noise and view into a single figure — then you multiply that figure by your own square footage and call the result a valuation. Every one of those variables has been silently assumed equal.
It also behaves counterintuitively with size. Smaller homes almost always carry a higher price per square foot than larger ones in the same neighbourhood, because the land, the kitchen and the bathrooms — the expensive parts — are spread across less area. Applying a small home's rate to a large home overprices it, reliably.
Use it as a sanity check at the end. If your price implies a rate wildly outside what the street has produced, something in your reasoning needs re-examining. That is the whole of its legitimate job.
The adjustments an appraiser makes
Once the comps are chosen, the work is adjusting them. The comp is never identical, so its sale price is modified up or down to answer one question: what would that house have sold for if it were yours?
The direction confuses people, so it is worth stating plainly. If the comp is better than your home, its price is adjusted down to reach your value. If it is worse, adjusted up. You are not adjusting your house — you are asking what that sale would have looked like with your house's attributes.
The defensible way to size an adjustment is paired sales: find two sales that differ mainly in the one feature, and let the difference between them price it. That is why a well-supported valuation is narrow and a hand-waved one is wide.
| Usually moves the number | Usually moves it less than owners expect |
|---|---|
| An additional bedroom, or a bathroom on the main floor | A kitchen remodel that has already been superseded by taste |
| Usable flat lot and genuine outdoor space | Solar owned outright — it helps, but rarely dollar for dollar |
| Roof, foundation, sewer lateral and electrical panel age | Pools, which are a positive for some buyers and a cost for others |
| Garage and off-street parking where the street has none | High-end appliances a buyer did not choose themselves |
| Which school attendance area the parcel falls in | Fresh paint and staging, which change speed more than price |
Why two blocks change the number
The Bay Area punishes generalisation harder than almost any market in the country. "Comparable neighbourhood" is doing an enormous amount of work in a region where the boundary between two of them can be a single street.
Four boundaries move prices sharply and are invisible on a map of the city:
- School attendance areas. Two houses that share a fence can feed different elementary schools. For a large share of buyers this is the first filter applied and the last one compromised on.
- Hills and flats. Elevation buys light, view and quiet, and costs you level yard, driveway grade and sometimes insurance.
- Transit and commute geometry. Walking distance to a station is a different product from a ten-minute drive to the same station, and prices accordingly.
- Arterial roads and rail. A house on the through-road and one two streets in are not comps for each other, no matter how alike they look.
This is why an automated estimate from a national site drifts so far here. Those models are trained to be roughly right across millions of homes; they are structurally unable to know that your side of the street is in the other attendance area. A local advisor is not being sentimental about that — it is the part of the job a model cannot do from parcel data alone.
Pricing into the way buyers search
The comps give you a supportable range. Choosing the number inside that range is a separate skill, and it has less to do with valuation than with how search filters work.
Buyers do not enter your price. They enter a ceiling, and the ceiling is almost always round. A home at $1,525,000 does not appear for anyone who set their maximum at $1.5M — and that group is far larger than the group who set it at $1.55M. You have made twenty-five thousand dollars invisible to your best-matched audience.
That produces a rule with real teeth: price at the top of a band, not just over it. Sitting just below a round number puts you in front of two audiences at once — everyone shopping under it, and everyone shopping above it who now sees you as the affordable option.
The cost of the opposite move is easy to underestimate. A listing priced above what the comps support does not simply sit and wait for the right buyer. It ages. Days on market accumulate in public, the price cut becomes a story buyers can read, and the eventual offers arrive lower than they would have on day three — because a listing that has been available for six weeks is answering a question nobody had to ask about a new one.
What the first ten days actually tell you
The first week and a half is the only free feedback the market ever gives you. It is worth knowing how to read it, because each signal points at a different problem and the fixes are not interchangeable.
| What you see | What it usually means | What to change |
|---|---|---|
| Strong online views, few showings | The photos or the price are contradicting each other | Photography and the first three images, before the price |
| Showings, no second visits | The home is showing worse in person than online | Condition, smell, light and clutter |
| Second visits, no offers | Buyers like it and think they can wait | Price, or a deadline that creates a reason to act |
| Almost no views at all | You are outside the search band your buyers use | The price, immediately |
The mistake is waiting a month to act on a signal that was legible in eight days. Attention is highest at launch and never returns to that level; a correction in week two reaches an audience that a correction in week six has already lost.
The appraisal gap, and why your price has to survive one
Accepting an offer is not the end of the valuation argument. If the buyer is financing, their lender sends an appraiser to do the exercise in this article independently — and the lender will lend against the appraised value, not against what your buyer agreed to pay.
When the appraisal lands under the contract price, that difference is the appraisal gap, and somebody has to close it. The buyer brings additional cash, the price is renegotiated, the appraisal is challenged with better comps, or the deal falls apart. All four happen.
This is the practical reason not to accept the highest number in the room without asking what supports it. An offer that cannot appraise is a delay dressed as a win — and by the time it unravels you have lost the weeks when your listing was new.
Disclosure sits in the same territory. California requires sellers of most one-to-four unit homes to give buyers a Transfer Disclosure Statement, and known material facts affecting value belong in it. Beyond the legal obligation there is a pricing one: problems found by a buyer's inspector after they are emotionally committed get renegotiated at a worse number than problems disclosed up front and priced in.
How Stelo prices a home
We do this the way it is described above, which is to say without shortcuts, and we do not charge a percentage for it. A licensed local advisor pulls the closed sales, walks the house, makes the adjustments and tells you the number they can defend — including when it is lower than the one you were hoping for.
What is different is the fee. A traditional listing takes a percentage of the sale, so the cost of selling a more expensive home rises with the price while the work stays roughly the same. With Stelo, choose a flat $2,500 paid upfront for our guided selling plan, or 1% of the sale price due at close for full agent representation.
The pricing conversation is the first thing that happens, before any commitment. You can see the rest of what happens after it — preparation, listing and marketing, showings, and closing — on the walkthrough of how a Stelo sale runs.
Sources and further reading
The Parkside example links to the reported sales and property descriptions above. For the appraisal framework behind choosing and adjusting comparable sales, see:
- Fannie Mae: Comparable sales — selecting comparable properties and assessing their relevance.
- Fannie Mae: Adjustments to comparable sales — supporting adjustments with market evidence.
Frequently asked questions
How many comps do you need to price a house?
Three to six closed sales is the working standard, and it is what an appraiser will typically build a valuation from. The number matters far less than the quality: three genuinely comparable sales on your street in the last two months are worth more than twenty loosely similar ones across the city. Pending sales are useful as a check on where the market has moved since those closings, and active listings are competition rather than evidence — they show what other sellers hope for, not what any buyer has agreed to pay.
How old can a comparable sale be?
Three to six months is the usual window, and the faster the market is moving the tighter that window has to be. In a quarter where prices are shifting, a five-month-old sale is describing conditions that no longer exist, and using it without a market-conditions adjustment will mislead you in whichever direction the market has travelled. If there is genuinely nothing recent nearby, an older sale can still be used — but the adjustment for time has to be explicit and defensible rather than assumed away.
Is Zillow's estimate accurate for Bay Area homes?
Automated estimates are built to be roughly right across millions of homes, which is a different goal from being right about yours. They work from parcel and transaction data, so they cannot see condition, light, layout or noise, and they systematically miss the boundaries that matter most here — school attendance areas, hillside versus flat, and which side of an arterial road you sit on. Treat one as a starting range and never as a listing price, particularly in neighbourhoods where two adjacent streets sell differently.
Should I price my home high and negotiate down?
It is the most common instinct and it usually costs money. Attention is at its highest in the first week and never returns to that level, so an overpriced launch spends the best audience it will ever get. What follows is visible: days on market accumulate publicly, a price reduction becomes part of the listing's story, and offers arrive lower than they would have at the start because buyers now assume something is wrong. Pricing to sit just under a round search band typically produces more competition, not less.
What happens if the appraisal comes in below the offer?
The lender lends against appraised value rather than the contract price, so the difference — the appraisal gap — has to be closed by someone. In practice there are four outcomes: the buyer brings extra cash, the price is renegotiated, the appraisal is challenged with better comparable sales, or the deal collapses. This is why the highest offer is not automatically the best one, and why it is worth asking what comps support a number before accepting it.
Does staging change what my home is worth?
Staging generally changes how fast a home sells and how many buyers compete for it, rather than moving the underlying value the comps support. That is still worth real money, because competition and speed are what produce a strong price within the range. But it does not substitute for condition: a staged home with an ageing roof, an old panel and a kitchen at the end of its life will still be adjusted against comps that do not have those problems.
About the author
Brayden Benz is the founder of Stelo Inc.
