Almost every homeowner I talk to has checked their Zestimate before calling me. It is a fine place to start. It is not a valuation, and treating it like one is where sellers leave money on the table.

The number on Zillow is the least accurate version they publish

Zillow is more transparent about its accuracy than most companies would be. Its published figures put the Zestimate's median error at roughly 2% for homes actively listed for sale — and roughly 7% for homes that are not on the market. On a million-dollar home, a 7% median miss is $70,000 in either direction. And "median" means half of homes miss by more.

There is a reason the on-market number looks so tight: once a home is listed, the model can see the list price and market activity. It is, in part, copying the homework of the agent who priced it. If your home is sitting off the market while you casually check its value, you are seeing the version of the number Zillow itself is least confident in. Even Zillow's own pricing models struggled at scale — the company shut down its Zillow Offers home-buying program in November 2021.

What the algorithm has never seen

A model works from public records, past sales, and listing data. It has never stood in your driveway. In San Diego, the things it cannot see are often the things that move value the most:

Condition and renovation quality. Two homes with the same square footage can be a full tier apart — one with a twenty-year-old kitchen, one with a thoughtful renovation. The model sees "remodeled" in a listing description. It cannot tell the difference between a flip and a craftsman-level build.

Micro-location. In La Jolla, a home one street off the bluff can price meaningfully differently from its neighbor two blocks inland. View quality, orientation, canyon versus street side, and whether the view is protected or one development away from disappearing — none of that is in the tax record.

Lot usability. Usable outdoor space, ADU potential, and how a lot actually lives matter enormously to San Diego buyers, and they vary block by block in ways a model flattens out.

What a real San Diego home valuation actually examines

When I value a home, the work is a comparative market analysis — comparable closed sales first, then judgment. I pull sales from roughly the last 90 days of properties that are the same kind of home: similar architecture, quality tier, and setting. A 1960s ranch that was fully rebuilt competes with rebuilt homes, not with original-condition ranches nearby.

Then I adjust for the differences: condition, view, lot, upgrades, and micro-location. I look at what is active and pending right now, because that is your competition — the buyer will compare you to the listings they can walk into this weekend. I also watch off-market activity and expired listings, which tell me where sellers misread the market recently. For the latest countywide picture, see my Fall 2026 market brief — but your price is set at the property level, never from a county median.

The three pricing mistakes that cost sellers the most

1. Chasing the highest number. Interview three agents and you will get three numbers. The one you want is not the highest — it is the one with the clearest evidence behind it. An inflated number wins the listing and loses the sale; the buyers' agents will show their clients the same comps I see.

2. Pricing to "test the market." Every pricing study in the industry says the same thing: the first two to three weeks are when a listing gets its most attention. Testing a premium eats that window, and a price reduction later is a signal buyers read as weakness. You cannot get the launch back.

3. Pricing from what you spent. Buyers do not pay your cost basis. They pay what the market supports for what they can see. A $200,000 kitchen does not add $200,000 — it adds what a buyer values a $200,000 kitchen at, in your neighborhood, this month. Pricing works backwards from the buyer, not forwards from your receipts.

Questions to ask any agent before you set a price

Before you agree to a number, ask the person recommending it to show their work. The questions I welcome from sellers:

Show me the comps — and tell me why each one counts. Which sales are most like your home, and what did you adjust for?

Who is my competition this week? Which active listings will buyers compare us against, and how do we beat them?

Where would you price versus where would you list? Strategy matters — list price and expected sale price are not always the same number, and you should understand the plan.

What happens if we don't get traction? You want an agent with a day-14 plan, not just a day-1 plan.

Browse my neighborhood guides to see how differently value behaves across La Jolla, Del Mar, Pacific Beach, and the rest of the coast — or start with the La Jolla guide, where micro-location matters more than anywhere else in the county.