How Much Does It Cost to Sell a House in San Francisco? A Seller Net Proceeds Guide

2026 median SF house

$2.105M

1,640 closed sales through October 31

Transfer tax at that price

$15,788

$2,105,000 ÷ $500 × $3.75

The number that matters

Net proceeds

Price less payoff and every selling cost

Data: Current SF transfer-tax law checked November 2, 2026; owner-supplied closed-sale medians through October 31, 2026.

The sale price is the number everyone repeats. The net proceeds are the number that changes your life.

My view is that a San Francisco seller should build the net sheet before deciding how much to spend, which listing proposal to accept or whether an offer is good enough. Starting with the headline commission rate misses the larger decision. Price, transfer tax, preparation, credits, timing and execution all move the final check.

At the 2026 year-to-date house median of $2.105 million, the City transfer tax alone is $15,787.50. A quarter-point change in compensation is $5,262.50. A $40,000 preparation plan needs to create more than $40,000 of additional net value after any delay and carrying cost. Those are three different decisions, and they belong on one page.

What sellers usually pay in San Francisco

The payoff deserves its own line. Your online mortgage balance is not necessarily the closing payoff. The Consumer Financial Protection Bureau explains that a payoff statement includes the amount required to satisfy the loan on a specified date.

San Francisco transfer tax is a cliff, not a marginal bracket

The City applies the rate for the sale-price tier to the entire value or consideration. Crossing a threshold can therefore change the tax sharply. This is not like a marginal income-tax bracket where only the dollars above the line receive the higher rate.

The $5 million line is the most important cliff for many higher-end residential sellers. At $4,999,999, the calculation is $37,500. At $5 million, it is $112,500. That $1 of price changes the transfer-tax estimate by $75,000.

This does not mean a seller should manipulate a contract price or reject a stronger offer to avoid tax. It means the listing team and escrow should model the real net before a threshold-crossing price is accepted. At $10 million the same issue becomes much larger: $225,000 just below the line versus $550,000 at the threshold.

What the 2026 medians imply

These medians are useful scale, not a pricing shortcut. A recent public market report also described September inventory as 25% below the prior year for houses and 34% lower for condos, with both property types selling quickly. Strong market conditions can improve the price side of the net sheet. They do not reduce transfer tax or make every preparation expense worthwhile.

Brokerage compensation is negotiable, but the cheapest proposal is not automatically the best net

Agent compensation is not set by law. The listing agreement defines what the seller pays the listing brokerage, and a seller may separately choose whether to offer buyer-agent compensation or another concession. NAR’s listing-agreement guide describes compensation as fully negotiable.

A quarter point equals $5,262.50 on a $2.105 million sale. That is real money. But a fee comparison without a common sale-price assumption, preparation scope, marketing plan and service list is incomplete. The right question is not “Which percentage is lowest?” It is “Which plan has the strongest defensible net after every cost?”

Preparation should earn back more than it costs

The current SF market rewards presentation, but there is no reliable rule that every staged or remodeled home earns a fixed premium. The 2025 NAR staging survey asked agents, not San Francisco buyers, and reported opinions rather than matched-property returns. It is useful evidence that presentation can help buyers visualize a home, not a promise that a particular staging invoice pays for itself.

Use a break-even test. If paint, repairs, staging and two extra months of carrying cost total $55,000, the plan needs to add more than $55,000 to expected net proceeds to compensate for the work and execution risk. A higher gross price is not enough if the cost and delay consume it.

This is where our San Francisco pricing guide and Victorian and Edwardian seller guide become useful. Price the home you actually have, then decide which uncertainty or presentation problem is worth paying to remove.

Use the calculator before you approve the budget

The default calculator example uses the 2026 house median, a $900,000 payoff, 2.5% listing compensation, 2.5% buyer-agent compensation, $40,000 of preparation and $10,000 of other closing costs. The math is: $2,105,000 − $900,000 − $52,625 − $52,625 − $15,787.50 − $40,000 − $10,000 = $1,033,962.50 before income-tax withholding and final prorations. These are teaching inputs, not a fee quote or expected transaction.

Withholding is not the same as the tax you owe

California real-estate withholding often causes unnecessary alarm because it appears as a large reduction on the closing statement. The Franchise Tax Board calls it a prepayment of income tax. If no exemption applies, the sales-price method is 3 1/3% of the price; sellers may also qualify for an exemption or elect an alternative gain-based calculation.

At $2.105 million, 3.33% is about $70,097. That does not mean the sale created a $70,097 California income-tax bill. It means the escrow calculation and Form 593 need to be handled before closing rather than discovered when the final statement arrives.

Federal gain is separate again. IRS Publication 523 explains the main-home exclusion of up to $250,000 for qualifying individuals or $500,000 for qualifying married couples filing jointly. Basis, improvements, prior rental use, depreciation, ownership and residence history can change the result.

Compare offers by net proceeds and closing risk

A higher offer can produce a lower expected net if it includes a large credit, weak financing, a long close or a repair demand that is likely to reopen negotiations. Our preemptive-offer analysis makes the same point from the timing side: price is one term, not the whole decision.

The three net sheets I would insist on

1. Before preparation: compare an as-is plan with the proposed work, including delay and carrying cost.
2. Before listing: update the expected price range, payoff, transfer tax, negotiated compensation and every approved vendor cost.
3. Before accepting an offer: replace assumptions with the actual price, credits, compensation, close date and material execution risks.

I would rather explain a conservative net sheet early than celebrate a sale price that leaves the seller disappointed at closing. Build the model before spending, update it before listing, and make the final decision from the offer that is actually on the table.