In the same thirty-day window, one South Austin pocket watched its tracked median sale price triple, while a mile away in Garrison Park, the twelve-month median sat at $407,000, down 14 percent from the year before. Both numbers came from inside the same zip code. Both are technically accurate. Neither one describes 78745.

That's the problem with treating 78745 as a single market. It isn't one. It's a loose federation of small, thinly traded pockets, Garrison Park, Westgate, South Manchaca, Sweetbriar, St. Elmo, the Sunset Valley-adjacent streets near Deer Park, each behaving on its own schedule, and a zip-wide median that averages all of it into a number that describes none of it particularly well. If you're using that median to decide whether 78745 is a good place to buy, or to price a listing inside it, you're reading a blended average of markets that don't actually compete for the same buyer.

What a 180 percent jump on four sales actually tells you

In the thirty-day window that closed out in early January 2026, South Manchaca's tracked median sale price hit $1,260,000, up 180 percent year over year. That figure came from four recorded sales, down from seven in the same window a year earlier.

Four sales is not a market. It's a coin flip that happened to land on one expensive property and skip the modest ones. When a pocket that small only closes a handful of deals in a month, one large custom build or one estate sale on a big lot can drag the median wherever it wants to go, and the swing back down is just as likely the following month. That's not appreciation. It's arithmetic doing what arithmetic does when the sample is tiny.

This matters because a buyer scanning 78745 data might see South Manchaca's number and assume the pocket has repriced overnight. A seller in the same pocket might see it and assume their 1,700-square-foot ranch is suddenly worth seven figures. Neither read holds up once you know the transaction count behind the headline.

The zip that's actually six markets

Even the zip-wide number for 78745 doesn't sit still long enough to trust on its own. Pull it from three sources in the same season and you'll get three different answers: a median list price of $480,000 in July 2026, a median sale price of $467,000 for the three months ending in April 2026, and an average price closer to $527,000 as of May 2026. None of those sources is wrong. They're measuring different windows, different mixes of homes, and in one case average instead of median. The instability isn't a data error. It's what happens when you try to compress six different pockets into one line.

Here's roughly what those pockets look like on their own terms right now:

A buyer comparing "78745" to a neighboring zip on the strength of a single median is comparing an average of averages. A seller pricing off that same number is anchoring to a figure that has almost nothing to do with their specific street.

The lots quietly resetting the entry point

There's a second force pulling on this zip's numbers that has nothing to do with existing homes changing hands. Austin's HOME ordinance, in its second phase, has made it possible to subdivide and build multiple units on lots that used to hold one house, and 78745 has some active examples of exactly that.

One listing currently on the market describes five contiguous lots facing Armadillo Neighborhood Park, with residential plans in process through DesignTrait Architecture to build thirteen units across roughly 18,251 square feet of the assemblage. Another, in the Deer Park pocket zoned to Sunset Valley Elementary, is a fully demolished lot where the listing notes plainly that the HOME ordinance's second phase means the site can potentially be subdivided for smaller homes rather than rebuilt as one.

Neither of these is a hypothetical. They're active projects, and they matter for the median conversation because they're adding a type of inventory 78745 didn't have much of before: smaller-footprint, new-construction units landing at a price point below the existing single-family stock around them. When those units close, they'll pull the zip-wide median in a direction that has nothing to do with whether a 1965 ranch in Sweetbriar is appreciating or not. A buyer comparing price per square foot between a new 1,700-square-foot unit near Armadillo Park and an older three-bedroom a few streets over isn't comparing similar products, even if both sit inside the same zip code.

What the zip-wide numbers are still good for

None of this means the broader 78745 numbers are useless. They're just better read as a temperature check on the whole zip rather than a price tag for any specific home.

By that measure, the zip has cooled. As of the three months ending in April 2026, homes in 78745 were taking an average of 62 days to sell, and buyers were putting in roughly one offer on average, a sign of far less competition than the multiple-offer environment of a few years ago. As of July 2026, price per square foot in the zip was down about 5 percent year over year. That's consistent with what several of the pocket-level numbers above are showing too: Garrison Park down 14 percent, Westgate's estimate softer than its recent peaks. The exception, South Manchaca's headline spike, is the one that needs the asterisk, not the other four.

What this means depending on where you're actually looking

If you're comparing 78745 to another zip code using a median you found online, ask which pocket that median is actually weighted toward, and over what window. A median built from a busy pocket like Garrison Park behaves differently than one skewed by a few sales in a quiet pocket like South Manchaca.

If you're buying in 78745, get comps from the specific pocket and street, not the zip. Ask whether any nearby lots have HOME-enabled infill plans in process, since new lower-cost units landing near you can either widen your options or shift what "comparable" means for your own purchase.

If you're selling, resist pricing off the zip-wide number in either direction. A seller in South Manchaca watching that pocket's tracked median triple should not assume their own home just did the same thing. A seller in Garrison Park seeing a 14 percent pocket-level decline should still get a comp set pulled from actual recent closings on their block before assuming the whole neighborhood dropped that much.

A short list worth keeping on hand before you write an offer or set a price:

• Ask for the transaction count behind any pocket-level median before reacting to it
• Compare price per square foot only against homes of a similar build era and lot size, not across new infill and older stock
• Check whether any nearby lots have active HOME-ordinance subdivision or multi-unit plans filed
• Pull the most recent 60 to 90 days of closings for your specific pocket, not the zip average

FAQ

Why did South Manchaca's median jump 180 percent in one window? It reflects four recorded sales in a thirty-day period, not a broad repricing. When a pocket closes only a handful of deals, one higher-priced sale can move the median sharply in either direction the next month.

Does the HOME ordinance mean every lot in 78745 can be subdivided? No. The examples here, the Armadillo Park assemblage and the Deer Park teardown, are specific active projects tied to particular lots and zoning conditions. Whether a given property qualifies depends on its own lot size and zoning, which is worth confirming property by property rather than assuming zip-wide.

Is 78745 a buyer's market right now? The broader trend points that way. Longer average time on market and softer price per square foot as of mid-2026 both suggest less competition than in recent years. But that trend plays out differently pocket by pocket, which is exactly why the zip-wide answer isn't the whole answer.

If you're trying to figure out what a specific 78745 pocket is actually doing, or whether an infill project down the street changes how your home should be priced, that's a conversation worth having before you set a number. Cody Hobza works these South Austin blocks street by street, not just zip by zip. Let's Connect.