ZIP codes are not areas
A ZIP code is a routing device, not a place. The Postal Service defines ZIPs as collections of delivery routes and addresses, publishes no boundary files for them, and changes them continuously as routes change. A "ZIP map" is always someone's approximation.
That creates a problem for media planning: campaigns buy ZIPs, but demographics live in census geography. Something has to bridge the two.
The ZCTA is that bridge
A ZIP Code Tabulation Area (ZCTA) is the Census Bureau's area version of a ZIP code. The Bureau takes every census block, looks at the addresses inside it, and assigns the block to the ZIP code most of those addresses use. The blocks assigned to one ZIP form that ZIP's ZCTA. The current vintage is built from the 2020 Census and stays fixed until the next one.
Because a ZCTA is made of census blocks, every census statistic can be tabulated for it exactly: population, voting-age population, households, and every ACS estimate. That is what makes ZIP-level planning against census data possible at all.
Five misconceptions that cause bad numbers
1. "Every ZIP code has a ZCTA."
No. A ZCTA exists only where people live. PO-box ZIPs, single-building ZIPs, and large-organization ZIPs have no residential footprint, so they have no ZCTA. About one in five ZIP codes has none. When the List Grader reports "unmatched ZIPs," this is why. Those targets are real people, but their mailing ZIP has no census geography to score.
2. "A ZCTA boundary is the ZIP boundary."
Close, but not equal. The Census Bureau builds each ZCTA from whole census blocks, and assigns each block to the ZIP most common among its addresses. Edges differ from the carrier routes, so an ad platform's ZIP audience never matches ZCTA population exactly. The CTV Planner's ZIP match rate (default 90%) prices this in.
3. "ZIPs nest inside districts, counties, and states."
They do not. ZIPs follow delivery routes, and routes ignore political boundaries. ZCTAs routinely cross district and county lines, and a few cross state lines. This is why every list in Badger Pass carries an allocation factor: the share of the ZCTA's population that falls inside the boundary you selected.
4. "Vendor "ZIP demographics" describe the ZIP."
Almost all commercial ZIP-level demographics are ZCTA data relabeled. If a vendor sells you "ZIP population," it is nearly always the 2020 ZCTA figure. Badger Pass uses the same geography and says so.
5. "New ZIP codes are covered."
ZCTAs are frozen at the 2020 Census. USPS creates and retires ZIPs continuously. A ZIP created after 2020 matches no ZCTA, and its residents are counted inside the ZCTAs that existed when the census was taken.
Why we use ZCTAs anyway
There is no better option. ZCTAs are the only ZIP-shaped geography with authoritative demographics attached, and crosswalking them to districts through census blocks (Geocorr 2022) is the standard practice in redistricting and market research. The errors are known, bounded, and mostly at the edges. Pretending a vendor's ZIP polygon file is exact does not remove those errors; it hides them.
What this means for a buy
Expect ZIP-targeted delivery to differ from ZCTA numbers at the margins, and plan for it:
- Check the unmatched ZIPs note when you grade a list. Those targets exist, but no census geography can score them. Decide separately whether to keep those ZIPs in the buy.
- Use the ZIP match rate in the CTV Planner to price geolocation slop. If your platform publishes its own IP-to-ZIP accuracy, enter that instead of the default.
- Read allocation factors before trimming a list. A ZIP that is 40% in-district is not a bad ZIP; it is a ZIP where 60% of impressions buy someone else's voters.
Full sources and vintages are on the methodology page.