How the restaurants in our city guides list delivery

A national look at delivery status across 1,416 restaurant records in 210 indexable city guides: what we checked, how we label it, and what a shortlist of a handful of places per city cannot show.

What this is

Chopstick Compass is built as a careful, sourced Chinese food guide. Each city page is an unranked shortlist drawn from operator sites, linked menus and platform listings, with a check date on every restaurant. This piece reports one field from those records: how delivery was listed when we checked.

It is not a census of Chinese restaurants in the United States. It is not a ranking of cities. It does not claim that we visited kitchens, placed orders or tasted food. The numbers describe the restaurants already in our indexable guides, nothing more.

What each status means

Every restaurant record carries one of four delivery statuses (from the city schema and the badges on each city page):

Delivery status definitions
StatusBadge textMeaning in our data
directDelivery: directDelivery appeared on the operator’s own ordering path.
third-partyDelivery: via appsDelivery appeared through a checked platform listing, not as operator-direct.
pickup-onlyPickup onlyThe checked sources supported pickup (or dine-in planning) without a verified delivery offer.
none-listedNo delivery listedDelivery could not be verified on the operator ordering page or checked listing; the schema uses this when delivery is unconfirmed.

On the site, “lists delivery” means direct or third-party — the same set our hasDelivery filter uses. Availability still depends on the reader’s address and time; a listed channel is not a promise that a given address will be served.

The national split (indexable cities)

Across 210 indexable city guides and 1,416 restaurant records (pinned commit ab114f5, export date UTC):

National delivery statusOne bar of 1,416 restaurant records, left to right: Delivery: direct 594 (41.9%), Delivery: via apps 229 (16.2%), Pickup only 375 (26.5%), No delivery listed 218 (15.4%).

Left to right

  • Delivery: direct (solid)594 (41.9%)
  • Delivery: via apps (stripes)229 (16.2%)
  • Pickup only (dots)375 (26.5%)
  • No delivery listed (lines)218 (15.4%)
1,416 restaurant records in 210 indexable Chopstick Compass city guides. Pinned commit ab114f5; export 2026-10-27 UTC. Statuses as listed on checked pages dated 2026-09-27 through 2026-10-09. Not a census of all Chinese restaurants.
National delivery status counts
StatusRecordsShare of 1,416
Direct59441.9%
Third-party22916.2%
Pickup-only37526.5%
None-listed21815.4%

So 823 records (58.1%) list delivery in our sense, and 593 (41.9%) are pickup-only or none-listed. Among the 823 that list delivery, 72.2% are direct and 27.8% are third-party. No record in this export had a missing or non-enum status (unknown = 0).

Lists delivery82358.1%
Pickup-only or none-listed59341.9%

Of those that list delivery, 72.2% direct / 27.8% via apps.

City guides were checked between and (dates on the checked field). Most indexable rows carry .

What holds at a small per-city sample

Each indexable city guide holds 6 to 11 restaurants (median 7; all but one guide hold 6 to 8). That is enough to plan a meal from a shortlist; it is not enough to say how a metro “does delivery.” Findings we treat as honest at this grain:

  1. The national channel mix above — pooled across 1,416 records, not inferred from any one city.
  2. Mixed statuses are normal. In 201 of 210 indexable cities, at least two different statuses appear in the same shortlist; 166 cities show three or more. A single city page often includes a place that delivers direct, one that uses apps, and one that does not list delivery.
  3. “None-listed” is common, not rare. 128 of 210 indexable cities include at least one none-listed record. That usually means we could not verify delivery on the checked page, not that the kitchen never delivers.

We do not rank cities by delivery share. With a handful of places per guide, city-level percentages swing on one or two records.

Methodology

  • Scope of this export: one CSV row per city the site publishes (badgeGuides in src/data/featured.ts). JSON guides use the matching record in cities (src/data/cities.ts). Austin uses austinRestaurants from src/data/austin.ts. Counts are restaurant records on those guides.
  • Indexable filter: cityIsIndexable in src/data/cities.ts (indexable !== false). Same rule city-page metadata and the sitemap use. Headlines here use only indexable=true rows.
  • How status is assigned: from the operator site, an operator-linked ordering page, or a platform listing fetched on the check date. If delivery cannot be verified, status is none-listed with a dated detail string. We do not assume a platform serves a branch.
  • Pinned commit / export date: ab114f5 (PR #39 head; files verified at that commit), export date (UTC), per data/exports/delivery-by-city/ on that commit.
  • Selection: restaurants are an editorial shortlist for each city guide, not a random sample of all Chinese restaurants in that city.

Limitations

  • Per-city samples are small (most guides 6–8 restaurants; median 7). Do not treat city percentages as market shares.
  • Check dates run from through ; menus and platform listings change.
  • “Lists delivery” is about what the checked page showed, not whether a reader’s address is in range, and not tip, fee or markup behavior.
  • Thirty-one cities with indexable: false remain in the CSV (201 records) but are excluded from these headlines.
  • This page summarizes an internal export; a downloadable dataset is not published on chopstickcompass.com.