Two hundred and seventy-five chamber-of-commerce domains, every name published, every reading repeatable

This company has published measurements before without publishing the list of domains behind them. That is a standard we hold other people to, so we now hold it to ourselves. Below is the whole population, named, with the number a receiver would compute for each one, and a link that re-reads the record live rather than replaying this table.

How this list was built, stated so you can rebuild it

1. Take the 645 city and village names in Wisconsin, from the English Wikipedia lists of cities and of villages in Wisconsin (read 4 September 2026).
2. Lower-case each name and strip everything that is not a letter or a digit. For each, form five candidate host names: <name>chamber.com, <name>chamber.org, <name>areachamber.com, <name>areachamber.org, <name>chamberofcommerce.com. That is 3,225 candidates.
3. Keep a candidate if it publishes at least one MX record on either of two independent resolvers. 407 survive.
4. Keep it if its own web site identifies itself as a chamber of commerce — the word appears in the page title, or the phrase “chamber of commerce” appears in the served HTML. 275 survive, and those 275 are this list.
5. For each, read the SPF record on two resolvers and count the DNS-querying terms a receiver would perform, per RFC 7208 section 4.6.4, expanding every include and redirect to the bottom.

What this list is not

It is not a directory of Wisconsin chambers, and we are not going to call it one. The names were built from Wisconsin place names, but a place name is not owned by a state: brooklynchamber.com and columbusareachamber.com are in this list because Brooklyn and Columbus are Wisconsin villages, not because those two sites are in Wisconsin. Of the 275, 67 carry evidence of a Wisconsin location in their own served HTML — the word Wisconsin, a “, WI”, a 53xxx or 54xxx postcode, or a Wisconsin area code. The other 208 are marked not confirmed, which means exactly that and nothing more: most chamber sites of this kind render their address in the browser rather than in the HTML, so an absence of evidence here is not evidence of absence. Both columns are given so you can use whichever population your question needs.

The numbers

The last row is the one that matters operationally, because it is the only one anybody can act on without asking a vendor for anything. A record at 11 whose whole budget is spent inside one vendor include cannot be fixed by its owner. A record at 11 that spends four lookups on names its own owner controls can be.

The nine that are broken right now

These nine publish more than ten DNS-querying terms. RFC 7208 section 4.6.4 says a receiver that exceeds the limit MUST return PermError, so for these domains SPF currently produces no useful result at all, whatever else the record says. Each was counted three times: twice by this system on two resolvers, and once more by a separately written counter reading only dns.google. All three agreed on all nine.

The whole list

Sorted by cost, highest first. — in the cost column means the domain publishes no SPF record on either resolver. Every domain name links to a page that reads that domain again, now, and shows its working.

Honest limits of this measurement

• The frame catches only chambers whose domain follows one of five naming patterns. A chamber at visitsomewhere.org is invisible to it. This is a sample with a stated rule, not a census.
• DNS is read at a moment. Every figure here is what two resolvers returned on 4 September 2026; the link on each row is there because the record may have changed since.
• The Wisconsin column under-counts, for the reason given above.
• We measured these domains from public DNS and public web pages only. Nobody was contacted, and nothing here was solicited from anyone.

The free lookup counter · The free record check · What each include costs · What else we do