Two residential IPs, both labelled United States, both delivered from the same pool in the same request. One belongs to a consumer broadband operator in Denver. The other belongs to a hosting company in Ashburn. On a defended site those two addresses get treated as different species: one is where customers browse from, the other is where servers live, and the difference is visible before your request is a millisecond old.
The country label could not tell you that, because the country is not what the site checked. It looked up the operator. Every IP address on the internet belongs to an autonomous system, the network operator that announces the block it sits in, and that mapping is public, cheap to query and stable. Country is a derived, fuzzy attribute. The ASN is a record.
This is a guide to buying and verifying on that axis: when the operator is the actual requirement, how to check what you have been given, and where ASN targeting is worth paying for rather than a spec-sheet ornament.
What the target looks up, in order
The first checks a defended site performs on a new connection are all cheap lookups that need no cooperation from you.
Which ASN announces it
public routing data, instant
What kind of operator is that
consumer ISP, mobile carrier, hosting, cloud
What is its recent history
abuse reports, prior traffic, block lists
Only the first is a fact. The second is a classification the vendor maintains, and the third is reputation accumulated by everyone who used that address before you. Country appears nowhere in that sequence, which is why it is a poor lever: it is one of the least load-bearing attributes of an IP in the decision that actually happens.
This is also the mechanism behind the entire ISP proxy product category. An ISP proxy is an address registered under a consumer operator's ASN but hosted on datacenter hardware, and it works precisely because the check reads registration rather than physical location. Same principle, used deliberately.
Four jobs where the operator is the requirement
Most work does not need ASN targeting. These four do, and in each one country targeting produces confidently wrong results rather than obviously broken ones, which is the expensive failure mode.
Ad verification. A campaign buys placements against a specific audience, and fraud frequently hides behind a network the campaign never intended to reach. Verifying what the ad looks like to a real subscriber of a named ISP or carrier means exiting through that operator. From a different operator in the same country you see a different auction and, often, a different creative. Our ad verification page covers the wider workflow.
Mobile carrier behaviour. Carriers do things nobody else does: captive portals, content filters applied at the network edge, zero-rated destinations, CGNAT mapping many subscribers to one address. Reproducing any of that requires exiting through the carrier's own network, and no amount of country accuracy substitutes.
Testing your own detection. If you run a signup flow, a payment step or a fraud model, you want to know how it behaves for traffic from a hosting ASN versus a consumer one. That is a deliberate experiment across operator types, and it is the fastest way to discover that your own product silently punishes people on a small regional ISP.
Reproducing an ISP-specific bug. Peering and CDN routing differ per operator. A support ticket that says the site is slow only for one provider's customers is not reproducible from any other network, and being able to enter through that operator turns a week of guessing into one measurement.
Check what you actually have
Never take the pool label as evidence. The label is a claim by the seller; the ASN is a record you can query yourself in a second. Three ways, all free.
# 1. What does the internet's routing data say about this address?
whois 203.0.113.10 | grep -iE 'origin|netname|org|country'
# 2. Same question via a JSON endpoint, through the proxy itself
curl -s -x http://USER:PASS@HOST:PORT https://ipinfo.io/json
# 3. Confirm the exit you got is the exit you asked for
curl -s -x http://USER:PASS@HOST:PORT https://ipinfo.io/json | grep -E '"org"|"country"|"city"'
Command two is the one that matters, because it asks the question from inside the connection you are actually going to use. A pool that advertises consumer ISPs and returns a hosting organisation on the exit has answered the only question worth asking, and it answered it against itself.
Our proxy checker does the same lookup in a browser and adds the anonymity grade and the exit country, so you can run it on any IP, ours or a competitor's, before spending anything. The same reasoning is why our free proxy list publishes per-ASN liveness instead of a single national figure: a claim about a country only means something once it is broken down to the operators inside it, and two providers in the same country routinely behave nothing alike.
What providers mean by ISP targeting, and what they do not
The phrase covers at least three different products, and the gap between them is worth one email before you buy.
A filter on the rotating pool. You request an exit from a given operator and the gateway serves one if it has capacity there. Availability is the catch: a large operator in a large country is easy, a small regional ISP at three in the morning may quietly fall back to something else, and whether the fallback is announced varies by provider.
A static ISP product. You rent a specific address under a consumer operator's ASN for a period. Availability is deterministic because the address is allocated to you rather than drawn from a pool.
A display name with looser matching underneath. The interface shows an ISP name, but the filter behind it is a broader category. This is the one to detect, and the detection is a single question: which ASN number does this correspond to? A provider filtering on ASN can answer with a number. A provider doing name matching will answer with prose.
Two follow-ups turn that into a real spec. First, what happens when the requested operator has no capacity, an error or a silent substitution. Second, how many distinct ASNs the pool actually contains in the country you care about, since a pool with millions of addresses concentrated in three operators is not the same product as one spread across two hundred. Our vetting guide covers how to test both claims rather than accept them.
Measure the pool instead of asking about it
Pool-size numbers are the least verifiable claim in this industry, and operator diversity is the part that actually decides whether targeting will work when you need it. You can measure it yourself in about ten minutes, on a trial or on a gigabyte, and the measurement is far more informative than the number on the pricing page.
Pull a couple of hundred exits with rotation on, record the ASN each one lands on, and count.
# 200 rotating exits, one line of ASN + org per request
for i in $(seq 1 200); do
curl -s -x http://USER:PASS@HOST:PORT https://ipinfo.io/json \
| grep -o '"org": *"[^"]*"'
sleep 0.3
done | sort | uniq -c | sort -rn
Three numbers fall out of that list, and each one answers a different question.
How many distinct operators appeared. This is the real diversity of the pool in that country, as opposed to the address count. A pool advertising millions of IPs that returns eleven distinct operators is eleven networks wide, and every filter you apply has to live inside those eleven.
What share the top operator took. Concentration is what makes targeting fail at the moment you need it. If one ASN carries most of your samples, then requesting any other operator is a request the gateway may not be able to fill.
How many samples were hosting rather than consumer. This is the honesty check on the word residential. A pool that returns cloud and hosting organisations under a residential label has told you what it is, in its own data, for the price of a few megabytes.
Run the same loop against a second provider and you have a comparison nobody publishes, built from measurements rather than marketing. It is the same method behind our own source quality study, which is where we applied it to the free-proxy ecosystem instead of to paid pools.
Four questions worth one email
If you would rather ask than measure, these four are specific enough that vague answers are themselves informative.
- Which ASN numbers can I select in country X? A number is a filter. A name is a category.
- What happens when the requested operator has no capacity? An error lets you handle it. A silent substitution corrupts a dataset quietly.
- How many distinct ASNs does the pool cover there? Compare the answer with your own sample count afterwards.
- Is targeting applied per request or per session? This decides whether a multi-step flow can hold one operator all the way through, which matters for anything with server-side state.
The trap: the ASN is registration, not physical truth
ASN targeting answers "which operator announces this address". It does not answer "is a human sitting behind it", and conflating the two leads people to over-trust the filter.
An address registered under a consumer ISP can be hosted in a rack, which is exactly what an ISP proxy is and why it is a legitimate product rather than a trick. A residential ASN can carry a compromised device, which is the supply problem behind the takedowns we documented in 911 S5 and RSOCKS. And a clean ASN can carry a burned address, because reputation lives per IP as well as per network, and the previous tenant of that address is not something the ASN can tell you about.
So the operator is the first filter, not the whole answer. What sits behind it is a sourcing question, which is why how websites detect proxies spends most of its length on the signals that come after the lookup.
A selection ladder that does not overspend
Targeting granularity costs money and availability. Buy the coarsest level that answers your question, then move up only when a measurement forces you to.
- Country. Right for most collection work, localisation checks and general access. Cheap and widely available.
- State or region. Right when the thing you are measuring varies inside a country: regulated pricing, regional catalogues, state-level availability.
- City. Right for local search results, delivery and store-availability checks, and any market where the city is the unit of the business.
- ISP or ASN. Right when the operator is the subject of the test rather than a detail of it. The four jobs above, and little else.
Our residential tiers follow that ladder, from country targeting at the entry level up to city targeting, with ISP and ASN selection on the higher tiers. The residential page lists which level sits with which tier, and targeting is selected through parameters on the connection rather than through a separate product, so moving from country to city does not mean changing integration. Pricing starts at $0.50/GB for a single gigabyte on a balance that does not expire, which makes testing a targeting level cheap enough to do before committing to it.
The honest closing note is the one that applies to every filter on this page. ASN targeting removes one specific reason to be refused. It does not make a client that behaves like a robot look like a person, and if your requests are being blocked at the same rate from a perfectly targeted consumer ASN, the problem moved to a layer this lever does not reach.