Proxies solve a specific class of problem: accessing data or services that block, rate-limit, or geo-restrict automated requests. Here are the most common use cases.
Real Estate Data
Rents, availability and price history are city-level objects. A national average is the one number in this market that describes nothing anybody buys or rents.
Traffic Arbitrage
A media buy is a chain of promises: the network says the placement ran, the tracker says the click landed, the offer says the flow works. Each link is checkable only from inside the country you paid for.
Job Market Data
Job boards decide what to show from where you are: a search from one country returns that country’s listings, its salary bands and its currency, and nothing tells you the rest exists.
MAP Compliance
A seller who breaks your minimum advertised price rarely shows it on the listing. It appears in the cart, for one country, to a shopper who looks local.
App Store Monitoring
App store catalogues, rankings, prices and screenshots are built per country. Unlike a web shop, there is no query string that shows you Brazil from Berlin: the store is chosen for you.
Review Monitoring
The same product carries different reviews in different countries, and most platforms show a visitor only the ones written near them. Collect from one place and you report on one market.
Affiliate Marketing
A funnel that converts in one country can be broken in another: the wrong landing, a currency nobody uses locally, a redirect that never fires. From your own desk it always looks fine.
Multi-Accounting
Agencies, marketplace sellers and QA teams run dozens of profiles. What links them is rarely the password: it is the exit address, the timezone and the profile drifting apart.
Web Scraping
Websites ban scrapers by IP. Rotate through millions of residential IPs and they never see the same address twice.
Price Monitoring
E-commerce sites block scrapers that check prices too often from the same IP. Rotate through residential IPs and pull pricing data at any frequency.
Ad Verification
Ad networks serve different creatives by location, device, and time of day. Check from any market without flying there.
SEO Rank Tracking
Google personalizes search results by location, search history, and device. Get clean SERP data from any market with residential proxies.
Social Media Automation
Social platforms link accounts by IP. Assign a dedicated proxy to each account and they never appear connected.
Market Research
Market data is geo-locked, localized, and rate-limited. Pull it from the source with proxies that look like local users.
Sneaker Bots & Retail Automation
Nike, Adidas, and Supreme use Akamai and Cloudflare to stop bots. ISP and residential proxies have the trust score to get through.
Travel Fare Aggregation
Airlines and OTAs show different prices by IP location and browsing history. Residential proxies from the origin market return the actual local fare.
Brand Protection
Counterfeit listings and unauthorized resellers appear in markets where your team has no presence. Residential proxies give you eyes in every market.
Software & QA Testing
Your app works on your office network. Does it work in Brazil on a 4G connection? Proxies let you test from any network condition.
The catalogue below is arranged by job, but every job shares one question, and it is better settled before opening any card. The address type is chosen by the target rather than by the budget. Paying for residential where datacenter passes is the most common way to overspend here; taking datacenter where a home address is required is the most common way to collect nothing.
One ladder, the same for every job
The order is identical whether you are monitoring prices or verifying advertising. What changes is the step you stop at.
- Start with a datacenter address at $1.15 per address per month. Publishers, reference sites, government resources and open data serve it calmly, and paying for residential there is money wasted.
- A 403 on the very first request means the target filters hosting ranges. That is neither a ban nor a reason to change tools; it is the reason to move to residential.
- Country-level residential covers retail, marketplaces, classifieds and search results. You pay for the gigabytes you pull, so the bill follows page weight rather than request count.
- If the work has to run on one address, add a sticky session with a lifetime. If the address must not change for weeks, price up ISP: above two gigabytes a month on a session it is already cheaper.
- A mobile address is for targets that expect a carrier: apps and the platforms grown out of them. It is billed per day, so it is taken selectively rather than for volume.
The bill follows page weight, not request count
The most common mistake in an estimate, and it has nothing to do with the tariff.
Residential is billed by the gigabyte. So the cost of a job is page weight multiplied by page count, and that weight differs several times over between a careful and a careless collector. Twenty thousand listings a day with images blocked is around 60 GB a month. The same twenty thousand without media blocking can reach 900 GB for exactly the same data at the end.
So blocking images, fonts, video and ad scripts is part of the initial setup rather than an optimisation for later. If you work through a headless browser, disable media loading at the request interception level.
And measure before scaling: fetch twenty pages, look at the actual volume and multiply. An estimate by page count is most wrong precisely where pages are heavy.
Three mistakes common to every job
They appear regardless of what you collect, and all three spoil conclusions rather than collection.
- Measuring your own scraper’s schedule instead of the market. If yesterday’s slice was taken in the morning and today’s in the evening, the difference is yours rather than the market’s. Store the observation time, not only the date.
- Averaging what cannot be averaged. A monthly average price in a country where prices move fast hides the movement. Currency, region and language are their own columns rather than a label on the export.
- Treating a timeout as evidence of blocking. A refusal carrying a status code and a drop without one are different events with different causes, and not separating them in the logs inflates your blocking statistics.
Questions about choosing for a job
Which type should a new job start with?
A datacenter address at $1.15 a month. If the target returns a 403 on the first request, move to residential. The reverse order costs more and verifies nothing.
When is residential more expensive than ISP?
When the work stays on one address. Above two gigabytes a month on a single session, a static address at $1.90 works out cheaper and its traffic is unmetered.
Why is my bill higher than the estimate?
Almost always page weight. Budget in gigabytes rather than requests, and measure the real volume on twenty pages before scaling.
Do I need city targeting?
Usually not: it narrows the pool and adds failures. It earns its keep where region drives prices or search results, as in Belgium, Mexico and Russia, and the pages for those countries say so.
Can one job cover several countries?
Only if the target does not substitute content by visitor country. International platforms almost always do, so that is a separate pass per country with country as its own column in the data.