A residential proxy plan that looks cheap on the pricing page can quietly turn into the most expensive line item in your data budget. The trap is almost always the same: teams compare headline numbers, sign up, and only discover the real cost weeks later when the bandwidth bill lands or the "unlimited" plan throttles them at the worst possible moment.
The honest answer to "which is cheaper, metered or unlimited?" is that it depends entirely on how you use bandwidth, not on the sticker price. This article gives you a method to work out your true cost per gigabyte before you commit, so you can compare two very different pricing models on equal footing.
Two Pricing Models, Two Different Bets
Residential proxy providers overwhelmingly sell in one of two shapes, and each one is a bet about your behaviour.
Metered (per-GB) plans charge for the traffic you actually push through the network. You buy a bucket of gigabytes, or you pay a rate per gigabyte consumed. The provider is betting that you will use enough bandwidth for the margin to work, and you are betting that your usage stays predictable.
Unlimited (or flat-rate) plans charge a fixed fee, usually per port, per thread, or per concurrent connection, and let you move as much data as you can within some fair-use ceiling. Here the provider is betting that most customers will not saturate their connections, and you are betting that you will.
Neither model is inherently better. They serve different consumption profiles, and the entire exercise of comparing them comes down to translating both into the same unit: your real, all-in cost per gigabyte delivered.
The Metric That Actually Matters: Effective Cost Per GB
List price per gigabyte is not your cost per gigabyte. The number you care about is effective cost per GB: the total you pay in a billing period divided by the volume of usable data you actually received.
The word "usable" is doing heavy lifting there. Not every gigabyte you are billed for does productive work, and that gap is where most cost estimates fall apart.
Start with billed bandwidth, not requests
Residential proxies bill on bytes transferred, not on requests made. That means every byte of overhead counts against you:
- Response payloads including HTML, JSON, images, and any assets you did not need but still downloaded.
- Request and response headers, which are small individually but add up across millions of calls.
- Retries, which duplicate the full cost of a request every time a target returns an error or a soft block.
- Failed sessions that consume bandwidth handshaking before they get blocked and return nothing.
If you scrape image-heavy pages without stripping assets, your billed bandwidth can be several times larger than the data you actually parse and store.
Build your true cost formula
Work it out per campaign or per month:
- Total spend for the period, including any per-port or subscription fees, not just the bandwidth line.
- Total billed GB the provider recorded.
- Useful ratio: the share of that traffic that produced data you kept. Measure it, do not guess. Log successful responses against total transferred.
Divide total spend by (total billed GB multiplied by useful ratio). That is your effective cost per usable GB. A plan advertised at a low per-GB rate but running a fifty percent useful ratio is, in reality, twice as expensive as the number on the page.
When Metered Wins
Metered plans reward low and irregular consumption. They are usually the cheaper choice when:
- Your volume is modest or spiky: a few gigabytes for periodic price checks, ad verification sweeps, or a monthly SEO rank audit.
- Your workload is lightweight per request: text-only scraping, API responses, or targeted lookups where you strip images and unnecessary assets.
- You need strict cost predictability and would rather cap a bucket than risk a runaway thread.
- You are still testing a target and do not yet know your steady-state volume.
The metered model also enforces discipline. When every gigabyte has a visible price, teams optimise: they compress, they cache, they stop downloading assets they never parse. That optimisation frequently pays for itself.
When Unlimited Wins
Flat-rate plans reward heavy, sustained throughput on a fixed number of connections. They tend to win when:
- Your consumption is high and continuous, running the same scrapers around the clock.
- Your work is bandwidth-dense: pulling large pages, media, or full documents at scale.
- You can saturate the ports you pay for, keeping concurrency high so the per-connection fee is spread across a large volume of data.
The catch is the fair-use ceiling and the concurrency limits. "Unlimited" almost never means unrestricted. Read the fair-use clause, the thread cap, and the throttling policy. A flat plan that quietly slows you after a threshold has an effective cost per GB that climbs the moment throttling kicks in, because your delivered volume stops rising while the fee stays flat.
The Hidden Variables That Break Naive Comparisons
Before you trust any per-GB number, account for the factors that distort it.
Success rate on your actual targets. A pool that succeeds ninety-five percent of the time on your sites is worth far more per gigabyte than one that succeeds seventy percent, even at a higher list price, because you waste less bandwidth on retries. Test against your real targets, not a generic benchmark.
Retry amplification. Every retry multiplies cost. A five percent block rate does not add five percent to your bill: with retries and re-authentication it can add considerably more.
Session control efficiency. Poor session persistence forces re-authentication and re-fetching, burning bandwidth on repeated groundwork. Sticky sessions where you need them cut waste.
Minimum commitments and expiry. Some metered buckets expire monthly. Unused gigabytes you paid for are pure loss and belong in your effective cost.
Geo and pool premiums. Specific-city targeting or mobile pools carry higher rates. If half your work needs premium geo-coverage, blend those rates into your average rather than quoting the cheapest tier.
Where Proxies Fit In
Once you understand that effective cost per GB is a function of success rate, retries, and session efficiency (not just the list price), the choice of provider becomes a question of infrastructure quality rather than headline pricing. A cheaper gigabyte that fails often is not cheaper at all.
This is where pool diversity earns its keep. Running the right pool for each job keeps your useful ratio high: lightweight lookups over efficient exits, heavy geo-targeted work over a pool built for it. A provider like EnigmaProxy offers residential, ISP, datacenter, and mobile pools under one account, so you can match pool type to workload instead of forcing everything through a single expensive tier and watching your effective cost climb.
Ethical sourcing matters to the calculation too. Cleanly sourced residential proxies tend to hold better IP reputation, which lifts success rates and cuts the retries that quietly inflate your per-GB cost. Predictable, transparent pricing plans also let you model both metered and flat-rate scenarios honestly before you commit, and validating exits with a proxy tester before a large run helps you confirm the success rate your cost model assumes.
Strategic Insights and Where This Is Heading
Several shifts are reshaping how teams should think about proxy cost.
Hybrid billing is becoming normal. More buyers now run a flat-rate plan for their steady baseline workload and keep a metered bucket for spikes and experiments. Splitting consumption this way often beats forcing everything onto one model.
Efficiency tooling is the real cost lever. As pages grow heavier, the biggest savings come from your own stack: request filtering, asset blocking, response caching, and smart retry logic. Teams that treat bandwidth as an engineering problem, not just a procurement problem, consistently pay less per usable gigabyte.
Transparency is a buying signal. Providers that publish clear fair-use terms and give you granular usage reporting let you build accurate cost models. Vague "unlimited" language is increasingly a red flag rather than a selling point.
Per-GB rates are compressing, but success rate is diverging. As raw bandwidth prices fall, the meaningful difference between providers is moving toward reliability and pool quality, exactly the variables that decide your effective cost.
Conclusion
The unlimited-versus-metered debate has no universal winner, only a right answer for your specific consumption profile. Metered plans reward disciplined, spiky, lightweight usage; flat-rate plans reward sustained, high-throughput work that saturates the connections you pay for.
The way to decide is to stop comparing list prices and start calculating effective cost per usable gigabyte: total spend divided by the volume of data that actually did useful work, adjusted for success rate, retries, and expiry. Do that math against your real targets before you sign anything.
When you are ready to model both options on genuinely reliable infrastructure with the pool diversity to keep your useful ratio high, EnigmaProxy is a solid, ethically sourced provider to run those numbers against.