The choice between paying for proxies "by gigabytes" and "by the number of IPs" directly affects the project's budget. The same pool of proxies can cost significantly less or more depending on how you use it: farming Facebook Ads accounts, managing 30 Instagram profiles, or scraping prices on Wildberries every 10 minutes. In this article, we will analyze both models with numbers and show which is more profitable for four typical tasks.
How Payment by GB (Traffic) Works
When paying for traffic, you purchase a pool of gigabytes that are consumed with each request through the proxy: loading a page, images, videos, API responses. The "heavier" the content, the faster the balance depletes. This model is most often applied to residential proxies, as they are needed for tasks with unpredictable data volumes: browsing social media feeds, loading media, working with dynamic websites.
The advantage of this model is that you pay exactly for what you actually consumed. If today you warmed up 5 accounts, and tomorrow 50 — the consumption scales automatically, without the need to purchase new IPs. The downside is that it is difficult to predict the budget in advance, especially if there is video content involved (TikTok videos, Reels, YouTube Shorts consume traffic significantly faster than text and images).
Average traffic consumption for popular actions:
- Opening the Instagram feed for 5 minutes — approximately 15-25 MB
- Watching 10 Reels/TikTok videos — approximately 40-80 MB
- Uploading one post with a photo on Instagram — 3-8 MB
- Scraping one product card on Wildberries (HTML + JSON without images) — 0.1-0.3 MB
- Rendering the Facebook Ads Manager — 5-12 MB per session
How Payment by IP (by Number of Addresses or Ports) Works
The "payment by IP" model means that you purchase a fixed number of addresses (or ports tied to a specific IP) for a certain period — a day, a week, a month. The traffic within this package is either unlimited or includes a very large reserve, which is more than sufficient. This approach is typical for data center proxies and some mobile proxies.
The advantage of this model is predictability: you know exactly how much you will pay for 20 or 50 accounts, regardless of how many videos the bots watched or how many pages were scraped. The downside is that if an IP is idle (the account is not used today), the money is still deducted as if it were an active address.
This model is especially convenient where the number of accounts is known in advance and stable: for example, an SMM agency manages 30 client profiles every day, or an arbitrageur keeps the same set of ad accounts for weeks.
Comparison Table of Payment Models
| Parameter | Payment by GB | Payment by IP |
|---|---|---|
| Budget Predictability | Low — depends on activity | High — fixed amount for the period |
| Profitability with a Small Number of Accounts | Yes | No, overpayment for idle time |
| Profitability with High Load on One IP | No, traffic depletes quickly | Yes, unlimited within the package |
| Suitable for Video Content (TikTok, Reels) | Expensive | Profitable |
| Suitable for Scraping Text Data | Profitable | Excessive |
| Typical Type of Proxy | Residential | Data Center / Mobile |
Calculation for Arbitration: Farming Facebook Ads and TikTok Ads Accounts
A typical scenario: an arbitrageur farms 20-30 Facebook Ads accounts through Dolphin Anty or AdsPower, each account is tied to a separate IP to avoid chain bans. Activity on the account includes logging into the dashboard, viewing statistics, editing campaigns, and logging into the feed 1-2 times a day for warming up. This averages 30-60 MB of traffic per account per day.
When paying by GB: 30 accounts × 50 MB × 30 days ≈ 45 GB per month. This is a small amount, and the "traffic-based" model will be profitable if there are few accounts and low activity. But if you are simultaneously warming up TikTok Ads accounts by watching video feeds to simulate live behavior, the consumption can rise to 150-300 MB per account per day — then 30 accounts already yield 150-270 GB per month, and payment by IP with unlimited traffic within the package becomes significantly more profitable.
Practical conclusion: for cold farming without video warming — payment by GB. For full warming with simulation of a live user (browsing the feed, Stories, videos) — payment by IP with a fixed number of addresses for the entire campaign duration. For TikTok Ads, where video content constitutes the main part of the traffic, it is often wiser to choose mobile proxies with payment by IP — they trigger anti-fraud systems less frequently and are not limited by a thin traffic cap.
Calculation for SMM Multi-Account Management: 30 Instagram and TikTok Profiles
An SMM agency manages 30 client accounts: posting updates, Stories, Reels, responding to comments, and analyzing statistics. Each account spends 10-20 minutes of activity per day, resulting in about 80-150 MB of traffic (considering image uploads during posting and viewing competitors' feeds).
Traffic calculation: 30 accounts × 120 MB × 30 days ≈ 108 GB per month. IP calculation: 30 permanent addresses tied to accounts for the entire month. Here, the choice depends on the stability of the number of accounts. If the agency regularly gains and loses clients (accounts fluctuate between 20 and 45), payment by GB is more profitable — you pay for actual usage. If the client list is stable for several months ahead, fixed payment by IP is more predictable and often cheaper per account.
An important nuance for SMM: in anti-detect browsers (Dolphin Anty, GoLogin, Multilogin, Octo Browser), proxies are usually tied to the profile once and used for a long time, so it is physically more profitable to assign a permanent IP to each client account — this reduces the risk that Instagram or TikTok will see a "jumping" address and suspect multi-accounting. Here, residential proxies with session binding, paid for with a traffic package calculated with a 20-30% buffer, are more commonly used.
Calculation for E-commerce: Scraping Prices on Wildberries and Ozon
A seller monitors prices of 500 competitor products on Wildberries every 30 minutes, which means 48 checks per day per product. One request for a product card without images is 0.1-0.3 MB (HTML/JSON response). Calculation: 500 products × 48 requests × 0.2 MB ≈ 4.8 GB per day, or about 144 GB per month.
For such a task, payment by GB is almost always more profitable because scraping involves many short requests with minimal weight, rather than long sessions with video watching. It makes no sense to overpay for fixed IPs that will be idle between checks. For this scenario, data center proxies are excellent — they are cheaper and faster at processing many parallel requests, which is critical when monitoring hundreds of SKUs simultaneously.
An exception is if Wildberries or Ozon starts banning data center IPs in batches due to aggressive request frequency. In this case, it becomes necessary to switch to residential proxies with payment by traffic, but with increased intervals between requests (once every 2-3 hours instead of 30 minutes), to reduce overall GB consumption while avoiding blocks.
Calculation for Marketers: Testing Ads from Different Regions
A marketer tests how Google Ads and Yandex.Direct ads appear in 15 different countries to check localization and bids. The task is short-term: a few hours of work per day for 1-2 weeks, without a constant binding of accounts to specific IPs.
Here, the logic is the opposite of the previous cases: payment by GB is more profitable because the number of regions and sessions is unstable, while the traffic volume per session is predictably small (loading several pages of results and ads — 5-15 MB per check). It is irrational to overpay for 15 permanent IPs for the entire month, when the actual work is done for 2 weeks at 2-3 hours per day.
However, if testing is regular — for example, weekly checks of positions and ads in the same 15 countries over a quarter — then a stable pool of IPs for a long term starts to save budget due to a fixed price, rather than hourly/per megabyte billing.
Checklist: How to Choose a Payment Model for Your Task
- Calculate traffic for one session: if less than 50 MB per account per day — consider payment by GB
- If using video content (TikTok, Reels, YouTube) — payment by IP is almost always more profitable due to traffic volume
- If the number of accounts is unstable month to month — choose payment by traffic, do not overpay for idle IPs
- If managing the same accounts for more than 2-3 months in a row — fixed payment by IP reduces the risks of address changes and chain bans
- If scraping large volumes of text data (prices, stock levels, product cards) — payment by GB is almost always cheaper
- If maximum similarity to a live user is needed (warming up, SMM) — residential or mobile proxies with a traffic buffer of 20-30%
- If the task is short-term and one-off (ad testing, one-time scraping) — payment by GB without long-term commitments
Conclusion
There is no universal answer to "which is more profitable" — it all depends on the structure of your activity. One-off and unpredictable tasks (ad testing, unstable number of accounts, text scraping) are cheaper to pay for by traffic. Permanent, long-term scenarios with high load on one address (farming ad accounts, SMM multi-accounting with video content, warming up TikTok Ads) are almost always more profitable to switch to fixed payment by IP.
If your task involves mass scraping of marketplaces or testing ads from different regions, consider data center proxies with payment by traffic. If you are managing multi-accounting on social media or farming ad accounts with full warming, we recommend residential proxies or mobile proxies with a fixed pool of IPs — they will provide stability and reduce the risk of blocks over the long term.