A/B Testing Hosting Cost Formula
Learn how to estimate monthly A/B testing hosting cost from experiment traffic, page transfer size, included bandwidth, overage pricing and fixed fees.
This calculation estimates the monthly infrastructure cost attributable to an A/B test. It converts test participant activity into estimated bandwidth use, applies any charge above the included allowance, and adds the recurring platform or hosting fee.
- 100% Free
- No Sign-Up Required
- Private & Secure
- Mobile Friendly
Estimated Monthly A/B Testing Hosting Cost
Where:
Start with the fixed monthly fee. Estimate the test's data transfer, subtract the included bandwidth, charge for any positive excess, and add that overage charge to the fixed fee.
Variables Explained
| Variable | What It Means | Unit |
|---|---|---|
| monthlyPlatformFee - Monthly testing or hosting fee | The recurring fixed charge for the testing platform, hosting service or edge delivery service. | currency |
| monthlyVisitors - Monthly site visitors | The estimated number of visitors to the site during a typical month. | visitors |
| testTrafficPercent - Traffic included in the test | The percentage of total monthly visitors assigned to the A/B test. | percent |
| pageviewsPerTestVisitor - Pageviews per test visitor | The average number of tested page loads generated by each experiment participant during the month. | pageviews |
| averagePageSizeMb - Average tested page size | The average amount of transferred data for one tested page load, including relevant assets. | MB |
| includedBandwidthGb - Included monthly bandwidth | The bandwidth allowance available before overage charges apply. | GB |
| overageRate - Bandwidth overage rate | The charge for each GB used above the included bandwidth allowance. | currency |
Step-by-Step Calculation
Estimate experiment participants
Multiply total monthly visitors by the percentage allocated to the experiment.
testVisitors = monthlyVisitors * (testTrafficPercent / 100)
Estimate tested pageviews
Multiply test participants by their expected tested page loads.
testedPageviews = testVisitors * pageviewsPerTestVisitor
Convert page transfer into bandwidth
Multiply tested pageviews by average page size in MB, then divide by 1024 to estimate GB.
testBandwidthGb = (testedPageviews * averagePageSizeMb) / 1024
Find bandwidth above the allowance
Only usage above the included allowance is treated as overage.
bandwidthOverageGb = max(0, testBandwidthGb - includedBandwidthGb)
Calculate the variable bandwidth charge
Multiply excess bandwidth by the provider's per-GB overage price.
bandwidthOverageCost = bandwidthOverageGb * overageRate
Calculate the monthly hosting estimate
Add the fixed recurring fee to the estimated bandwidth overage charge.
monthlyHostingCost = monthlyPlatformFee + bandwidthOverageCost
Example: 100,000 visitors and a 50% A/B test allocation
Test visitors
100,000 × (50 / 100)
50,000 visitors
Tested pageviews
50,000 × 3
150,000 pageviews
Test bandwidth
(150,000 × 2 MB) / 1,024
292.97 GB
Bandwidth overage
max(0, 292.97 − 100)
192.97 GB
Overage charge
192.97 × $0.09
$17.37
Monthly hosting cost
$49.00 + $17.37
$66.37
Final Result
Estimated monthly A/B testing hosting cost: $66.37. The estimated cost per test visitor is about $0.0013.
Assumptions
- ✓Each participant is assigned to one variation rather than loading every variation.
- ✓The average page size represents transferred data per tested page load under typical caching behavior.
- ✓The entered bandwidth allowance and overage rate apply to the experiment traffic being budgeted.
- ✓A GB is estimated as 1,024 MB in this calculation.
- ✓The fixed monthly fee is charged regardless of the amount of experiment traffic.
Limitations
- !Actual provider billing may use decimal GB, GiB, rounding rules or measurement intervals that differ from this estimate.
- !CDN delivery, origin transfer, requests, serverless execution, image processing and database activity may be priced separately.
- !Bot traffic, prefetching, retries and monitoring can increase actual transferred data.
- !A larger variation, additional testing scripts or a cache miss rate different from the estimate can change real usage.
- !The calculation does not include experiment design, development, quality assurance, analytics or staff costs.
Common Mistakes to Avoid
Entering all site traffic when only a smaller share is assigned to the A/B test.
Using the uncompressed source-file size instead of the average data actually transferred to visitors.
Forgetting assets introduced by a variation, such as images, fonts, video or third-party scripts.
Treating MB and GB as the same unit instead of converting the page transfer total.
Entering an overage rate for the wrong service when CDN and origin bandwidth have separate prices.
Assuming the result includes request, compute or analytics charges when only fixed fees and bandwidth overage are modeled.
Related Formulas
Frequently Asked Questions
How is A/B test bandwidth calculated?
Test visitors are multiplied by pageviews per visitor and average page size. The resulting MB total is divided by 1,024 to estimate GB of monthly test bandwidth.
Does an A/B test automatically double hosting bandwidth?
Usually no. In a normal split test, a participant receives one assigned variation. Bandwidth can rise if the test adds larger assets, scripts or cache misses.
Why does the formula use max(0, bandwidth minus allowance)?
It prevents a negative overage. If estimated test bandwidth stays within the included allowance, the bandwidth overage cost is zero.
What is the cost per test visitor?
It is the estimated monthly hosting cost divided by the number of experiment participants. It is a useful allocation metric, not necessarily a provider billing metric.
Should page size include cached files?
Use an estimate of the data transferred under normal visitor conditions. If browser or CDN caching reliably avoids some transfers, reflect that in the average page size where possible.
Ready to calculate your result?
Use the calculator to get instant results with your own inputs.