A/B Testing Database Storage Calculator Examples
Worked A/B test storage examples for small, high-traffic, and event-heavy experiment scenarios.
These examples show how traffic, duration, event volume, record size, overhead, and replication affect an A/B testing database storage estimate.
Small product experiment
A product team runs a 14-day two-variant experiment for a feature page.
Input Summary
Daily visitors
10,000
Duration
14 days
Events per visitor
3
Event size
0.50 KB
Assignment size
0.15 KB
Overhead and replicas
20%; 1 copy
Calculation Breakdown
- 1Visitors10,000 × 14140,000
- 2Events140,000 × 3420,000
- 3Raw data(420,000 × 0.50 + 140,000 × 0.15) ÷ 1,048,5760.22 GB
- 4Provisioned storage0.22 × 1.20 × 10.26 GB
Result Summary
Provisioned storage
0.26 GB
A/B Testing Database Storage Calculator
Estimated storage is about 0.26 GB.
High-traffic marketing experiment
A marketing site tests landing-page messaging for 100,000 visitors per day.
Input Summary
Daily visitors
100,000
Duration
30 days
Events per visitor
5
Event size
0.75 KB
Assignment size
0.20 KB
Overhead and replicas
25%; 2 copies
Calculation Breakdown
- 1Visitors100,000 × 303,000,000
- 2Events3,000,000 × 515,000,000
- 3Raw data(15,000,000 × 0.75 + 3,000,000 × 0.20) ÷ 1,048,57611.30 GB
- 4Provisioned storage11.30 × 1.25 × 228.25 GB
Result Summary
Provisioned storage
28.25 GB
A/B Testing Database Storage Calculator
Estimated storage is about 28.25 GB across replicas.
Event-heavy mobile experiment
A mobile team tracks many interaction events during a 60-day onboarding experiment.
Input Summary
Daily visitors
50,000
Duration
60 days
Events per visitor
20
Event size
1.50 KB
Assignment size
0.25 KB
Overhead and replicas
30%; 3 copies
Calculation Breakdown
- 1Visitors50,000 × 603,000,000
- 2Events3,000,000 × 2060,000,000
- 3Raw data(60,000,000 × 1.50 + 3,000,000 × 0.25) ÷ 1,048,57686.55 GB
- 4Provisioned storage86.55 × 1.30 × 3337.55 GB
Result Summary
Provisioned storage
337.55 GB
A/B Testing Database Storage Calculator
Estimated storage is about 337.55 GB across replicas.
How to Read Your Results
Raw data storage is the estimated event and assignment volume before database overhead and replication.
Estimated database storage is the total across all configured copies, not the size of one primary database.
Average daily storage growth divides the provisioned estimate by experiment duration.
Compare scenarios by changing one input at a time to identify the largest storage driver.
Assumptions & Important Notes
- Daily visitor counts are treated as new experiment assignments.
- All visitors generate the stated average number of stored events.
- Examples exclude backups, logs, data warehouse exports, and retention after the experiment.
Related Examples
Frequently Asked Questions
Which input usually has the largest effect on storage?
Event count per visitor and average event-record size commonly have the largest effect because they apply to every stored event.
Can I use these examples for several simultaneous experiments?
Use them as a starting point, then estimate each experiment separately and combine the results while accounting for shared events or shared storage.
Why can a mobile experiment use more storage than a web experiment with similar traffic?
Mobile instrumentation may retain more interaction events or larger property payloads, but actual use depends on the tracking design.
Ready to calculate your own result?
Use the live calculator with your own inputs, timing, and preferences.