SimpleVote is used for binding elections. This page summarises load tests we have run on STV counts, voter registration, and email sends. Current upload and email limits are shown clearly where they still apply.
Benchmark results
Last updated May 2026
| Test | Scale | Duration | Status |
|---|---|---|---|
STV calculate (single position) Production platform | 50,000 ranked ballots | 12.3 s | Stable 12 candidates, 5 seats |
STV calculate (multi-ballot) Production platform | 80,000 cast rows, 4 positions | 19.0 s | Stable Four ballots in one election |
STV calculate (stress test) Staging environment | 50,000 ranked ballots | 13.3 s | Stable Includes database fetch time |
Voter register Staging environment | 4,000 invitees | 5.2 s | Stable Batch upload, no email send |
STV engine (CPU only) Isolated benchmark | 50,000 ranked ballots | 58 ms | Stable 12 candidates, 5 seats |
Scottish Rule 51 tie-break Isolated benchmark | 20,000 ranked ballots | 9 ms | Stable Verified against reference implementation |
Magic-link invitation send Per-request cap today | 2,000 recipients | - | Interim cap Higher volumes send automatically in batches |
CSV voter confirm upload Per-request cap today | 500 invitees | - | Interim cap 4,000 in one action is the Tier 2 target |
Magic-link invitation send Planned scale test | 50,000 recipients | - | Not yet tested Scheduled before general availability at this size |
What we guarantee today
These are the scales we design and test against. Interim caps on upload and email (500 voters and 2,000 emails per action) are safety limits while background processing scales up. They are not the long-term promise.
Voter upload
Tier 2 guarantee: 4,000 per action
Next target: 100,000 per action
Email send
Tier 2 guarantee: 4,000 one click
Next target: 100,000 one click
STV per position
Tier 2 guarantee: 4,000 ranked ballots
Next target: 50,000 ranked ballots
Ballot positions
Tier 2 guarantee: 15 positions
Next target: 15 positions
Total cast rows
Tier 2 guarantee: 60,000 rows
Next target: 600,000 rows
How tests are run
- STV engine benchmarks measure pure count speed at large ballot sizes.
- End-to-end tests use realistic voter lists and full calculate runs on staging environments.
- Email sends are tested in controlled batches before we raise public limits.
Planning an unusually large election? Contact us before you schedule open voting. We will confirm your tier, any temporary limits, and a realistic timeline.
Step-by-step guides
For instructions on running a vote, see the ERO Guide. For product walkthroughs, see How it works on the home page.