Evaluation protocol
Evaluation Methodology 2026
A shared rubric so Multilogin vs alternatives comparisons stay auditable across operators and time.
Purpose
This methodology exists so partner content on MultiAccountManager stays comparable: same weights, same evidence artifacts, same go/no-go language. Browse the full Guides hub or jump to the compare hub when you are ready to apply the rubric.
Scoring weights
| Pillar | Weight | What we measure |
|---|---|---|
| Profile integrity | 40% | Drift across repeated sessions, restore success, isolation |
| API reliability | 25% | Start/stop latency, error rates, automation attach stability |
| Governance & handoff | 20% | Roles, token hygiene, recovery playbooks |
| Cost of failure | 15% | Blast radius when a profile or batch fails |
Scores are 0–10 per pillar, then weighted. Publish the date, environment notes, and proxy class with every score.
Evidence pack (minimum)
- Profile IDs used in the pilot
- Proxy provider / exit country (no secrets)
- Repeat-run count and critical drift flag
- API trace IDs for failed starts
- Rollback steps that actually worked
Store packs outside chat logs. Chat is not an archive.
Affiliate-safe buying rule
This site is commercial, so the methodology deliberately slows the purchase moment down:
- Validate the stack first.
- Score the environment second.
- Buy the smallest correct plan third.
- Only then apply the matching Multilogin code.
That protects both the reader and the site from low-trust “buy first, diagnose later” content patterns.
Decision bands
| Band | Meaning |
|---|---|
| Ready | Controls support controlled rollout |
| Pilot-Only | Bound spend and volume until penalties clear |
| No Annual Commitment | Stay monthly / trial until stability improves |
| No-Go | Do not scale procurement |
Map interactive scores via the fingerprint readiness tool.
How we use Multilogin in tests
Multilogin is our primary reference stack as a partner publication. That does not auto-win comparisons: Dolphin Anty or others can win a pilot when evidence says so. When Multilogin wins, checkout uses partner codes 50MAM / MAMOFF via /go/multilogin.
Change control
Material rubric changes are logged in the methodology changelog with impact notes on previously published scores.