Our 8-Point Review Methodology
Last verified: recentThe published testing protocol behind every VaultPremium flagship review. We publish this so readers can audit our work — and so the testing process is reproducible by anyone who wants to verify our findings independently.
The 8 Points
1. Testing Window — 14 Days Minimum
Every flagship review is based on a minimum 14-day active subscription window. Short-window reviews miss release cadence, support behavior, and rebill timing — so we don't publish them.
2. Personal Card — No Comp Accounts
The editor pays for the subscription using a personal credit card. We do not accept review accounts, comp passes, or pre-arranged access. The reason: comp accounts may receive different treatment than real subscribers, and they make billing descriptor and cancellation testing impossible.
3. Device Matrix — 3 Devices
Every test is run on three devices to capture cross-platform performance:
- Desktop: Chrome on Windows 11
- iPhone 15 Pro: Safari on iOS
- Pixel 8: Chrome on Android
4. Connection Matrix — 3 Connection Types
Streaming performance varies dramatically with bandwidth. We test on:
- 100 Mbps symmetric fiber (best case)
- 25 Mbps DSL (typical home)
- LTE mobile (out-of-home)
5. Sample Size — 30+ Scenes Streamed, 10+ Downloaded
Minimum 30 scenes streamed end-to-end across all sub-sites/categories. Minimum 10 scenes downloaded in full. Subset re-streamed across connection types for buffering comparison.
6. Resolution & Bitrate — MediaInfo Verified
The most common deception in premium adult video is relabeled resolution — a 1080p stream tagged "4K" in the player. We verify peak resolution and bitrate using MediaInfo against actual downloaded files, not the player's UI label. If a network claims 4K and delivers 1080p, our review will say so.
7. Billing Descriptor — Pulled From Real Statement
After the first billing cycle completes, the editor pulls the actual descriptor from the credit card issuer's statement (online or paper). We record exactly what appears — verbatim — so readers know what their own statement will show.
8. Cancellation — Timed & Screenshotted
At the end of every test window, the editor cancels the subscription, times the cancellation flow with a stopwatch, screenshots every step, and logs the confirmation email. The cancellation experience is the most reliable trust signal in the category.
What We Measure & Publish
| Metric | Tool / Method | Why It Matters |
|---|---|---|
| Peak resolution | MediaInfo | Verifies actual file resolution, not player label |
| Peak bitrate | MediaInfo | Distinguishes true 4K from upscaled 1080p |
| Time-to-first-frame | Manual stopwatch | Real streaming UX latency |
| Buffering events / 30 min | Manual count | Stream stability across connections |
| Download throughput | Browser network panel | Real-world download speed |
| Billing descriptor | Card-issuer statement | Privacy / household discretion |
| Cancellation time | Stopwatch + screenshots | Trust signal — how the network behaves when you leave |
| Support response time | Manual ticket + timestamps | Service quality signal |
What We Do NOT Do
- ❌ We do not accept paid placements, sponsored reviews, or "featured network" fees.
- ❌ We do not publish reviews of networks we haven't personally subscribed to.
- ❌ We do not anonymize our reviewers — every byline is real and contactable.
- ❌ We do not publish unverifiable claims. If we can't measure it, we say "as of writing."
- ❌ We do not delete negative findings. Networks that score badly stay scored badly.
Review Update Cadence
- Flagship reviews (PornWorks, Seduced.com): re-verified every 90 days.
- Discount pages (PornWorks, Seduced.com): re-verified every 14 days.
- Comparison pages: re-verified every 120 days.
- Trust pages ("Is X Legit?"): re-verified every 60 days.
- Informational pages: re-verified every 180 days or when industry shifts (codec, processor, regulation).
Every page carries a "Last verified" stamp under its H1.
How to Audit a Specific Claim
If you want to verify a specific finding in any review, the protocol above gives you the tools to reproduce it. If you reproduce the test and get a different result, email the editor. We will re-test and update the page if your methodology is sound.