How we review

We follow a 7-step process across categories. We pay for products, run quantitative tests, do hands-on use, add clinician review where needed, and score independently. We revisit results on a schedule and when products change. We disclose how we fund our work and how we keep it separate from rankings.

1. Real subscriptions

We buy the same plans you can buy, using our own credit cards and identities. No press accounts. No vendor-controlled demos. We track onboarding, billing, renewals, refunds, and cancellation friction. Examples: Therapy — 11 platforms reviewed with 22 paid sessions; GLP‑1 weight‑loss programs — 9 programs enrolled using 3 distinct test identities to check access and pricing differences.

2. Quantitative testing

We collect repeatable, numeric data and keep the raw measurements. For VPNs, that means download/upload, latency, jitter, and leak checks; we’ve logged 3,400+ speed tests across providers. We define loss and improvement clearly and show the math. Example: on a 200 Mbps line, a 12% speed loss is still 176 Mbps. Where stable automation is possible, we run tests for multiple weeks to reduce noise and catch regressions.

3. Qualitative review

Numbers do not capture everything. We live with each product to assess reliability, setup, usability, and support. We note dark patterns, forced upsells, privacy surprises, and accessibility gaps. Examples: Therapy — 22 therapist sessions documented for intake clarity, wait times, and fit; onboarding scripts and cancellation flows archived with timestamps and screenshots.

4. Clinical review where applicable

When products touch health, licensed clinicians review medical claims, safety guidance, and care models. A board‑certified obesity‑medicine physician and a registered pharmacist review GLP‑1 content; a licensed clinical psychologist reviews therapy content. We check dosing protocols, contraindications, lab requirements, and follow‑up cadence against guidelines. For compounded medications, we confirm the pharmacy’s 503A status and that any compounding aligns with FDA shortage allowances; we flag when a drug is not on shortage. We call out insurance explicitly and note when programs are in‑network or cash‑pay only. Crisis content references the 988 Suicide & Crisis Lifeline and directs readers there when appropriate.

5. Independent scoring

We use category‑specific rubrics with weighted factors set before testing. Examples: VPN scoring weights performance, privacy, security audits, app quality, and streaming unlock; therapy weights clinical oversight, access (response time and scheduling), price transparency, and therapist credentials; GLP‑1 programs weigh clinical screening, medication access, total monthly cost (medication plus program), and follow‑up care. We convert raw measures to normalized scores, apply weights, and produce a composite. If a brand makes a claim we could not verify, we either exclude it from scoring or label it clearly as “the brand claims” in our copy.

6. Re‑evaluation cadence

Products change. We re‑test on a cadence and when major updates land. Fast‑moving software (like VPNs) gets periodic spot checks and scheduled re‑runs; longer‑cycle services (like therapy platforms) get freshness checks on network size, wait times, and pricing. Health programs are monitored for medication availability, pricing shifts, and policy changes that affect access. We also trigger out‑of‑cycle reviews for security incidents, feature removals, price hikes, or reader‑reported issues with evidence.

7. Commercial transparency

Editorial and commercial are separate. Our writers and testers decide rankings and scores; sales does not. We may use affiliate links, which can earn a commission if you buy, but they do not change our scores or placement. We do not sell placements, accept paid “review fees,” or let brands preview or edit our findings. We decline gifts and “lifetime” accounts that normal buyers cannot get. For full details on how we make money and manage conflicts, see our Disclosure page at /disclosure.

Updated May 15, 2026.