The Methodology
How We Score
Every AI tool gets the same 7-dimension audit. Here's the rubric — and the conflict-of-interest policy that keeps it honest.
Last updated
1. The 7-Dimension Rubric
Each tool earns a 0–10 score on every dimension. Weights below produce the BestStacked Score you see on every review.
| Dimension | Weight | What we measure |
|---|---|---|
| Output Quality | 25% | Accuracy, depth, voice — does it produce work you'd ship? |
| Speed | 10% | Response time, throughput, latency under load. |
| Ease of Use | 15% | Onboarding, UX, time to first value. |
| Pricing Value | 15% | Price-to-output ratio, transparency, hidden fees. |
| Integrations | 10% | Native, no-code, API quality. |
| Reliability | 10% | Uptime, error handling, support responsiveness. |
| Future-Proof | 15% | Roadmap velocity, funding, ecosystem momentum. |
2. Score Bands
Best in class
9.0 – 10.0
Strong choice
8.0 – 8.9
Good with caveats
7.0 – 7.9
Skip
Below 7.0
3. How We Test
Every tool goes through a minimum 10–20 hours of hands-on testing across real-world projects — not vendor demos, not 20-minute fly-bys. We sign up like any other customer (no press accounts, no review-only seats), set up the workflows our readers actually run, and use the tool inside a working creator/founder/marketer stack for at least two weeks before scoring it.
We test against documented use cases: ship a real piece of work with each tool — a published newsletter, a deployed app, a recorded podcast, a launched campaign — and measure output quality, time-to-result, error rates, and how often we hit the wall on the free or starter plan. We benchmark every tool against the two or three most credible alternatives in the same category, and we re-run pricing math against the vendor's live pricing page on the day we publish. If we can't get a tool to do the thing it claims it does, that lands in the cons section in the reviewer's exact words — not a softened paraphrase.
4. Conflict-of-Interest Policy
Affiliate revenue does not influence scores, rankings, or what we choose to cover. Vendors cannot pay for placement, cannot edit drafts, and never see scores before publication. We will publish a negative review of a tool we earn commission on, and we have.
Every page that contains affiliate links carries an inline disclosure, and our full affiliate disclosure details exactly which relationships exist. If a vendor sponsors content elsewhere on the internet (newsletter, podcast), it is disclosed there too — and it still does not buy them a higher score here.
5. Update Cadence
- Every review is re-audited quarterly at minimum.
- Pricing is reviewed monthly against the vendor's live pricing page.
- Major product launches or pricing changes trigger an immediate re-review.
- The "Last updated" date in every byline reflects the most recent material edit.
Section 6 · The Author
Who Reviews
Marketing thought leader, LinkedIn Top Voice, bestselling author of The End of Marketing, and Editor-in-Chief of BestStacked.com (AI tools) and BestSends.com (email infrastructure). Carlos personally tests every tool that earns a BestStacked Score, and his name is on every review.
Read more about Carlos →