Testing Method

Cross Shore Tools reviews global SaaS products from the perspective of Japanese users and Japanese companies. This page explains how we test, what we measure, and what we will not do. If a review on this site says something, this page is how we got there.

1. We use the product ourselves

Every review here is based on hands-on use. We create our own account, run the product through real work, and record what happens — including the parts that go wrong. We do not publish a review of a product we have not used.

2. Japanese is the core test

Most SaaS products are designed and tested in English first. A tool can work flawlessly in English and still fall apart in Japanese: broken spacing, mis-segmented words, speaker confusion, mishandled names, or an interface that gives a Japanese beginner no way in.

So we test in Japanese. Japanese speech, Japanese text, Japanese names, Japanese meeting conventions. Where a product also matters in English, we note the gap between the two.

3. Japan Readiness Score

Each product receives a Japan Readiness Score out of 100. The current version is 0.1.

CriterionPoints
Japanese Language Performance25
Setup & Onboarding15
Core Product Usability15
AI Output Quality15
Integrations10
Japanese Business Fit10
Pricing & Billing5
Beginner Friendliness5

Every criterion is published with three things: the score, the evidence behind it, and a short comment explaining the judgement. A score without evidence is not a score.

4. Evidence rules

  • Scores come from tests we ran, not from vendor marketing.
  • A vendor claim such as “Japanese supported” is treated as a claim until we verify it ourselves.
  • We state the conditions of each test: device, environment, number of speakers, length, and language.
  • We publish results that are bad for a product exactly as readily as results that are good.
  • We never publish confidential material from any meeting or document used in testing.

5. What we do not do

  • We do not raise a score because a product pays a higher affiliate commission.
  • We do not generate reviews with AI in place of testing.
  • We do not describe untested features as tested.
  • We do not blend vendor documentation into our own findings without saying which is which.

6. This method will change

Version 0.1 is a starting point. As we test more categories, some criteria will prove too blunt and others too generous. When we change the method, we will say so on this page rather than revise scores silently.

Last updated: August 25, 2026.