How we evaluate tools and apps.
Every YTC score is earned through hands-on testing, independent research, a published framework, and rules that stop one strong area from hiding a serious problem. Never influenced by advertisers or affiliate commissions, and never for sale to companies that commission our research.
Four principles behind every score
Our scoring system exists to give you a fast, trustworthy signal, not to manufacture a number that makes a product look good.
Independently evaluated
Scores rest on our own hands-on testing first, then on verified data, official documentation, and third-party benchmarks. No score is based solely on press releases or vendor marketing claims. Where we rely on someone else’s evidence instead of our own testing, the review says so.
Weighted, not averaged
Different dimensions matter differently. A tool’s output quality carries more weight than its onboarding experience. Our weights reflect what actually matters to real users making real decisions.
Serious problems can’t be averaged away
A tool can be excellent at its job and still mishandle your data. Our score caps make sure a serious failure in quality, privacy, or integrity limits the final score, whatever the other dimensions say.
Africa-aware scoring
Every score includes an African Accessibility dimension, a differentiator unique to Your Tech Compass. A tool that works beautifully in San Francisco but fails on Nairobi mobile data scores lower here.
One score. Five dimensions.
Each dimension is scored 0 to 10 in half-points, multiplied by its weight, and summed to a maximum of 100. Totals are rounded to the nearest whole number, and three score caps can then apply. Always read the dimension scores in a review, not just the total.
YTC Tool Intelligence Score
Applied to AI tools, productivity apps, software platforms, and consumer technology products across AI Unboxed and Apps & Tools. Fintech apps, wallets, lenders, and remittance services are scored under a separate framework: see How We Rate African FinTech.
What we measure and how
Open each dimension to see what we check, the minimum evidence behind it, and what each score range means.
Extra checks by product type
These inform the dimension scores above. They are not separate scores.
AI assistants and chatbots
Accuracy on questions with verifiable answers, source citation behavior, handling of long or complex context, refusal behavior, and whether your conversations are used for training.
Creative and generative tools
Output consistency across repeated prompts, commercial-use and licensing terms, watermarks and export limits, and how well the tool follows detailed instructions.
Productivity and workflow apps
Sync reliability, integration quality, offline behavior, and how easily you can export your data and leave.
Three rules that protect readers
Caps apply to the final total, whatever the other dimensions say. If more than one applies, the lowest is used; caps are never added together. Both editors review the evidence first, and every cap is stated openly in the review. There is no cap for African accessibility: limited availability lowers the score through its 10% weight, but it doesn’t make a product unsafe or unreliable.
Quality cap: 54
Applies if the Output quality score is below 5. It caps the total at 54 (Below average).
A tool that is unreliable at its core job cannot score well because it is pleasant to use and cheap.
Privacy cap: 69
Applies if the Privacy & trust score is below 5, for example because of training on user data with no meaningful opt-out, or materially misleading privacy claims. It caps the total at 69 (Average).
Such a product cannot be rated Good or Excellent.
Integrity cap: 54
Applies when a serious security incident, undisclosed breach, or deceptive practice has occurred in the last 36 months. It caps the total at 54 (Below average).
It is lifted when the company has resolved the matter and we have verified the remediation.
See how the score is built
Adjust each dimension from 0 to 10, try a preset, or switch caps on and off to see how the final score is calculated. Dragging Output quality or Privacy & trust below 5 ticks its cap automatically.
Final YTC Score
What the numbers mean
Scores are honest signals, not marketing tools. A score of 70 is genuinely good; it means the product is worth your time. Scores within about 3 points of each other are broadly equivalent.
How we keep scores consistent
Half-points and rounding
Dimensions are scored in half-points. A score at the top of a band means all criteria are met with margin; the bottom means at least one is met only narrowly. Totals are rounded to the nearest whole number.
Minimum evidence
At least 10 representative tasks across three or more task types, in two sessions on different days, for Output quality. A full setup from a fresh account for Ease of use, and prices verified on the review date for Value. Each review states its test counts, the version tested, and the date. Features, plans, or markets we could not test are labeled “claimed, not verified”.
From first login to published score
Every score follows the same rigorous path. No shortcuts. No exceptions.
Hands-on testing
We use the product ourselves, then gather official documentation, third-party benchmarks, and community feedback. Each review states how many tests we ran and which version we tested.
Verification
Every claim is cross-checked against multiple independent sources.
Independent dimension scoring
Each editor scores every dimension on their own, before discussion. Dimensions are then weighted and summed.
Editorial review
Both Oscar and Diana review the evidence. Any dimension where our scores differ by 2 points or more is re-reviewed with the evidence side by side before publication.
Right of reply
Before we publish any dimension score below 5, apply a cap, or make a factual allegation of misconduct, deception, or hidden charges, the company has 7 days to respond. Material corrections are reflected in the score, and we note if a company declined to respond.
Living score
Scores are re-verified at least every 6 months, and sooner after a major release, pricing change, privacy-policy change, reported breach, significant outage, or ownership change. Every update is date-stamped, and reviews not re-verified within 12 months are labeled as possibly out of date.
No scores for sale
Advertising, affiliates, and free access
Scores are never influenced by advertising relationships, affiliate commissions, or sponsored partnerships. If a product we have reviewed later advertises on Your Tech Compass, its score is not revisited unless the product itself has materially changed.
Free access to a product does not guarantee a favorable score. Our framework applies identically however we obtained access. When vendor-provided information is used, it is cross-referenced against third-party sources and clearly attributed.
If you believe a score is incorrect or outdated, contact us. We take corrections seriously and respond within 48 hours.
Commissioned research
Some companies pay us to research and publish an agreed piece (see Work with Us). Payment covers research and production. It never buys a score, a verdict, or a positive outcome.
Commissioned pieces are labeled as commissioned at the top of the page. The same framework, evidence minimums, caps, and two reviewers apply as for any other product. Any narrative advice happens before testing begins and does not touch the scoring.
The company gets the same right of reply as anyone else: factual corrections backed by evidence are reflected in the score, and a score is never negotiated. We only take on products we believe can stand up to testing; if testing shows we were wrong, the score says so.
A company may withdraw a commissioned piece before publication. A fee applies, and it grows as the work progresses. A withdrawn piece is never published. We may still cover the product independently, using only our own testing and public information, never the company’s confidential material.
What a score can’t tell you
A score is a snapshot. Our tests use limited samples, AI tools change quickly, and we may not be able to test every feature, plan, or market a product claims. A good score is not a guarantee, and products can change after we publish.
The YTC Tool Intelligence Score is editorial opinion, not professional, legal, or financial advice. Check a provider’s current terms before relying on any tool for sensitive work.
See our scores in action
Browse our AI Unboxed and Apps & Tools reviews to see the YTC scoring system applied to real products.
Browse AI Unboxed → Apps & Tools reviews FinTech methodology