How we grade
The methodology, in the open
Every grade on this site can be traced to this page. If a rule is not written here, it is not part of the score.
A grade you cannot audit is just an opinion with better typography.
Two numbers, never blended
Test Grade versus Fit Score
Test Grade
An objective editorial evaluation of the test itself, scored against the published criteria below. It is the same for every visitor.
Expressed as a letter, a number out of 100, and a plain-language verdict, always with a confidence level and a verification date.
Fit Score
A personalized relevance estimate built from your stated goal, budget, state, sample preference, and how much support you want.
A high Fit Score never makes a weak test strong. It only tells you which acceptable test matches your situation best.
The rubric
Eight dimensions and their weights
Weights reflect how much each dimension changes a buying decision. Bars pair with printed percentages, never color alone.
- Measurement quality18%
- Target relevance15%
- Coverage14%
- Interpretability13%
- Transparency12%
- Practical access10%
- Value10%
- Privacy8%
Missing data. A dimension with no usable evidence scores at the category floor, not the average, and drops the record one confidence level. Absence of proof never earns points.
How sure we are
Confidence levels
Primary documents, provider confirmation, and at least one verified experience all agree.
Documents are solid but confirmation or experience is missing, or sources disagree on minor points.
Material gaps remain. The record says so plainly and the grade carries a visible caution.
What counts as evidence
Source hierarchy
- Primary lab documentation: validation summaries, detection-limit tables, method papers.
- Written provider confirmations with a named contact and a date.
- Firsthand Labbrary purchase and completion records.
- Verified customer experience with artifacts on file.
- Public provider marketing pages, used only when nothing above contradicts them.
Labels you will see
Four evidence states
Document Verified
A claim traced to primary documentation that Labbrary has read and archived.
Provider Confirmed
A claim confirmed in writing by the provider, with the confirmation dated and on file.
Customer Experienced
A verified customer walked the flow and shared artifacts Labbrary checked.
Labbrary Purchased
Labbrary bought and completed the test itself. The strongest state a record can carry.
Policies with teeth
Independence, corrections, and stale data
Commissions are handled by people who never touch grading. Graders do not see revenue data. A provider ending or starting an affiliate relationship triggers no grade review, and this policy is itself audited on a schedule.
Anyone can flag an error. Confirmed errors are corrected on the record with a dated note that explains what changed and why. Grade changes keep their full history visible. Nothing is silently edited.
Every record shows a last-verified date. Past a category-specific window, the record is flagged like this: Needs reverification. Stale records stay published with the flag, because hiding them would hide the uncertainty too.
Documentation, method class, coverage, stated pricing, privacy policy, and ordering flow can all be evaluated from documents and confirmations. Report quality and fulfillment reality cannot. Records without a purchase say so and cap their confidence accordingly.
Without firsthand or verified customer evidence, Labbrary will not describe report quality, support responsiveness, turnaround reality, or collection experience. And with any evidence at all, it will never claim a test diagnoses a condition or that a result should change your medical care.
Worked example
Scoring walkthrough with demo data
The Demo Mycotoxin Panel R11 record, computed step by step. Dimension scores multiply by their weights, sum to 82, and map to B+.
86
18%
15.5
84
15%
12.6
72
14%
10.1
80
13%
10.4
88
12%
10.6
78
10%
7.8
76
10%
7.6
82
8%
6.6
82 / 100
100%
Grade B+
All values are demo data. Letter bands: A is 85 and above, B is 70 to 84, C is 55 to 69, D is below 55, with plus and minus thirds inside each band.
Questions
Frequently asked
No. There is no paid placement, no sponsored grade, and no fee to be included. Some outbound links may earn a commission after you choose, and that money never reaches the rubric.
The grade judges the test against public criteria, not against your experience of it. Your Fit Score can stay high for a test with a modest grade. Both numbers stay visible so you can weigh them yourself.
Yes. Grades update when documentation, methods, pricing, or verified experience changes. Every change is logged on the record with the date and the reason.
In this prototype, nothing leaves the page. In production, fit answers would compute a score in your session and would never alter a public Test Grade.
No. Labbrary evaluates tests as products. It does not diagnose conditions, interpret personal results, or recommend medical action. Bring health decisions to a qualified clinician.