中文 Brand login
For pet food and pet supplement brands · Efficacy validation platform

Every efficacy claim
should stand up to science

PETEFF helps brands answer three questions: What evidence does this claim need? Is the existing evidence enough? If not, which study should be run, and how large? From evidence review and study design to feeding trials, every step is traceable.

9 typesStandardized validation studies
L0–L5Six-level evidence grading
AAALACPartner laboratory accreditation
CNAS · CMA · GLPPartner laboratory quality systems

Claims need evidence, not copywriting

Functional claims for pet food are moving from marketing-driven to evidence-driven. "Rich in essence of X" convinces neither regulators nor consumers; "coat gloss score improved in a 28-day controlled feeding study" does. PETEFF grades every claim on six evidence levels and shows exactly which level is missing.

L0Formulation calculation
L1Finished-product testing
L2Published literature
L3Palatability, digestibility and similar tests
L4Controlled feeding study
L5Replication and mechanism research

Ordinary pet food may not claim to prevent or treat disease. Such claims are marked "prohibited" on the platform, and we do not design studies for them.

Validation studies

Each study pre-specifies its primary endpoint, design and statistical method. Sample sizes are computed exactly by software, and the protocol is reviewed and locked by the chief scientist before the study starts.

L3

Palatability: two-bowl test

Offered two diets side by side, which one does the pet prefer? Endpoints: intake ratio and first choice.

Typically 2 days
L3

Palatability: one-bowl test

Offered a single diet, does the pet accept it? Endpoints: daily intake as a share of food offered, and days of refusal.

Typically 3–5 days
L3

Apparent digestibility

Apparent digestibility of dry matter, crude protein, crude fat and energy, plus metabolizable energy.

Adaptation period plus fecal collection
L3

Stool quality

Are stools normal after a diet change? Standardized fecal scores and fecal dry matter.

Often combined with palatability or digestibility
L4

AAFCO feeding trial: adult maintenance

Demonstrates nutritional adequacy through a feeding trial: completion rate, body weight change and end-of-study blood values.

Typically 26 weeks
L4

Skin and coat

Blinded assessors score coat gloss and scaling; instrument gloss readings optional; photos taken under fixed lighting and angle.

Typically 8–12 weeks
L4

Weight management

Does a lower-energy formula help control body weight? Percent body weight change and body condition score.

Typically 8–16 weeks
L4

Gut microbiome

Effect on microbiome composition: diversity indices and relative abundance of target genera; short-chain fatty acids optional.

Duration set by the chief scientist
L4

Controlled feeding study (general)

Joint, urinary, oral, immune, growth, hairball and other claims, with endpoints defined by the claim.

Duration set by the chief scientist

How validation works

From request to locked protocol, brands can follow every step online. Once locked, a protocol cannot be changed; any revision becomes a new version, with hashes kept on record.

  1. Submit a validation request

    Product, control product, the claims to validate, existing evidence and target markets.

    Brand
  2. Evidence gap review

    For each claim: the evidence level required, the highest level already held, a conclusion and the recommended study type.

    Platform + nutrition consultant
  3. Study design

    Exact sample size (noncentral t distribution); AI drafts the protocol text and statistical analysis plan, with every number filled in and checked by software.

    AI study designer + nutrition consultant
  4. Review and lock

    The chief scientist locks the protocol once all pre-lock checks pass; a stratified block randomization schedule is generated at the same time.

    Chief scientist
  5. Feeding study

    Run at the partner laboratory according to the protocol, with blinded packaging and ethics approval in place before the start.

    Partner laboratory
  6. Analysis and report

    Analysis follows the pre-specified statistical plan, and a validation report is issued.

    In development

Laboratory and accreditation

Co-founding technology partner Huateng Biomedical provides pharmaceutical-grade animal facilities and quality systems. PETEFF is responsible for study design, data management, statistical analysis and reporting.

AAALAC accreditedLaboratory animal care and welfare program accredited by AAALAC International
CNAS / CMAChinese laboratory accreditation and inspection body certification
FDA-GLP / OECD-GLPGood Laboratory Practice quality systems: traceable data that withstands inspection
SPF animal facilitySpecific-pathogen-free laboratory animal facility
200+ large-animal modelsGene-edited minipigs, rabbits and other large-animal models, physiologically closer to dogs and cats
10,000 m² laboratoryMulti-platform technology center that has served more than 40 pharmaceutical companies

Report independence is written into our governance: no shareholder may influence the conclusion of any study.

Beyond validation: formulation and compliance in one place

Get the formula and the label right before you validate. The platform's formulation and compliance engine is already in use.

Nutrient calculation

Dry matter, as-fed and per 1000 kcal bases, with data coverage shown for every nutrient; unknown values are never treated as zero.

Standards comparison

Checked against AAFCO dog and cat nutrient profiles and Chinese national standards: meets, fails, needs review or no data.

Formula optimization

Linear programming returns four candidate formulas: lowest cost, robust margin, high animal protein and fewest ingredients. If no solution exists, it shows which constraint to relax.

Label and claim checks

Guaranteed analysis, ingredient list, naming thresholds, and the risk and required evidence level of each written claim.

Independent audit

Nutrient levels and costs are recomputed independently with integer arithmetic; any mismatch blocks release.

Human sign-off

AI and software only produce drafts. Formulas and data are signed off by a nutrition consultant, and every action is written to a hash-chained audit log.

FAQ

What does PETEFF do?

PETEFF assesses what evidence a pet food or supplement claim needs, designs and organizes feeding studies, and checks formulation nutrients and label compliance.

Where are the studies conducted?

Feeding studies are carried out at the laboratory of co-founding technology partner Huateng Biomedical. PETEFF handles study design, data management, statistical analysis and reporting.

What does AI do in a study?

AI drafts the protocol text. Sample sizes, randomization and every other number are computed and checked by software, and the protocol is locked by the chief scientist before the study starts. Locked protocols cannot be changed.

Can you validate disease prevention or treatment claims?

No. Ordinary pet food may not claim to prevent or treat disease; the platform marks such claims as prohibited.

How is sample size determined?

From the pre-set significance level, power, expected difference and variability, computed exactly with the noncentral t distribution and then inflated for expected dropout. The calculation sheet is exported with the protocol.

How do we get started?

Brands receive an account from the platform, log in and submit a validation request under "Efficacy validation". The platform interface is currently in Chinese.

Brands with an account can submit a validation request today

After logging in, create a request under "Efficacy validation" and the platform will show your evidence gaps and recommended studies.

Brand login