How Every BowlWise Grade Is Calculated

BowlWise analyses a product from publicly available label and listing information, gathered at the moment you run a scan. For URL inputs, we fetch the actual product page and extract the ingredient list and guaranteed analysis directly from it, so the analysis is anchored to the exact product you submitted rather than a generic web search.

The inputs we gather:

• The full ingredient list, in the exact order printed on the label (order reflects pre-processing weight).
• The guaranteed analysis: protein, fat, fiber, moisture, and any other listed nutrients (ash, phosphorus, taurine where provided).
• Brand and manufacturer details: parent company, country of origin, years in business, recall history, manufacturing ownership, and sourcing transparency.
• Customer reviews and ratings from public platforms (Amazon, Chewy, Petco, Trustpilot and others where available).
• Current retail pricing and bag sizes, scoped to your region (detected from the source URL or input).

We determine whether a product is a complete and balanced diet or a complimentary/supplementary feed by checking the label for statements such as "complete and balanced", "formulated to meet the nutritional levels established by AAFCO", or "for intermittent or supplemental feeding only". This is assessed against AAFCO/NRC (US) and FEDIAF (Europe) standards for the species and life stage.

• Complete and balanced: meets minimum nutritional requirements for maintenance or growth for the stated species and life stage.
• Complimentary: a treat, topper, mixer or supplement that is not nutritionally complete on its own. We provide feeding instructions (for example, "limit to 10% of daily caloric intake").

If the label does not state it clearly, we make a best assessment and explain the basis in the report.

Every product receives five sub-scores from 0-100, plus a separate Protein Digestibility index:

• Ingredient quality - quality of protein sources, presence of whole foods, by-products or meals as the primary protein.
• Protein source - biological value, species appropriateness, named vs generic proteins.
• Processing level - how heavily processed the product is (higher = less processed). This dimension uses an explicit quantitative framework (next section).
• Biological appropriateness - how closely the food matches the species' evolutionary diet, with carbohydrate evaluated on a dry matter basis.
• Filler and additive content - fillers (corn, wheat, soy), artificial preservatives and colours, unnamed meat sources, excessive plant protein.

The Protein Digestibility index (0-100) estimates biological availability from the number and variety of animal proteins, whether they are whole/fresh/raw or rendered, named vs generic, the protein-to-carb ratio, and anti-nutritional factors.

The five sub-scores are integrated into an overall 0-100 score, which is then mapped to a letter grade. Processing level carries an explicit calculation; the remaining dimensions are assessed and weighed by the veterinary-nutritionist analyst according to the severity of the findings for the specific product and pet profile. The overall score is then constrained by mandatory floor rules (see the adequacy section).

Processing level is the one dimension with an explicit, reproducible calculation:

1. Classify every ingredient into one of four buckets: Fresh/Raw/Whole, Meals and Rendered, Added Supplements, or Plant and Other.
2. Count the total ingredients and the number in each bucket, then calculate the percentage in each.
3. Start from a base of 50 points and apply the adjustments below.
4. Clamp the result to 0-100. Higher means less processed.

Adjustments applied to the base of 50:

• +2 for every 10% of ingredients that are Fresh/Raw/Whole (cap +30).
• -1 for every 10% that are Meals and Rendered (cap -20).
• -1.5 for every individual added vitamin, mineral or amino-acid supplement (cap -25).
• -10 if the product is extruded kibble (high-heat, high-pressure processing).
• +15 if raw, freeze-dried raw, or air-dried (low-heat processing).
• +5 if wet/canned with visible whole-meat chunks (gentler processing than extrusion).

The count of added supplements matters most: 25 added vitamins and minerals signals far more nutrient destruction than 5, because high-heat processing strips natural nutrients that must then be synthetically re-added. Two kibbles may both be extruded, but the one with 25 supplements is meaningfully more processed.

• Named animal proteins ("chicken", "salmon") score higher than generic ones ("meat", "animal fat").
• Whole foods and fresh or raw animal ingredients earn bonus points; by-products or meals used as the primary protein reduce the score.
• Every ingredient is flagged as beneficial, neutral, watch, or avoid, with a plain-English reason.
• Lentils, pulses and legumes are not automatically penalised - they are acceptable fibre and binder sources. They are only flagged "watch" when they displace animal protein excessively.
• Synthetic supplement load is captured under the processing dimension, so it is not double-counted here.

Carbohydrate is rarely listed on a guaranteed analysis, so we estimate it using the Nitrogen-Free Extract (NFE) method on a dry matter basis:

1. Convert each guaranteed-analysis value to dry matter: DMB% = (as-fed% / (100 - moisture%)) x 100.
2. Estimate carbohydrate: Carb (DMB) = 100 - protein - fat - fiber - ash, all on a dry matter basis.

Missing values: if ash is not listed we assume 5%; if crude fiber is not listed we assume 4%.

Wet food (moisture above roughly 20%) is never penalised for lower as-fed protein or fat - that is just water weight. On a dry matter basis, wet foods often have superior nutrient profiles to kibble.

• When a nutrient cannot be found on the page or link analysed (common for EPA, DHA, phosphorus and ash), we mark it unknown. We never assume the nutrient is absent from the product. The report note states the data was not found on the analysed source and may be available elsewhere on the manufacturer's website or in technical datasheets.
• Taurine is almost never listed on guaranteed-analysis labels, so its absence is never flagged as a concern or a transparency problem.
• Estimated values (assumed ash or fiber, derived carbohydrate) are clearly labelled as estimates in the report.

Two mandatory rules apply after the sub-scores and overall score are computed, and they take precedence over the standard aggregation:

In other words, nutritional adequacy can outweigh ingredient-list preference: a mediocre-ingredient but complete diet will not fall below C-, and great ingredients will not rescue an unbalanced or unsafe food from D.

BowlWise aggregates customer sentiment from public review platforms found via web search at scan time. For each product we report:

• The aggregate rating (out of 5) across platforms where data is available.
• An approximate total review count.
• A short sentiment summary and the most commonly mentioned pros and cons.
• The source platforms, with their URLs and per-platform ratings where data was found.

When collected: at the moment a scan runs. How calculated: a simple average of publicly displayed platform ratings, weighted by availability. Sources: named in the report with direct links where available.

• Caching: once a product is analysed for a given species, life stage, weight and activity profile, the result is cached. Re-scanning the same product and profile returns the cached grade for consistency.
• Canonical key: results are stored under a stable key built from brand + product line + format + species + life stage + activity, so the grade stays consistent regardless of how you enter the product (barcode, brand and name, URL, or ingredient list).
• Reformulation: scores reflect the label at the time of analysis. We do not continuously re-evaluate stored scans; a reformulated product is re-scored when a user scans it again, and a changed formulation may produce a new canonical id.
• Reviews and pricing: these are point-in-time web snapshots, refreshed only on a new scan.

• BowlWise assesses from label information - ingredients and guaranteed analysis - and does not laboratory-test the food. Actual nutrient levels and bioavailability may differ from the label.
• Estimated values (carbohydrate via NFE, assumed ash and fiber, ingredient-composition percentages) are approximations, not measurements.
• Bioavailability, sourcing quality beyond what is disclosed, and contamination cannot be fully determined from a label alone.
• Review and pricing data are public web snapshots that change over time.
• A high grade reflects strong label-level nutrition for a general pet profile; it is not a guarantee of outcomes for an individual animal.
• BowlWise is an informational tool and does not replace veterinary advice.

Therapeutic and veterinary diets (renal, hydrolysed allergy, urinary, weight-management and similar) are graded with the same framework - we do not apply special clinical scoring. Two things to understand:

• Their ingredient-list scores may look lower than expected because they are intentionally restricted (for example, limited protein for kidney disease, modified fat for pancreatitis). That restriction is clinically purposeful, not a quality flaw.
• The BowlWise grade reflects general nutritional quality from a label, not clinical appropriateness for a medical condition.

Therapeutic diets should only be fed under veterinary supervision. Use the grade as one input alongside your vet's guidance, not as a reason to switch away from a prescribed diet.