How this guide scores a calorie tracker
What a score is
Each app can receive an overall score and seven dimension scores: user feedback, pricing value, accuracy, feature depth, ease of use, beginner friendliness, and coachability. The scale is 10 unless the record says otherwise. A missing score stays missing. This page does not turn a blank into zero, and it does not invent a price.
Store stars are shown on fact pages as context. They are a different number, from a different source, and they are not copied into the editorial score.
Three layers
A score is not one input. The record has a quantitative layer, a qualitative layer, and an editorial layer. The editorial layer is allowed to disagree with the store rating. When it does, the verdict has to say why, using a field that is actually on the record.
Quantitative
Numbers already stored for the app: store rating and rating count, list price, monthly, yearly, and weekly amounts or ranges, trial text, feature availability, and any quality, coverage, usefulness, or confidence value on a feature. Goal-fit scores, theme mention counts, and the date the listing was updated sit here too. If a field is empty, the page says it is unknown.
Qualitative
Words already stored for the app: liked and disliked theme summaries, best-for and not-ideal-for lines, feature notes, and short evidence lines. These describe a pattern. They are not a paste of store reviews. One disliked theme does not set the score by itself.
Editorial
The editorial layer reads the first two and writes the dimension scores and the verdict. A feature marked unavailable cannot be praised as if it were present. A thin record — a score with no price, no themes, and no feature notes — does not get an editorial review URL. The fact page can still exist.
What each dimension is allowed to use
User feedback. Quantitative: theme mention counts and the store rating count as context. Qualitative: liked and disliked theme summaries. A high score means the themes are specific and repeated. A low score means complaints are repeated or the record is too thin to trust. The editor must not treat a 4.8 store average as this score.
Pricing value. Quantitative: monthly, yearly, and weekly figures, and the trial line. Qualitative: pricing notes in the summary. A high score means the listed price is legible and the paid tier matches the features that are marked on. A low score means the price is missing, contradictory, or high relative to features that are marked off. Unknown price is not scored as free.
Accuracy. Quantitative: the accuracy feature flag, its confidence, and any coverage value. Qualitative: accuracy notes and evidence lines, plus complaint themes about wrong calories. A high score means the record describes how logging is checked. A low score means accuracy is unmarked or the notes describe gaps. A high store rating is not an accuracy score.
Feature depth. Quantitative: how many tracked features are marked available, plus quality and coverage where those exist. Qualitative: feature notes. A high score means several logging and nutrition features are marked on, with notes. A low score means the record is a calorie counter with little else, or the flags are on without notes. Flags without notes are not treated as proof of quality.
Ease of use. Quantitative: none beyond the score itself. Qualitative: themes about logging friction, and best-for lines that mention a simple diary. A high score means the themes describe a flow people finish. A low score means the themes describe friction. The editor must not infer ease of use from a download count. This catalog does not store one.
Beginner friendliness. Quantitative: the beginner-friendliness score when present. Qualitative: best-for lines that name first-time trackers, and themes about setup. A high score means the record says a new tracker can start without a coach. A low score means setup themes are negative or the field is absent. Absence is not a low score. It is unknown, and the ranking falls back to the overall score.
Coachability. Quantitative: the coach-connection flag. Qualitative: notes on that feature, and themes about feedback. A high score means a coach or check-in is marked on and described. A low score means the product is marked as self-serve only, or coach themes are negative. A coach flag is not a clinical program.
How the layers combine
Comparisons use the same fields on both apps and talk about deltas: price, features marked on one side only, and dimension scores that differ. A single overall number is not the verdict. Best-of pages exist only when a filter or a sort changes who ranks, and the page says which field it uses. The page at “calorie counter app” redirects to “calorie tracker” because those two lists are the same order.
Rankings are not paid placements. A commercial relationship does not buy a score.
Worked example
This walkthrough uses Bite AI: Photo Calorie Scanner, because that record currently has the most of the fields this method cites. Fact page.
Overall score on the record: 7.3/10. Method healthier-v1: weighted blend of 9 data-derived sub-scores; see scoring.md.
Monthly price: $7.99
Feature rows used here: Photo logging (quality: mixed) — The app offers instant AI analysis via photo, but user reviews indicate frequent inaccuracies requiring manual correction. Evidence: description screenshots. Audio logging (quality: good) — Supports voice logging for meals and workouts. Evidence: description.
One theme: Flexible Macro Tracking (1 mentions). Users appreciate the ability to track macros without limits and monitor them over the week rather than just daily.
- Feature depth: 8.9/10. Quality-weighted share of tracked features available (excellent 1.0 .. poor 0.2), x1.6, capped at 10.
- Ease of use: 9/10. 5 + 5 x best quality of fast-logging features (AI image/text/audio, barcode).
- Traction: 4.3/10. log10 of 390 total ratings (iOS+Play), 1M ratings = 10.
- Maintenance: 9/10. Last release 3 days ago (half-life 90d, 70%) + median 12 days between releases (30%).
The verdict is allowed to cite only the lines above. If a later record drops the monthly price or a feature note, the review has to drop that sentence too.
How a score is written
- 1. Freeze the record. The editor works from the stored fields for that app: prices, feature flags, quality and coverage, theme summaries and mention counts, goal fit, store rating as context, and the date the listing changed. A fact that is not in the record is not used.
- 2. Mark the gaps. Empty price, empty feature note, empty theme, and empty dimension stay empty. They are not filled with zero, with the category average, or with a guess from a similar app.
- 3. Score one dimension at a time. Each dimension uses only the fields named in the section above. The reason line has to point at one of those fields. A reason that only restates the number is not enough.
- 4. Write the overall score after the dimensions. The overall score is the guide’s summary judgment. It is not a hidden weighted formula published as if it were a lab result, and it is not the average of the store stars. If the dimensions disagree with each other, the overall summary has to say which disagreement mattered.
- 5. Decide whether a review URL exists. An editorial review is published only when the record has an overall score and at least two of: a real price line, feature evidence, review themes, or a fit line. A thinner record keeps its fact page and does not get a second URL that repeats the score.
How to read the scale
The scale runs from 1 to 10. These bands describe the editor’s reading. They are not grade cutoffs from a validated instrument.
8 to 10. The fields the dimension is allowed to use are present, specific, and point the same way. A 9 still has to name the field. A round number is not a compliment.
5 to 7. The record supports the feature or the fit, with a gap: a price that is legible but high for what is marked on, themes that split, or notes that describe a limit. This is a usable app with a tradeoff the verdict should state.
1 to 4. The fields the dimension uses are negative, contradictory, or so thin that a higher score would be inventing evidence. A low score is not assigned because a field is missing. Missing stays unknown and the dimension is left blank.
Store stars use a different scale, usually 1 to 5, from a different author. They are shown beside the editorial score so a reader can see the disagreement. They are never copied into it.
Confidence, comparisons, and rankings
Some records store a confidence value next to the score. When it is present, the fact page and the review show it. Confidence is the editor’s note on how complete the record was, not a statistical confidence interval and not a margin of error from a sample of users.
A comparison essay is allowed only when the pair is on the priority list, or when the two records actually diverge: at least two tracked features differ, or both have a pricing summary and those summaries differ. The essay talks about those deltas. It does not declare a winner from the overall score alone.
A best-of page exists when a filter changes who can rank. Beginners sort on beginner friendliness when that score exists. Free keeps apps whose pricing summary mentions a free tier. AI logging keeps apps marked for photo, audio, or text logging. The page names the field. The duplicate “calorie counter app” list redirects to “calorie tracker” because the order is the same.
Nothing in this procedure is a blinded meal test, a panel of dietitians, or a published error rate in calories. If a field reads like a measurement, it is only a measurement when the record itself stored one. The guide does not manufacture that number.
Freshness
Fact pages show the record’s update time. An essay shows the day it was published and should be revised when a price, feature, or verdict would change. This methodology page was revised on 22 October 2026.
Limits
Personal fit varies. The guide is consumer wellness information, not medical advice, and it does not diagnose, treat, or replace a clinician. Store text can be wrong. Theme counts are not a census of every user. A score is a reading of the record, not a guarantee that a person will lose weight.
Send a correction if a price, feature, or verdict on this page is wrong. It is emailed privately and is not published.