The classic features of the category, re-read through “what user need does this actually serve” rather than “what do trackers usually have.” Seven needs, not thirty-two features — most requested features turn out to be one of these seven wearing a different name.
| Need | What people actually ask for | Verdict |
|---|---|---|
| N1 · Capture cost | Faster logging, widgets, shortcuts, barcode | Build Android shortcuts/share_target are cheap, unused wins — no native app, no home-screen widget on either platform. |
| N2 · “Tell me what it meant” | Insights, summaries, “what changed” | Build, without interpretation Contributions per ingredient instead of a processing/health score. Show what a food contributed, not whether it was good. |
| N3 · “Am I missing something?” | Adequacy, deficiency, “is this enough” | Build The coverage read. Focus-nutrient selection already exists in preferences — the same primitive a professional would use on the pro platform. |
| N4 · “Why do I feel like this?” | Mood/energy correlation, symptom tracking | Unresearched Real need; no evidence has been gathered on it in any round of this collaboration. Not assessed, rather than assessed and deprioritised. |
| N5 · “Someone else needs to see this” | Export, sharing, a clinician or coach handoff | Build The shareable weekly summary — user-initiated, no second account, no GDPR Article 9 complication. See the pro platform. |
| N6 · “Don’t make me feel bad” | No red, no streaks, no guilt | Kept, reframed as permission See softness for the evidence and the reframe. |
| N7 · The practical side | Recipes, favourites, barcode-as-repair | Bend the rule Favourites with custom titles already solve reuse. Barcode as a repair path for packaged foods (not a search screen) is defensible — blocked on licensing, not design. |
Every “build” above is a generic mechanism, never a special case: contributions work for any nutrient, focus-selection works for any nutrient, the shareable summary is the same object for a friend, a professional, or nobody. The one place this discipline matters most is N3 → the pro-platform settings screen on the pro platform: a professional’s “raise the protein target” suggestion has to be the same generic target-setting primitive a user’s own goal-picker uses, not a parallel code path.