Evidence, not claims
Impact & evidence
Palatify's public-benefit case — better nutrition transparency for restaurant meals, especially at the independent restaurants federal menu-labeling rules don't reach — will be made with measured evidence. This page defines the measurements before the pilot, and reports nothing that hasn't happened.
Standing commitment. No users, restaurants, partners, outcomes, revenue, or health effects are claimed here or anywhere in the product unless they are recorded, dated, and reproducible. The current catalog is demonstration data; orders are prototype orders; every simulated state is labeled simulated.
Pilot metrics — pre-registered
Defined now, measured when the Boston independent-restaurant pilot runs. Definitions won't move after the fact.
Pre-order decision rate
Defined · not yet measuredShare of logged restaurant meals whose dish was viewed in Palatify before the order was placed.
Why it matters: Measures whether the product changes the decision, not just records it.
Verified-menu coverage
Defined · not yet measuredPercent of listed dishes whose nutrition data has been reviewed and confirmed by the restaurant.
Why it matters: The transparency gap for independent restaurants is the core problem.
Correction turnaround
Defined · not yet measuredMedian time from a flagged nutrition value to a versioned, restaurant-approved correction.
Why it matters: Trust depends on how fast bad data gets fixed, visibly.
Logged-meal accuracy
Defined · not yet measuredDifference between order-confirmed meal contents (ground truth) and engine estimates, per nutrient.
Why it matters: The order-to-log loop creates a validation dataset no retrospective logger can build.
Independent-restaurant onboarding cost
Defined · not yet measuredHours and dollars to digitize, estimate, and verify one independent restaurant's full menu.
Why it matters: National importance requires a protocol that scales beyond one city.
Diet-quality delta
Defined · not yet measuredChange in sodium and calorie intake per restaurant meal for active users vs their own baseline.
Why it matters: The public-benefit claim must eventually rest on measured behavior change.
Source ledger
Where every nutrition number in the catalog comes from, by restaurant.
| Restaurant | Dishes | Data source |
|---|
0 restaurants · 0 dishes · 0 dishes under the demo partner-verified pattern. Rows marked restaurant-published quote real public nutrition disclosures (retrieval dates in each listing); rows marked engine estimate cover real restaurants that publish no nutrition, clearly labeled. No listed restaurant has authorized or partnered with Palatify — prices, ratings, and delivery estimates are placeholders. This table becomes the verified ledger as pilot restaurants sign on.
Recorded activity
Corrections recorded via the demo partner terminal on this device: 0.
No aggregate usage statistics are published because none have been measured. When the pilot produces real numbers, they will appear here with methodology.
How this product accumulates evidence
Dated, tagged releases with changelogs. Orders stamped as prototype so demo data can never masquerade as traction. Meal-log entries that carry source, order reference, portion, and two confidence dimensions — exportable as JSON. Menu corrections that are versioned and public. Pre-registered pilot metrics. Each is a by-product of using the product, not a claim written after the fact.