MenuWright FAQ
POS support, data boundaries, AI posture, billing, and reports — answers to common MenuWright questions.
Do I need a POS system to use MenuWright?
No. The fastest path is connecting Square, which syncs your menu and sales automatically. Without a POS, add menu items in the dashboard and import sales history from CSV or Excel — or explore instantly with realistic sample data. See Square connection and CSV import.
Which POS systems are supported?
Square today, in sandbox mode while the production app is under review. The POS adapter interface is designed so Toast and Clover can be added without touching the analysis or UI — but they are not implemented yet.
What if I don't have food cost data?
MenuWright still works — it classifies by popularity and revenue. Once items have food costs, it unlocks true contribution margins and the full profit-based matrix. See the matrix methodology.
How are the AI recommendations generated?
MenuWright analyzes your menu matrix and sales data and produces specific, prioritized recommendations — reprice, promote, bundle, or cut — with a projected monthly dollar impact where one can be estimated. The AI layer uses an OpenAI-compatible interface to a primary and fallback provider; generation is a candidate step, and you decide what to act on. Recommendations fail gracefully: if the provider is unreachable, the analysis completes without recommendations rather than blocking.
What is the "observed revenue change" on accepted recommendations?
When you accept a recommendation, MenuWright anchors a closed-loop measurement: it compares the item's average daily revenue in matched calendar windows before and after the action day (the action day itself is excluded). Once both windows have at least 7 days with recorded sales, the recommendation card shows an observed revenue change alongside the original projection — for example, "+$380/mo observed" next to "+$410/mo projected."
The windows use a growing matched calendar span: span = min(30, days_since_action). The comparison stays weekday-mixed unless the span is a whole week, so matching lengths keeps it symmetric but does not by itself remove weekday-mix bias. Averages divide by the calendar span, not by the number of days with records, so sparse sellers aren't inflated.
This is an observed before/after revenue change, not an attributable impact. An uncontrolled comparison on a single item cannot separate the recommendation's effect from a concurrent price change, promotion, or seasonal shift. The label says "observed revenue change" deliberately.
The number only appears when the data honestly supports it — if either window has fewer than 7 days with recorded sales, no figure is shown rather than a fabricated one. Negative deltas are included honestly.
Is my restaurant's data secure?
Every record is scoped to your restaurant, and Square credentials are encrypted at rest. Tokens are revoked and deleted when you disconnect. The connection is read-only — MenuWright never modifies your Square data. See the privacy policy.
What does each plan include?
Free: 1 analysis per month plus a preview of the single highest-impact AI recommendation. Core ($29/month): unlimited analyses and the complete AI recommendation set with dollar impact. Pro ($99/month): everything in Core plus priority support and early access to new features. All paid plans are flat-rate — no usage meters, no overage. See billing and plans for the full contract.
Can I cancel anytime?
Yes. Core and Pro are month-to-month subscriptions with no long-term commitment — manage your plan from your billing settings.
Why can't I connect my real Square account?
Production Square onboarding waits for Square's app review of MenuWright's production app. The integration is implemented and tested against the sandbox environment today; this documentation will be updated when production approval lands.
What reports will I get?
Every restaurant receives a weekly email digest (the dispatcher runs hourly and sends during the restaurant's local 8 AM hour; successful dispatches are not repeated until at least 7 days have elapsed) and a monthly PDF deep dive (1st of the month, 8 AM local), both scheduled per-restaurant timezone. Free accounts receive the same reports with only the single preview recommendation and an upgrade note; paid plans get the full content. See reports.
What is the menu engineering matrix?
It's based on the Kasavana–Smith menu engineering method (1982): every item is classified by profitability (contribution margin) and popularity (menu mix), landing in one of four quadrants — Stars, Crowd Pleasers, Hidden Gems, and Leftovers. See the methodology.
What happens to my data if I leave?
Disconnecting Square revokes the OAuth grant and deletes the stored credentials immediately. Data is scoped to your restaurant account; reach out to support@menuwright.com for account deletion.
Does MenuWright write to my Square account?
Never. The integration is strictly read-only: catalog and sales aggregates only. MenuWright never stores card or customer payment details.
What's on the roadmap?
Toast and Clover adapters are the headline directions — not yet available.