ACCURATE COVER COUNT — Confirm at Check Close, Not Just Table Open

[KaylaRippee]

If you run a full-service restaurant and your cover count has never quite added up, upvote this. The more voices, the better.

After six years running a full-service restaurant on Square for Restaurants, our cover count data has never been fully accurate — and it's not a workflow problem, it's a product gap.

Here's how our floor works: we use seat-based ordering. When we open a table we enter the cover count — a 6-top gets 6, a 2-top gets 2. Every item is assigned to a seat so any server can run food without auctioning plates. It's a clean system and Square handles it well.

The problem is what happens at close. If only 5 of 6 guests show up, seat 1 stays empty and no food rings on it — but Square still reports 6 covers because that's what was entered at open. We have no way to correct it without disrupting service. Multiply that across 15 tables on a Friday night and our cover data is consistently inflated.

What I'm asking for is simple:

1. Confirm or edit cover count at check close or payment — not just at table open. The server knows exactly how many guests ate at that moment. Two seconds, one number, accurate data.

2. On split checks, assign covers at the split — if a 6-top splits into 3 checks, each check should carry its guest count so per-check cover metrics are meaningful.

3. Flow the confirmed number into daily and weekly cover reporting — net sales per cover, covers per turn, revenue per cover by daypart. Right now those metrics are unreliable because the input is wrong.

Why it matters:

Cover count is one of the most fundamental metrics in a full-service restaurant. It drives staffing decisions, menu pricing analysis, and year-over-year performance comparisons. Right now our net sales per cover is either understated on tables with empty seats or overstated where the party was lighter than the seat count. We're making staffing and scheduling decisions on numbers we can't fully trust.

The seat-based ordering workflow is already there and it works. This is just a confirmation step at a moment that already exists — check close — that would make the data coming out of it accurate.