The Restaurant Owner's Guide to Reading a Verdict Breakdown

For most of restaurant history, feedback meant waiting for someone to write a review, and only about 3% of guests ever do. A verdict breakdown, a quick thumbs up or down per dish that far more guests are willing to leave, roughly 41% of tables in practice, gives you something closer to a real signal. But a new kind of data is only useful if you know how to read it without either ignoring it or overreacting to it.

Start with volume, not just the ratio

A dish with three verdicts and a rough ratio tells you almost nothing yet. A dish with two hundred verdicts and a consistent pattern tells you something real. Before reacting to any number, check how many guests it's actually based on. Small sample sizes swing wildly and a single bad night can look like a trend when it isn't one.

Watch for patterns, not single data points

• A dish trending down over weeks is more meaningful than one bad day
• A dish with a mixed verdict pattern that correlates with time of day may be a kitchen consistency issue, not a recipe issue
• A newly launched dish needs more time before its verdict pattern means anything
• A dish with strong verdicts but low order volume might just be poorly placed on the menu, not disliked
• A sudden negative shift right after a recipe or supplier change is worth investigating immediately

Don't let one bad batch become a permanent verdict

If a dish gets a rough patch of negative verdicts, check what was actually happening in the kitchen during that window before deciding the dish itself is the problem. A short-staffed night, an ingredient substitution, a new cook still learning a recipe, all of these can dent a dish's numbers temporarily without meaning anything about the dish long-term. Read the trend line, not the worst week.

Turn a good verdict into something more

A dish with consistently strong verdicts and low visibility, like Rohit Malhotra's Kadhai Paneer at Spice Route, which had the best like rate on the whole menu while sitting buried on page two, is worth moving, promoting, or simply pricing with more confidence. The verdict data is telling you something your menu layout hasn't caught up to yet. In Rohit's case, the fix wasn't complicated, tagging the dish as a Chef's Special moved it higher for every guest, and the like rate had already done the hard work of proving the promotion was deserved rather than just a hunch.

How the recommended-action categories map to a verdict pattern

A verdict breakdown becomes far more useful once you stop treating it as a raw number and start reading it against a small set of categories. A dish with a high like rate and low order volume is usually a Promote candidate, the Kadhai Paneer pattern above. A dish with a concentrated dislike pattern tied to one taste profile, too sweet, too oily, too bland for guests who tagged themselves as preferring bold spice, is a Review spice or recipe candidate, pointing at the kitchen rather than the dish concept itself. A dish showing allergen-conflict strikeouts, guests being blocked from ordering it because of a flagged conflict, gets flagged as at risk, which is worth checking even if its verdicts look otherwise fine, since it's a different kind of problem entirely.

How often to actually check this

A verdict breakdown isn't something to stare at daily, day-to-day swings are mostly noise. A weekly glance is usually enough to catch a real shift before it's been running for a month, and a monthly deeper look is a reasonable cadence for reconsidering menu placement or descriptions based on accumulated patterns. Checking too rarely means you miss problems early. Checking obsessively just means reacting to noise that would have resolved itself.

TasteGraph's verdict system is built to make this kind of pattern easy to see, giving guests a fast way to weigh in on a specific dish and giving you a dashboard that separates real signal from noise, sorting dishes into clear recommended actions like Promote, Review spice or recipe, and Flag as at risk, rather than a pile of raw numbers you have to make sense of yourself.