Glossary

Review & Feedback Analysis

In one line: Review and feedback analysis uses AI to read every customer review across platforms — Google, GrabFood, Foodpanda, and more — and surface what’s recurring, so nothing slips by.

What it does

Instead of a manager skimming a handful of reviews when they remember to, AI reads all of them — continuously, across every platform you’re listed on. It groups feedback into themes (slow service, a dish that’s drawing complaints, staff who get named repeatedly for good reasons) and tells you what’s trending up or down over time.

Why it matters

Reviews shape your reputation whether or not you read them. A recurring complaint you catch in week one is a fixable operational issue; the same complaint caught in month three is a pattern that’s already cost you customers. Automated analysis turns scattered feedback into an early-warning system — and surfaces what people love, so you can do more of it.

What to watch for

The value is in the aggregation and trend-spotting, not any single review. Make sure whatever tool you use covers all the platforms your customers actually use, and that it distinguishes a one-off gripe from a genuine recurring theme.

Related terms

Sentiment Analysis, Menu Engineering

Glossary

Sentiment Analysis

In one line: Sentiment analysis is AI that reads text — reviews, social posts, support messages — and decides whether the writer is happy, unhappy, or neutral.

Why it matters

You can’t read every review or DM a customer writes. Sentiment analysis automates the reading — classifying every customer message so you can track how perception of your business is changing day by day, by product, by location.

Modern aspect-based sentiment

Old sentiment tools labeled a whole review positive or negative. Modern tools (powered by large language models) extract sentiment per topic within a single message: “food was excellent, service was slow.” That’s two data points from one review — one positive on food, one negative on service. The granularity is what makes the tool useful.

Where retailers use it

• Monitor product reviews for emerging quality issues
• Track restaurant reviews for service or food trends
• Prioritize support tickets — angry customers get human attention faster
• Measure marketing campaign reception on social media

Related terms

Menu Engineering, Large Language Model (LLM), Generative AI

Glossary

Menu Engineering

In one line: Menu engineering uses sales and margin data to design menus that nudge customers toward your most profitable dishes — AI makes the analysis continuous.

The classic framework

Every menu item is rated on two dimensions: popularity and profit margin. The combinations:

• Stars — high margin, high popularity. Feature them.
• Plowhorses — low margin, high popularity. Either raise price or reduce cost.
• Puzzles — high margin, low popularity. Reposition or promote.
• Dogs — low margin, low popularity. Cut.

What AI adds

Traditional menu engineering is a quarterly spreadsheet exercise. AI version runs continuously: it watches sales velocity, knows your cost of goods, and surfaces items that have shifted category — the dish that was a star last quarter but now sells less due to a new competitor, or the puzzle that’s gaining traction after a tweak. It also recommends placement, pricing, and bundling experiments.

Why it matters

For most restaurants, a 1–2% improvement in average menu margin compounds to meaningful annual profit. The math: a restaurant doing $50K/month in food sales with a 2% margin improvement adds $12K/year of pure profit — without spending anything on new customers.

Related terms

Sentiment Analysis, Demand Forecasting, Dynamic Pricing