How scanner-based newQC assessment of mangoes moves beyond pass/fail inspection to predict which fruit consumers will find appealing — before it reaches the shelf.
Key results
- An AI classifier that reliably predicts which products shopper will enjoy was created
- The classifier rating matches to > 90% with shopper ratings
- The validation was perform for 6 days inside 2 retail makets with over 240 consumer intervies and 850 mangoes tasted.
USE CASE — CONSUMER SATICFACTION INTELLIGENCE
The Limitation of Pass/Fail Quality Control
Conventional incoming inspection answers one question: does this shipment meet the norm?
It does not answer the question that determines repeat purchase and consumer satisfaction: will shoppers enjoy this product?
A mango can pass every size and defect check and still disappoint. Ripeness at the point of consumption — not at goods receipt — determines whether a consumer buys again.
The gap between “passes inspection” and “consumers enjoy it” is where category reputation is won or lost.
The newQC Approach: Classifying for Consumer Outcome
Using internal quality parameters measured by the newQC NIR scanner, such as sweetness, acidity, juice content, firmness, ripeness index, and remaining shelf-life, a consumer preference classifier was developed for mangos.
The classifier assigns each fruit to one of three categories:
| Tasty — Firm
Appeals to consumers who prefer fresh, firm mangos with a clean flavour. Typically, earlier in the ripeness window. |
Tasty — Sweet & Creamy
Appeals to consumers who prefer fully ripe, sweet, soft-textured mangos. Peak ripeness profile. |
Likely Unappealing
Fruit that is likely to be rated as dull, tasteless, or unappealing by consumers — regardless of whether it passed standard quality checks. |
This classification does not replace quality inspection. It adds a consumer-facing layer on top of it — enabling retailers and importers to make more precise decisions about timing, placement, and sourcing.
Consumer Validation Study
IN-STORE SENSORY EVALUATION — STUTTGART AND BAD KROZINGEN
The classifier was validated through a structured in-store consumer study conducted across two locations in Germany over six days.
| Samples evaluated | 850 individual mangos |
| Consumers participating | 240 |
| Study duration | 6 days across 2 locations |
| Locations | Stuttgart and Bad Krozingen |
| Method | Each mango scanned before evaluation. Consumer assessed ripeness by appearance and touch, then tasted and rated ripeness and flavour. |
| Classifier accuracy | ~90% — scanner classification matched consumer rating in approximately 9 out of 10 cases |
| Each mango was assessed by the scanner before the consumer evaluation. The consumer rated ripeness first by appearance and touch, then by taste. The scanner had no access to consumer feedback — the match between scanner classification and consumer rating was determined after the fact. |
What This Enables
- Sourcing decisions based on consumer outcome, not just norm compliance
- Shelf-life and placement decisions informed by ripeness profile at goods receipt
- Supplier conversations grounded in consumer preference data, not just defect rates
- Reduction in consumer disappointment from fruit that passes inspection but underdelivers on taste
A full description of the study methodology, locations, and results is available in our blog post:
Applicable Products
The consumer preference classifier has been developed and validated for mangos. The underlying scanner parameters — sweetness, acidity, juice content, firmness, ripeness, shelf-life — are available across:
- Mangos — full classifier validated
- Table grapes, peaches, blueberries — parameter measurement available; consumer classifier in development
- Avocados — firmness and dry matter; ripeness stage classification applicable
- Tomatoes — firmness and dry matter; ripeness stage, taste intensity by lycopene content
Related pages
Related pages: newQC and FreshScan applications · Expertise: Food Science, Robust Decisions from Limited Data.