Case · client 24TTL · 2026

Size data control

24TTL · lamoda

Scope: 150k SKU; 800 thousand rows of data

Task: find the discrepancies between Lamoda’s size charts and the brands’ own sites — by hand that is 800 thousand rows of comparison.

Two size charts, line by line the brand’s chart is fetched, rendered, read and normalised, then set against Lamoda’s row by row; a wrong product is turned away at the gate, a chart that is a picture goes through the vision-LLM, a row whose units cannot be read goes to quarantine — only what the system can answer for reaches the report

Lamoda · SKU

Brand’s siteidentity✓✗

Render · Playwright

Extract the tablegolden✓

Normalise · in → cmin→cm

Comparison

size	chest, cm		waist, cm
size	Lamoda	brand	Lamoda	brand
S	88	88	70	70
M	92	92	74	74
L	96	98	78	78
XL	100	100	82	82

Quarantine2 rows held back

vision-LLM · reads the picture

#1042?✗#1057?✗#1063?✗#1071?✗#1088?✗#1094?✗

01234checked01discrepancies012quarantined

How it works

DATA TRUSTWORTHINESS LAYER — identity check of the product against the DOM · quarantine for unverified rows · vision-LLM on non-standard tables · golden tests against regression

Results

150k: SKUs under automatic comparison
800k: rows instead of manual checking
6: stages in the pipeline

Discrepancies are visible by address

A dashboard in Lamoda’s brand book: filters on every dimension, line-by-line comparison, CSV export

The check repeats

The run starts again on any volume — comparison stopped being a one-off exercise

Stack