Classification or categorization?
- Categorization
Assigning products to categories in the store or marketplace tree. It depends on the channel and sales objective; one product can be assigned to multiple categories, e.g.:Home and Garden > Power Tools > DrillsandPromotions > Workshop. Answers the question: "Where is the buyer supposed to find it?" — however, it does not impose a full set of technical features. - Classification
Channel-neutral (domain-specific) product classification system, in which each product belongs to exactly one class. The class defines not only the position in the hierarchy but, above all, a set of technical attributes, units, and permissible values (e.g., ETIM, ECLASS). It answers the question: "What exactly is it and how to describe it technically?".
How marketplaces do it
In marketplaces classification hierarchy is part of the category tree - is not a separate, parallel tree. Usually last node (leaf category) fulfills the role of classification classes with an attached set of features. In some cases, the role of the class is fulfilled by parent category with respect to the leaf, if the attribute schema has been assigned to it.
How industrial classifications do it
- stable class identifiers (codes),
- normative attributes for each class,
- allowed dictionary values,
- multilingual labels and definitions, which facilitates data exchange in the supply chain,
- assignment of the product always to only 1 class
Why is it important for getName.ai?
In this text, we consistently use the term"classification"as system names and we talk about "classification" products - that is, assigning them to the correct classes. getName.ai works precisely with product classifications: it class determines, what technical parameters we will identify and verify. First, we indicate the class, then we recognize the attributes.
To remember:
-
Categorization = sales navigation; possible multiple categories for one product.
-
Classification = attribute structure and validation; exactly one class for one product.
-
In marketplaces class is part of the category tree (most often a leaf or its parent) and to this node a set of features is attached.
What class - the missing element in automation
Three elements needed to recognize attributes
getName.ai has offered automatic recognition of product attributes from the beginning. For the system to function correctly, it needs:
- Selection of classification (e.g. ETIM, ECLASS, Amazon, Otto)
- Indication of the product class (e.g., "257064" for "Gaming chairs")
- Text data about the product (description, name, specification)
POST https://api.getname.ai/v1/attributes
{ "classification":"allegro",
"class_id": "257064", // -- Where to get this from?
"description": "The ALPHA gaming chair is a swivel bucket-style seat in black, perfect for both adults and
children with a height from 150 to 185 cm. Made of high-quality materials, it features quilted
eco-leather with carbon inserts, a steel frame, rubber casters, a class 3 gas lift, adjustable
armrests, two adjustable and removable cushions (lumbar and neck), a reclining backrest with lock,
and a seat height adjustment mechanism ranging from 46 to 56 cm. The chair’s dimensions are:
height 130 cm, width 62 cm, depth 51 cm, seat width 49 cm, seat depth 51 cm, and the total weight
with packaging is 17 kg. The chair is delivered in its original packaging with a simple assembly
manual." }
Multi-marketplace listing
Product: "Gaming chair"
Product description: "The ALPHA gaming chair is a swivel bucket seat in black, ideal for both adults and children with a height from 150 to 185 cm. Made from high-quality materials, it features quilted eco-leather with carbon inserts, a steel frame, rubber wheels, a class 3 gas lift, adjustable armrests, two adjustable and removable cushions (lumbar and neck), an adjustable backrest angle with lock, and a seat height adjustment mechanism ranging from 46-56 cm. The chair's dimensions are: height 130 cm, width 62 cm, depth 51 cm, seat width 49 cm, seat depth 51 cm, and the weight with packaging is 17 kg. The chair is delivered in its original packaging with simple assembly instructions."
Automatic classification returns:
| Marketplace | Parent categories | Id | Class name |
|---|---|---|---|
| AMAZON UK | - | CHAIR | Chair |
| EBAY UK | Home, Furniture & DIY > Furniture | 54235 | Chairs |
| OTTO DE | Stühle | Gaming-Stuhl | Gaming-Stuhl |
| ALLEGRO PL | Electronics > Computers > Accessories (Laptop, PC) | 257064 | Gaming chairs |
Savings: 15 minutes of manual searching → 1 minute of automatic classification
Border products
Classification: "Allegro"
Product: "Gaming chair"
Product description: "The ALPHA gaming chair is a swivel bucket seat in black, ideal for both adults and children with a height from 150 to 185 cm. Made from high-quality materials (...)"
Automatic classification returns:
| Parent categories | Id | Class name |
|---|---|---|
| Electronics > Computers > Accessories (Laptop, PC) | 257064 | Gaming chairs |
| Home and Garden > Furniture > Workshop | 251011 |
Swivel chairs
|
Both classes are "good," which is confirmed in practice, as offers for this type of product can be found in both categories on the marketplace. The attributes describing both classes have many common attributes, but also many by which they differ. Based on this, a decision must be made about the final classification.
Below is a comparison of attributes specific to these classes. It is worth noting the subtle differences when attributes have the same name, different identifiers, or different data types.
| Id | Attribute name | Type | Id | Attribute name | Type | |
|---|---|---|---|---|---|---|
| 18648 | Depth of the furniture [cm] | float | 18648 | Depth of the furniture [cm] | float | |
| 202341 | Seat depth [cm] | float | 202341 | Seat depth [cm] | float | |
| 202353 | Maximum seat height [cm] | integer | 202353 | Maximum seat height [cm] | integer | |
| 128653 | Upholstery material | dictionary | 128653 | Upholstery material | dictionary | |
| 202357 | Minimum seat height [cm] | integer | 202357 | Minimum seat height [cm] | integer | |
| 18646 | Width of the furniture [cm] | float | 18646 | Width of the furniture [cm] | float | |
| 202337 | Seat width [cm] | float | 202337 | Seat width [cm] | float | |
| 18647 | Height of the furniture [cm] | float | 18647 | Height of the furniture [cm] | float | |
| 231921 | Upholstery color | dictionary | 128651 | Upholstery color | dictionary | |
| 202345 | Maximum load [kg] | integer | 9466 | Maximum load [kg] | float | |
| 237206 | Model | string | 10838 | Type | dictionary+ | |
| 218669 | Safety information | dictionary | 128650 | Body color | dictionary | |
| 232005 | Additional functions | dictionary+ | 128652 | Corpus material | dictionary | |
| 237214 | Line | string | ||||
| 249829 | Medical device | dictionary | ||||
| 245321 | Additional features | dictionary+ |
Extensive industry classifications (ETIM, ECLASS, GS1)
- ETIM: 5,600+ product classes
- ECLASS: 41,600+ classes
- GS1/GPC: 5,300 classes of type "brick"
Solution:The system seamlessly searches virtually any extensive classifications and returns the best matching classes. The employee only verifies the suggestions instead of manually searching through thousands of options.
Product: "Electronic washbasin faucet"
Product description: "The electronic kitchen faucet is an elegant, stainless steel sink faucet with a touchless function (infrared sensor) and a traditional side handle for adjusting water temperature and flow. The standing model, with a height of 400 mm, features a swivel spout with a reach of 221 mm, dual power supply (electric/battery) ensuring reliable operation, and a 5-year manufacturer's warranty, combining modern technology with practicality and hygiene in kitchen use."
Automatic classification for 3 API queries returns:
| Classification | Parent categories | Go | Class name |
|---|---|---|---|
| ETIM 10 | Sanitary taps | EC011431 | Kitchen mixing tap |
| ECLASS 15 | Construction technology > Sanitation > Fitting (sanitary) |
22360101
|
Fitting (sink) |
| GS1/GPC | Plumbing/Heating/Ventilation/Air Conditioning > Plumbing/Heating/Ventilation/Air Conditioning > Sanitary Ware | 10002602 | Faucets/Taps |
Savings: 5 minutes of manual searching → 20 seconds of automatic classification
Solution: automatic classification in getName.ai
How does classification work?
The new automatic classification feature (endpoint /classify) analyzes the product description and automatically suggests the best matching classes:
Request:
POST https://api.getname.ai/v1/classify
{
"classification":"allegro", // or "amazon_en", "etim_v10_en", "ebay"...
"description": "The ALPHA gaming chair is a swivel bucket seat in black, ideal for both adults and children with
a height from 150 to 185 cm. Made from high-quality materials, it features quilted eco-leather
with carbon inserts, a steel frame, rubber wheels, a class 3 gas lift, adjustable armrests, two
adjustable and removable cushions (lumbar and neck), an adjustable backrest angle with lock,
and a seat height adjustment mechanism ranging from 46-56 cm. The chair's dimensions are: height
130 cm, width 62 cm, depth 51 cm, seat width 49 cm, seat depth 51 cm, and the weight with
packaging is 17 kg. The chair is delivered in its original packaging with simple assembly
instructions."
}
Respons:
{
"classes": [
{
"id": "257064",
"name": "Gaming chairs",
"path": "Electronics > Computers > Accessories (Laptop, PC)"
},
{
"id": "251011",
"name": "Swivel chairs",
"path": "Home and Garden > Furniture > Study"
},
{
"id": "20274",
"name": "Chairs and armchairs",
"path": "Home and Garden > Furniture > Study"
}
]
}
Important: The system returns matching classes ordered by relevance - the first class is the best match. The result always includes at least one class, but not necessarily more - the system only shows classes that make sense for a given product, omitting those that would be incorrect. The final decision is up to you.
getName.ai - one tool, all classifications
The system unifies the way of working with all classifications - whether it's Allegro, Amazon, ETIM, or your own class structure. Instead of learning dozens of different tools or building your own solutions, you use one API for everything.
Key advantages:
- Uniform classification method for all platforms
- High recognition quality - test it on your own data!
- Full flexibility - support for custom classifications
- Simple integration via API - works with any IT system without time-consuming implementations
- Flexible plans - from 250 queries for small shops to tens of thousands for enterprises
Real applications in practice
Multi-marketplace listing
Product: "Gaming chair"
Product description: "The ALPHA gaming chair is a swivel bucket seat in black, ideal for both adults and children with a height from 150 to 185 cm. Made from high-quality materials, it features quilted eco-leather with carbon inserts, a steel frame, rubber wheels, a class 3 gas lift, adjustable armrests, two adjustable and removable cushions (lumbar and neck), an adjustable backrest angle with lock, and a seat height adjustment mechanism ranging from 46-56 cm. The chair's dimensions are: height 130 cm, width 62 cm, depth 51 cm, seat width 49 cm, seat depth 51 cm, and the weight with packaging is 17 kg. The chair is delivered in its original packaging with simple assembly instructions."
Automatic classification returns:
| Marketplace | Parent categories | Id | Class name |
|---|---|---|---|
| AMAZON UK | - | CHAIR | Chair |
| EBAY UK | Home and Furniture > Furniture | 38208 | Sofas and armchairs |
| OTTO DE | Stühle | Gaming-Stuhl | Gaming-Stuhl |
| ALLEGRO PL | Electronics > Computers > Accessories (Laptop, PC) | 257064 | Gaming chairs |
Savings: 15 minutes of manual searching → 1 minute of automatic classification
Border products
Classification: "Allegro"
Product: "Gaming chair"
Product description: "The ALPHA gaming chair is a swivel bucket seat in black, ideal for both adults and children with a height from 150 to 185 cm. Made from high-quality materials (...)"
Automatic classification returns:
| Parent categories | Id | Class name |
|---|---|---|
| Electronics > Computers > Accessories (Laptop, PC) | 257064 | Gaming chairs |
| Home and Garden > Furniture > Workshop | 251011 |
Swivel chairs
|
Both classes are "good," which is confirmed in practice, as offers for this type of product can be found in both categories on the marketplace. The attributes describing both classes have many common attributes, but also many by which they differ. Based on this, a decision must be made about the final classification.
Below is a comparison of attributes specific to these classes. It is worth noting the subtle differences when attributes have the same name, different identifiers, or different data types.
| Id | Attribute name | Type | Id | Attribute name | Type | |
|---|---|---|---|---|---|---|
| 18648 | Depth of the furniture [cm] | float | 18648 | Depth of the furniture [cm] | float | |
| 202341 | Seat depth [cm] | float | 202341 | Seat depth [cm] | float | |
| 202353 | Maximum seat height [cm] | integer | 202353 | Maximum seat height [cm] | integer | |
| 128653 | Upholstery material | dictionary | 128653 | Upholstery material | dictionary | |
| 202357 | Minimum seat height [cm] | integer | 202357 | Minimum seat height [cm] | integer | |
| 18646 | Width of the furniture [cm] | float | 18646 | Width of the furniture [cm] | float | |
| 202337 | Seat width [cm] | float | 202337 | Seat width [cm] | float | |
| 18647 | Height of the furniture [cm] | float | 18647 | Height of the furniture [cm] | float | |
| 231921 | Upholstery color | dictionary | 128651 | Upholstery color | dictionary | |
| 202345 | Maximum load [kg] | integer | 9466 | Maximum load [kg] | float | |
| 237206 | Model | string | 10838 | Type | dictionary+ | |
| 218669 | Safety information | dictionary | 128650 | Body color | dictionary | |
| 232005 | Additional functions | dictionary+ | 128652 | Corpus material | dictionary | |
| 237214 | Line | string | ||||
| 249829 | Medical device | dictionary | ||||
| 245321 | Additional features | dictionary+ |
Extensive industry classifications (ETIM, ECLASS, GS1)
- ETIM: 5,600+ product classes
- ECLASS: 41,600+ classes
- GS1/GPC: 5,300 classes of type "brick"
Solution:The system seamlessly searches virtually any extensive classifications and returns the best matching classes. The employee only verifies the suggestions instead of manually searching through thousands of options.
Product: "Electronic washbasin faucet"
Product description: "The electronic kitchen faucet is an elegant, stainless steel sink faucet with a touchless function (infrared sensor) and a traditional side handle for adjusting water temperature and flow. The standing model, with a height of 400 mm, features a swivel spout with a reach of 221 mm, dual power supply (electric/battery) ensuring reliable operation, and a 5-year manufacturer's warranty, combining modern technology with practicality and hygiene in kitchen use."
Automatic classification for 3 API queries returns:
| Classification | Parent categories | Go | Class name |
|---|---|---|---|
| ETIM 10 | Sanitary taps | EC011431 | Kitchen mixing tap |
| ECLASS 15 | Construction technology > Sanitation > Fitting (sanitary) |
22360101
|
Fitting (sink) |
| GS1/GPC | Plumbing/Heating/Ventilation/Air Conditioning > Plumbing/Heating/Ventilation/Air Conditioning > Sanitary Ware | 10002602 | Faucets/Taps |
Savings: 5 minutes of manual searching → 20 seconds of automatic classification
Key Benefits - Summary
Automatic product classification in getName.ai is the missing piece of the puzzle in e-commerce automation. The system does not replace humans in making business decisions but eliminates the tedious, time-consuming task of searching through thousands of classes in dozens of different classifications.
When does it really speed up work?
- Multichannel Sales - one product goes to 5 platforms, each with a different classification and class ID. Instead of learning 5 different structures, one query to getName.ai is enough.
- New Employees - the onboarding time is reduced from weeks to days. They don't need to know the details of each classification, just focus on the substantive evaluation of the system's suggestions.
- Mass Product Import - 1000 products × 3 platforms equals 3000 class assignment decisions. Manually, this takes 2-3 weeks, with getName.ai only 1-2 days (with verification).
Main Advantages of the Solution:
- Unification - one tool for all platforms and classifications
- Time Savings - 90% reduction in classification time
- Scalability - handling thousands of products without increasing the team
- High Accuracy - 92-95% accuracy of suggestions
Endpoint /classify in combination with the existing /attributes creates a complete system: from raw product description, through automatic classification, to a full set of technical attributes ready for export to any sales platform.
