tubeCatalog positioning model

DNA
Algorithm

HERO GRAPHIC Techniczna wizualizacja algorytmu DNA

Everyone can be selected as the first!

scroll
01

Equal positioning

Every creator has the same chance

A channel with 20 million subscribers receives no advantage over a channel with 12,000. Subscriber count, views and popularity are not used to determine which block appears first.

The same algorithm. The same rules. The same probability.

GRAPHIC 01 Porównanie identycznego prawdopodobieństwa dla dużego i małego kanału
02

Batching model

Results are divided into blocks

All matching channels are retrieved and divided into blocks containing up to six records. Each block becomes one page of the catalogue.

97 channels = 16 full blocks + 1 residual block.

GRAPHIC 02 Podział 97 kanałów na bloki po sześć rekordów
03

Random first position

Any block can appear first

Once the blocks are created, the algorithm randomly selects which one begins the result sequence. It may come from the beginning, middle or end of the dataset.

No block is permanently assigned to the top.

GRAPHIC 03 Losowa zmiana pierwszego bloku wyników
04

Mathematical equality

Every block has equal probability

If a category contains 17 blocks, every block has a probability of 1/17 of appearing first during a new search.

P(first block) = 1 / N

GRAPHIC 04 Wykres identycznego prawdopodobieństwa dla wszystkich bloków
05

Fresh sequence

A new seed is generated every time

Every category visit and every new search creates a fresh random seed. The sequence is recalculated so users do not repeatedly encounter the same channels at the top.

New visit. New seed. New first block.

GRAPHIC 05 Trzy wizyty generujące trzy różne sekwencje bloków
06

Independent ranking

Position cannot be purchased

Premium improves presentation, not position. Premium and Standard channels enter the same randomization process and receive the same chance of appearing in the first block.

Better presentation is available. Priority placement is not.

GRAPHIC 06 Kanały Premium i Standard w tym samym procesie pozycjonowania
07

Relevant discovery

Filters define relevance, not popularity

Category, country, language and status determine which channels match a search. They do not decide which matching creator deserves a higher position.

Filter first. Randomize second.

GRAPHIC 07 Przepływ od filtrów do losowego pierwszego bloku
08

Lightweight architecture

78 KB of focused code

The application is built with HTML, JavaScript and SQL. Indexed tables and composite indices keep retrieval direct, precise and fast.

Minimal code. Indexed data. Immediate response.

GRAPHIC 08 Schemat lekkiej architektury HTML JavaScript SQL
09

Precise catalogue

27 divisions. ~160 categories.

A broad catalogue structure combines with country and language filters to move users from a general subject to a precise group of creators.

Precision without creating a popularity hierarchy.

GRAPHIC 09 Techniczne drzewo kategorii krajów i języków
10

Premium recommended

3× cheaper than a coffee per month.

Teleport users straight to your channel and look Premium for less than per dollar a month - no hidden fees, no price increases. CANCEL ANYTIME.

Boost your visual identity. Look Premium.

GRAPHIC 10 Nowa sekwencja wyników tworząca kolejną szansę odkrycia