Yuri Litvinenko’s column on the role of data for YouTube’s competitors

Owned by Gazprom Media Holding (GPMH), Rutube can count on access to a larger pool of users than before: the government has included the mobile version of the application in the list of mandatory pre-installation on phones. But a video service that exists outside of ecosystems will still find it difficult to compete with the developments of multidisciplinary technology companies.

It’s not just about the resources that an owner can spend on a product or infrastructure expansion – for example, the same network of servers in different geographic locations. Companies that develop a whole portfolio of services are more likely to invest in a project whose results can be deployed in several directions at once. So about ten years ago, Google did it, creating the Knowledge Graph (“Knowledge Network”) – a database with information about different objects and relationships with each other.

Google positioned the Knowledge Graph as a way to improve search – it is from its data that cards with enriched answers in the issue are collected. For YouTube users, the new development brought precise recommendations. For example, if you watch a video about a specific game, the service will offer videos not just about games in general, but in the same genre or from the same developer. The dominance of Google helps to enrich this network: for example, Internet companies that want good positions for their sites in search, themselves add already marked up data to them.

GPMC is an association of media companies, not information companies, and the creation of a database of this level is hardly a priority for the holding. The experience of implementing recommendations in online cinemas (Premier) cannot be fully transferred to a service that lives mainly on user-generated content. The catalog of online cinemas is filled centrally, and films and series are already supplied with a set of data, in particular, about the genre and actors.

By comparison, VK, which is currently developing VK Video, has access to more data that can be used to improve the accuracy of recommendations. An important component that Google, among other things, tried to organize for itself is data on the social connections of users, their activity in interest groups.

Of course, creating a good video recommendation system without relying on third-party sources of knowledge about the world is quite possible; An example of this is TikTok. But due to its specificity, this service can quickly collect data about user preferences. The TikTok app collects almost all available information, from geolocation to a list of other apps on your phone. So the owners of any video service will have to fight for the user’s data as well.