The algorithms of social networks and digital platforms offer us personalized content related to our interests, but they threaten to increasingly limit our vision of the world. On the other hand, the suggestions that family, friends or colleagues can make may not suit our tastes so perfectly, but they can help us discover new horizons. In the middle of both worlds, the “algotorial” emerges, a concept that proposes a recommendation born from the symbiosis between the algorithm and the work of an editorial team.

This is the proposal launched by the Spanish streaming platform Filmin in an entry on its community blog signed by Helena Oller, Director of Strategy, and which was previously published on October 6 as an article in the magazine Retina. In it, he analyzes the value of combining technology and human judgment to offer truly enriching experiences and explains the formula that Filmin applies to offer personalized content to its subscribers.

Thus, it defends the usefulness of recommendation algorithms in the face of a volume of content that has become practically unaffordable for viewers. It ensures that, to carry out their function, the systems combine multiple signals, such as what has been previously viewed, what has been searched for or whose viewing has been abandoned, what users with similar behaviors choose or the popularity of the title. The goal, he says, is to reduce uncertainty and anticipate what the user is likely to choose.

However, they carry the danger of making probability the sole criterion for recommendation. “If everything we are shown is deduced from what we have already done, personalization risks returning us to an increasingly narrow version of our own taste.”, explains Oller, who also warns of the role that artificial intelligence can play in driving this logic towards hyperpersonalization.

As an alternative to this predominance of the algorithm, the “algotorial” model is proposed, in which the work of the platform’s editorial team comes into play. This, as he assures, begins long before the recommendation itself, when deciding which titles are acquired, choosing the right time to release them or generating additional content to understand what makes them special and how they relate to other titles, themes or authors. An example of this philosophy are, for example, the videos that Filmin periodically publishes and in which Jaume Ripoll, Co-Founder of the platform, announces the new features that are being added to the catalogue.

Also part of this work is the agreement that Filmin has established to bring part of its selected catalog to Netflix Spain. As announced in October, the collaboration is articulated through monthly collections of films and documentaries chosen by the Filmin editorial team, which will be available in a specific section within Netflix. In this way, on the first day of each month Netflix will incorporate a new collection made up of ten titles licensed and selected by Filmin, benefiting from the editorial curation of the Spanish service.

“It is artisanal and continuous work. The team watches the films, interprets what is happening in the world, listens to the community and builds connections that are not always evident“, says the Director of Strategy of Filmin. “When a row of content tells you ‘Based on what you’ve seen,’ the algorithm uses your history to recognize a starting point. But the movies it shows you don’t have to be there because other users have followed the same route. They are there because someone on the editorial team has manually linked them to what you saw”.

This vision does not intend to distribute the work or replace the algorithm, but rather to make it coexist with the work of curation to address what Oller considers one of the great current challenges: beyond producing content, selecting what deserves to be seen in the midst of the excess. “This selection requires judgment, and technology is what allows it to reach each person on a scale that would otherwise be impossible.”, he assures in his text.

In this regard, he explains that the “algotorial” model is a two-layer architecture, that is, the editorial criterion builds the relationships and the algorithm personalizes access to them. Thus, the human team selects, contextualizes and relates the works in a kind of map, while the technology uses the knowledge it has of each user to decide at which point on said map they can begin to explore it.

“A model based primarily on aggregate behavior says, ‘Those who watched this movie also watched this one.’ The algotorial model can say: ‘Since you have seen this movie, we show you another one that our team has related to it for a specific reason‘”, says Oller. “The algorithm knows where you are starting from; The editorial criteria propose how far you could go”.
So it’s about combining probability with relevance, context, diversity and discovery.

However, he admits that he is not infallible, since editorial work can make a bad choice or bet on a title that does not quite find its audience. However, Filmin’s management ensures that the key is that when the system fails, there is someone behind it to review and correct the process and explain the strategy behind the decision. something that he assures does not happen with opaque systems.

That is why he argues that the great contribution of the “algotorial” model is trust. “We do not trust a system that only predicts; We trust in a criterion that has a signature, that can be wrong and be held accountable for it. That signature is what, later on, gives us permission to take risks”, he explains.

Along these lines, it demands greater human intervention in the recommendation processes to prevent visions of the world and reality from being reduced to that which aligns with our ideologies, certainties and convictions. Thus, he considers that the future of the recommendation should combine the scale and relevance of the algorithm with the context, emotion and judgment of people. And these people can be the editorial team of a platform or other users who share recommendations.

“Therein lies, I believe, the real opportunity: not in competing with the immediacy of a viral video, but in building a community that heals, comments and dares to recommend what is not obvious.”, points out Oller. “When you trust someone, whether it’s an editorial team, a community, a friend whose taste you respect, you dare to go somewhere you would never have gone on your own. A system trained to reduce uncertainty will tend to move you closer to a safe choice. Trust can do the opposite: give you permission to take risks..”