Hey. Just a quick question for you: how much do you know of Yandex's social networks? I'm pretty sure most will answer 'zero'. Someone may recall the recent 'Aura', which is either in testing or frozen. Someone with a good memory will remember older things like the Ya.ru blogging platform (which Yandex was deliberately killed), the My Circle social network (which Yandex sold) and other smaller projects of the Russian giant.
And here's another run called Sloy. Its closed testing was announced half a year ago, and in mid-October there was an official release. True, so far only on iOS, the version for Android will arrive later.
The name 'Layer' is no coincidence. In the application, you can upload short videos, like the same 'Tiktok', but different layers can be superimposed on them: an AR layer with the addition of all kinds of effects and masks and a clothing recognition 'layer'. And if the first does not surprise anyone, the second looks quite interesting. The creators trained the algorithms of the 'Layer' on pictures from the Internet, so that he could learn to highlight individual clothing items, and, most importantly, their brands.
Imagine. Put on your new Balenciaga sneakers, Fenty hoodie, Balmain jeans, and stomp into your new ciderreria for a Poiret chat with a friend. And along the way, shoot a boomerang with your bow and upload it to Sloy, which will allow your subscribers to learn the ingredients of your image and find them on the web.
Did the previous paragraph sound like a ridiculous set of words? Don't worry, it's just that you are not the target audience of Layer, because the application is designed for fashionable youth (from 12 to 25 years old). But in general, the idea is sensible – sometimes you see something on a person and want to find something similar for yourself, but you won't approach every passer-by with a request to 'explain for a piece of clothing'. Or write in the comments to the video: 'Where did you find the T-shirt?'. Here, for a person, this problem is solved by an algorithm, which, hypothetically, will also be able to pick up something similar for him on sale.
One can imagine quite great prospects for such a service. Firstly, it's just a marketer's paradise, the same commercials take on a new dimension immediately. If the video has hooked the viewer with something specific, then he will have the opportunity to go to this product directly from the video screen. And after all, all this can be transferred not only to clothes and shoes, but also to any other categories of goods, if only their recognition in the application works well. And how convenient it will be for the same marketers to track various metrics: what things were most clicked on, how many of these transitions turned into real purchases, how many people have bookmarked an item or sent a bow and a bunch of other things to friends.
Secondly, it is very convenient for the buyer. Again, it will be possible to buy many of the things that he sees in the videos, or see recommendations for visually similar things. If, for example, a coat from Gucci is too expensive for him, then he will be able to see similar models on the conventional 'aliexpress'. And it happens that in general you like a thing, but some of its separate elements do not cause delight, and you want something else. As in the film: 'the same, but with mother-of-pearl buttons'. Technology will help here too, especially if there is some kind of 'similarity percentage' between products.
And if you 'tie' sensible augmented reality to this whole business, then you will be able to try on different products and images on yourself online, a kind of 'virtual fitting room'. Here, all sorts of new types of sensors used in modern smartphones will not be superfluous: radars, infrared point projectors, camera arrays working together, etc.
Similar ideas have been in the air for a long time, and we have already seen attempts to do something similar. Think of the home improvement service Houzz, which allows you to tag products from an internal marketplace in a photo and add them to other pictures. Or the IKEA Place application, with which you can place virtual goods from this store in a real room. Or the ZOZOSUIT project from a large Japanese clothing store Zozo Town, which was supposed to very accurately measure the owner's body and select the most suitable items for him. However, it failed miserably, and in general, nothing like this in the field of clothing has so far become real.
And here we come to the problem of implementing a new project from Yandex. To put it mildly, then everything I described above does not seem to be the closest prospect. Since the guys have so far only implemented iOS – the application, and I have no devices from Apple, all my impressions are based on a brief study of the application on a friend's smartphone. And I can't say that I was just delighted, since even the basic functions are still weak in it. Quite often, it correctly guesses the type and color of the item, but does not recognize the brand, even directly applied to it. Tech support promises that this is a temporary flaw and that brand recognition will improve. But something tells me that if in six months of testing the application has not learned to recognize quite famous brands, then the problem will not be solved overnight.
The Sloy project is headed by Daniil Trabun, media director of the famous Yandex Zen. At one time, Zen threatened to make a big breakthrough in recommendation services, but something went wrong, and he turned into a verbal trash heap of rewriting other people's articles, yellow headlines and worthless texts. In the early days of The Layer, it was argued that it was a “startup within a Zen startup” and that it was developed by the people involved in Zen. And the AR component of the application is completely based on the technologies of the Belarusian startup Banuba. All this suggests that the project, of course, can shoot (like Zen once did), but the probability of this tends to a minimum. And the chosen target audience is often insolvent.
Yandex's device is such that all services within it are, in fact, separate structures, united only by the ecosystem and the parent company. Some of them work and bring money ('Search', 'Taxi', 'Market'), some are not very profitable, but important for the ecosystem ('Maps', 'Kinopoisk', 'Music'), and which Some are in the start-up stage and have very little support ('Zen', 'Drive'). I suspect that the head office itself does not really believe in the success of 'Layer' and does not see it as a promising service, which is why development is proceeding so slowly. This is one of the first serious attempts by the Yandex project to come out in the field of augmented reality, and if it were critical for the leadership, then it would be fully supported.
Obviously, the guys did not even study the possibility of entering foreign markets, since they will not be able to offer anything very new there. And the name of the project is not the most pleasant for an English-speaking user (see Urban Dictionary). This indicates very little preparatory work before launching the project. In Russia, at first, you can rely on the popularity of Yandex, but this will not help in the long run.
What do you think about 'Layer'? Should he choose a different trajectory of development and become more popular, or is it an initially stillborn project? Share in the comments if you have already used it, what impression it left and what you think about its future future.