Yandex Blog

Yandex Provides Hyper-Targeting Opportunities to Advertising Clients

Yandex unveiled a new service for businesses advertising their products on the company’s websites, as well as websites in the Yandex Advertising Network, at its annual e-marketing conference Yac/m. The new service, Yandex.Audience, allows companies use their own customer information to segment audiences for hyper-targeted advertising, as well as target their ads to existing groups of customers to boost upselling campaigns, improve retention, and increase average spend.

After uploading customer information, such as email addresses, telephone numbers or device IDs, to Yandex.Audience, an advertiser receives anonymised IDs identifying their customers among visitors on Yandex’s pages and the YAN websites. These people can now be personally targeted with offers relevant to their previous customer experience – a pair of shoes matching the bag they bought a week ago, or a special loyalty program to recapture lost customers. The same data can be used to identify lookalike audiences – groups of people who exhibit characteristics similar to those of the existing customers and are likely to be interested in the offers that the existing customers were interested in – and target ads to them.

Yandex.Audience is available in English and Russian. To start creating hyper-targeted ad campaigns, an advertiser needs to sign into their account with Yandex’s auction-based service for contextual advertising Yandex.Direct and enter their customer data in the .txt or .csv file. The IDs returned by Yandex.Audience can not identify any individual user, but can be used for delivering personally targeted ads.

Medium- to large-scale businesses such as retailers, banks, car dealers, insurance companies, possessing reasonably large amounts of customer data and striving for customer conversion will appreciate this service most. The amount of data required for hyper-targeted advertising starts at 1,000 records. There is no upper limit for the number of records that can be uploaded to the service, neither is there a limit for the number of types of audiences or hyper-targeted advertising campaigns.

In addition to hyper-targeting opportunities, Yandex.Audience will soon be providing tools for marketing analysis. Yandex's proprietary behaviour analytics technology Crypta, which can identify web users’ interests, age, gender, family status, and even if they have a car or a pet, based on their behaviour online will soon be added to the service. Thanks to this technology, advertisers will be able to use social and demographic statistics of their audiences to plan their marketing strategies.

Yandex Unveils First Browser with Infinite Personally Targeted Recommended Content

Yandex builds personalised content recommendation technology Zen into Yandex Browser on all platforms in 24 countries and 15 languages. Based on the latest developments in artificial intelligence research, Zen recommendation technology uses the company’s vast global web index to pick stories, images, videos and other content for each individual user and offer it them right in the new tab of Yandex Browser.

zen 3.002 (1).png

The intelligent content discovery feed in Yandex Browser delivers recommendations based on the user’s location, browsing history, their viewing history and preferences in Zen, among hundreds of other factors. Zen uses natural language processing and computer vision to understand the verbal and visual content on the pages the user has viewed, liked or disliked, to offer them the content they are likely to like. Yandex’s recommendation technology Disco, based on the company’s machine-learning algorithm, MatrixNet, helps Zen choose which suggestions to offer to the user at any given point in time. Targeted to identify the user’s personal long-term interests and cater to them, Zen also delivers content not directly related to their immediate preferences. The more the user interacts with Zen, the better are the chances that they will see serendipitously interesting content.

‘With all the vastness of information available on the internet, something genuinely interesting isn’t easy to come by. Zen helps solving this problem,’ says Victor Lamburt, head of Yandex Zen. ‘It points each user to what’s interesting specifically to them. This is the future for all web browsers: providing personal internet experience and helping people discover something new’.

The infinite personally targeted content feed in Yandex Browser gives web users an opportunity to discover something they appreciate, but wouldn’t have found it otherwise. To start exploring this new internet experience, all one needs to do is download Yandex Browser and give Zen some browsing history to work with. Alternatively, liking or disliking a few websites on Zen’s start up page will help it understand your preferences on the outset. Users can also alter the type or topic of content they are offered later on by choosing to view more of similar content, less of it, or block specific sources altogether.

Zen first appeared as an experimental feature in Yandex’s launcher app for Android in Mexico and Brazil in 2015. The average time the users spent viewing Zen’s recommended content has increased since then from only 5 minutes to 20 minutes in May 2016. Zen is currently available both in Yandex Launcher and Yandex Browser for iPhone, Android mobile devices and Windows PC and laptops.

Yandex’ personal content recommendation technology can also be easily integrated into third-party mobile applications, such as browsers or launcher apps, and offers great monetisation potential for OEMs, app developers, and mobile carriers.

New Yandex Service Uses Machine Learning for Hyperlocal Weather Forecast

Machine learning is Yandex's core technology. We’ve long been using it in almost all of our services — to answer users’ search queries, for machine translation, ad targeting, personal recommendations, and plotting routes on maps, among others. Since last year, our MatrixNet machine learning algorithm has been utilised for the optimisation of business processes in real enterprises — weopened Yandex Data Factory for this purpose.

Today we announce yet another application of machine learning in a new field for us — weather forecasting. For this we have developed our own forecasting technology Meteum, which will now be used in the web service and mobile application Yandex.Weather available for iOS and Android.

Basic weather forecasts are traditionally constructed using the Navier-Stokes equations. Models for describing weather are extremely complex, as they depend on a multitude of factors. Programs for their calculation consist of hundreds of thousands of lines of code and run on huge supercomputers. Nonetheless, they still make mistakes, so their forecasts need to be fine-tuned. Besides that, the complexity and resource-intensiveness of traditional calculations results in a situation where forecasts are made for relatively large regions and cities. Constructing a precise forecast for, say, a small village would require taking into account a large number of local factors – such as, solar radiation, phase transitions of water vapour, or thermal radiation from the soil. Performing this task using traditional methods is not much less resource-intensive than for a large city, while the number of people using such a forecast is much lower.

Using machine learning allows collating a large volume of historical data about forecasts and actual weather, identifying causality in forecasting errors and correcting them. This is quicker and easier, as it doesn’t require factoring in laws of nature for each new forecast, but simply corrects traditional mathematical models and localises the forecast down to specific latitude and longitude. That’s exactly what Meteum does.

2015-11-24-meteum-640px-72dpi.png

Our new technology uses traditional meteo models to process the initial data, and works with intermediate results using Yandex’s machine learning technology MatrixNet. To calculate the weather, Meteum constantly compares forecast with actual weather conditions — more than 140,000 times a day. To learn about current weather conditions, we use meteorological station data, as well as weather information from other sources indirectly indicating the situation — about 9 terabytes of data every day. One of the sources is our users, who can let us know about discrepancies between forecasts and real weather conditions via the app. The more data we receive from them, the more precise Meteum’s forecasts will become.

Screenshot_2015-11-23-15-12-33.pngfeedback.png


Meteum calculates a new forecast each time a user consults Yandex.Weather on their desktop or mobile device. It locates a person and shows them a fresh forecast for precisely that location. The user can choose another place and time for the forecast to see what the weather will be like around their office in an hour or if it might rain when they go out of town in the evening.

Meteum currently works in 36 regions of Russia, with a possibility to expand to other regions or countries.

Tune in to Yandex.Radio

Today we’ve announced our newly hatched personal ‘jukebox’ media player Yandex.Radio. This service will be appreciated by anyone wishing to always hear ambient music matching their personal preferences, or serving as a backdrop to anything they do or how they feel at any time of their day or night. The user interface of Yandex.Radio is simple and intuitive – it just lets you play the music that you will enjoy.

Yandex.Radio webversion.png

Yandex.Radio shares a catalogue of more than 20 million tracks with Yandex.Music, Yandex’s recommendation-based music streaming service, which was first launched in 2010 and re-launched after a major overhaul in 2014. While Yandex.Music helps listeners discover new music based on their interests and preferences, Yandex.Radio offers them the music that matches their current mood or activity.

Out of 10 million people currently tuning in to Yandex.Music every month, one million do this on their mobile phones. This million of people just like to have some background music while they are working out in a gym, driving to work, or chatting at a party. Alternatively, people choose their background music to match how they feel at the moment – cheerful or vigorous, moody or relaxed. Yandex.Radio is a service for those who like to choose their music based on their current activity or state of mind. The new service taps into the catalogue of Yandex.Music to offer these users more than a hundred stations to choose from – not only depending on their mood or current activity, but also on a genre or time period of music.

Radio_1.pngoffline.png

Yandex.Radio is now available on desktop, as well as on iOS- and Android-based devices. The personal recommendation technology in use on the Yandex.Music service has also been implemented in Yandex.Radio to play music based not only on the listener’s settings, but also their personal preferences, streaming history, and previous likes or dislikes.

The new personal jukebox player is currently available free of charge only to users in Russia. The revenue from the service will come from audio advertising, which is a novel format for our business. Just like traditional radio broadcasting, Yandex.Radio will feature audio ads, which in the future will be complemented with a graphic image or text both in the app and on the desktop service. Advertisers on Yandex.Radio will be able to enjoy all the advertising and audience targeting possibilities available to them on Yandex.Direct, as well as banner advertising.

Well, search has been personalised already. How about the rest of the internet?

At Yandex we’ve long been striving to tailor search results especially for every individual user – and we can already do it pretty well.

Our Personalised Search fetches results and delivers search suggestions individually for each user based on the many things we know about them – including their geographical location, language preferences, search history and clicks in search results. The user's search history tells the search engine what may be currently relevant for this particular user, and whether he or she would appreciate getting search results in English, for instance. Our MatrixNet machine-learning algorithms allow our search engine to look at users as live, multi-faceted human beings: gender, age, sphere of activity and domestic status are just some of the qualities it knows how to consider when delivering personalised search results and suggestions.

Well, naturally we couldn’t stop there, and we started thinking about how to take this great idea one step further. Once we’d developed personalised search, another idea arose: if we can personalise search results, why not personalise the whole internet? Introducing …. (drum roll, please) …. Atom!

Atom is one of Yandex’s new technology concepts. It allows any web resource to be adapted (or personalised) for nearly any person, even if they have not visited that web resource before but have a search history at Yandex.

For example, a site selling package tours is more likely to satisfy a user (and make a sale) if its main page only shows those tours that are likely to be of most interest to that user, based on his or her past behavior online. If a site can work out how to reconfigure its front page or catalogue according to the interests of any given person – and deliver what’s needed right at the start -- then it follows that the person will return to the site again and again.

How does it all work? We “talk” to a site through an API, telling the site what to show, in what order, in what priority, for each individual. We’d like to emphasise that we don’t give third-party sites any private information about the user – none of their cookies, nor their search history. We process all that information ourselves.

At present Atom exists on the level of a concept that we will be developing over the next few years together with the internet community. It’s an ambitious plan, which will work only if it gets the support of everybody – users, web site owners, web masters. 

And who wins? First of all – users, who will get only relevant and useful information on their PC or tablet or smartphone screen. Imagine a newswire website where all new items are interesting for everybody. Nothing to be left unread. Or an e-commerce service delivering not only recommendations based on their own statistics, but considering much more extensive behaviour of a user in the internet. A personal internet – for each, their own – is coming. That would be the huge shift in upcoming years or even decades. Stay tuned!

It May Get Really Personal – we have rolled out our second-generation personalised search program

We all know what it takes to understand another person. It’s a lot. Even if all you need to understand is what a person is looking for online. We have been trying to do this for years. A person’s interests and preferences give a good clue as to what they want to find. We used to look into a user’s search history as far as a few months back to choose for them the search results that would be most relevant to their scope of interests.

Now, we have added to our search algorithm a search history of a few seconds – searches within the current search session. We can now deliver results and search suggestions based on the ‘full picture’ of the user’s search behaviour.


Updating our knowledge of users’ interests once a day allows us to understand their more-or-less stable interests, such as a love of books or football, or that they speak Russian and live in Saint Petersburg. More than half of all searches on Yandex, however, are about something that interests the searcher at the very moment of searching and stops interest them the moment after. To be able to cater to such momentary searches, we now analyse search sessions in real time.

Search queries begin to influence search results within seconds. The search engine can figure out whether a person is looking for a book or a film even before they have finished typing [The Great Gatsby] in the search box.

 

 

 

With the new personalised search program, we can offer relevant results even to those web users who don’t have any search history on Yandex. To instantaneously react to changes in users’ search behavior, we created a real-time data processing system, which processes more than 10 terabytes of data a day, continuously correcting its knowledge of users’ needs.

To make search results as personal as they can be, we first learnt to take into account users’ language preferences and permanent interests. Now, we have learnt how to tell current search intents of our users from their search footprint and give them what they are looking for. Personally.

Search results based on current search interests are available on all Yandex domains globally for all 93 millions of our users. Personalisation that takes into account long-term or medium-term interests works best for searches in Russian, but, just like all previous editions of Yandex’s Personalised Search, the latest version will learn to fetch personlised search results in other languages as well as it receives more search queries in these languages. 

Read more about our Personalised Search program in Products and Technologies.

Future of search, who really owns your content and why personalisation can backfire

Watch Yandex's CTO discuss all these and other important issues at DLD's Business and Technology Conference, TES 2013

Ilya Segalovich, Yandex CTO talks to Gerhard Thomas, Managing Director of Burda Direkt Services about the future of search technologies, who should own the user-generated content and who really owns it, and why strong personalisation is bad.

 

Yandex has rolled out personalisation for its search results and search suggestions

The search engine now fetches results and delivers search suggestions individually for each of its users based on their interests and preferences.Web users in Russia, for instance, typing on Yandex the query [nevermind] might just as likely be looking for Nirvana's album, as wishing to find out what this word means. Personalised Search would know the difference and would act accordingly.

Yandex’s Personalised Search is based on the user's language preferences, their search history and their clicks in search results. The user's search history tells the search engine what may be currently relevant for this particular user. Someone searching on the internet for free software, books or music is very likely to be interested in this type of content as such. Those users, who frequently visit websites in English, may very well appreciate search results in this very language. As personal preferences tend to change over time, Yandex considers only the relatively fresh search history spanning the period of a few months to offer the users personalised search results and make personalised search suggestions.  

Search suggestions

In contrast to regular search suggestions, personalised search suggestions are targeted individually at every single web user. When guessing what a searcher might want to find, Yandex suggests potential search queries, based on what other people with similar online preferences looked for on the internet. The search engine classes everyone into one of about 400,000 user groups with more-or-less common interests. This classification is fluid – it changes for every user according to the changes in their online behaviour.

In practice, web users repeat about 25% of their search queries and often click the same search results. This behaviour can be interpreted as going to frequently visited websites or viewing popular or personally relevant web documents. Yandex offers users a shortcut to favourite content by showing them recently made queries and their favourite websites in the search suggestions when they type the first letter of their new search. 

When choosing search suggestions for a specific user, Yandex also looks at what searches have been previously made during a whole search session. So, the search engine would know that Christopher Lloyd would probably be a better search suggestion for 'c' in the search box than any other if the searcher has looked for Back to the Future before.

Personal results

Other than making personalised search suggestions, Yandex helps its users to achieve their search goals by providing them with the most relevant search results. In doing that, the search engine uses a special personalised ranking algorithm, which it recalculates on a daily basis according to the ever-changing interests and language preferences of each user.

Personalised ranking algorithm allows the search engine to understand how well each of the fetched results matches the user's expectations. Search results are evaluated and ranked according to their usefulness for a particular web user. The same search query made by two different people would trigger the same results ranked differently to match their individual interests. An inveterate gamer and an art film enthusiast, for instance, will see on top of search results links to websites that are relevant to their respective interests even if they both look for 'Stalker'.


  

The personalised search algorithm works only for those queries that may be related to the user's search history. Yandex won't be able to use a personalised search formula based on the user's preferences if they are looking for something they have never searched before. This formula, however, does work, on average, for about 75-80% search queries. Which queries trigger personalised search depends on the user and on their personal search history.

Personalised search is enabled by default for every more-or-less frequent search user. The more queries the user makes, the better results and suggestions the search engine can provide. Personalised search deactivates if there aren't enough searches on which personalisation can be based, and activates again when queries start coming. Personalisation can also be enabled or disabled manually in the search engine's settings.

Currently, personalisation on Yandex works best for searches in Russian, but as the search engine accumulates search statistics in other languages other users will also be able to feel its effect.