Top recommendation: the art of offering customers what they want

Top recommendation: the art of offering customers what they want

ALEA’s Alexandre Tomic discusses recommendation engines: the complex set of algorithms that suggest slot titles to players that they will really enjoy and therefore keeping them engaged and entertained.

Understanding players’ likes and dislikes in detail, right down to distinguishing between types of features and slot volatility, is one of the most important aspects of the online gambling industry, and yet not a lot has been done so far in order to achieve this. However, one of this year’s major projects for operated SlotsMillion is a recommendation engine, designed for exactly that purpose.

Recommendation engines filter content by suggesting items to users based on what they have previously enjoyed, hopefully leading them to exactly the product they are looking for. They can offer users a much more personalised experience, which is why many companies, such as film and music sharing platforms like Netflix and Spotify, spend a lot of money on recommendation engines in order to provide their users with the best.

Two years ago, we developed a unique filter for SlotsMillion that allows players to browse our game portfolio based on the type of slot, theme, features, paylines, and game provider. It has been said that the best place to hide a dead body is the second page of Google results, and the same is true of this. Beyond the first four lines of slots in the filter lie hundreds of games, but only the blockbusters get pushed - despite the fact that out of twenty blockbusters there may only be three games to the player’s taste, when there may actually be three or four hundred games they would like on the site. We have over 1000 desktop slots and 500 on mobile, and yet the top 20 games represent 80% of the bets made on our website. As far as we know, nobody in the industry has explored this field yet, beyond simple game suggestions. The Long Tail principle, outlined in a Wired article by Chris Anderson and his book published in May 2006, described this concept ten years ago, and still very little has been done to explore this in the world of online gambling to date.

Next generation

We are now looking to release a second version of the filter where games can be ranked by more precise features, such as types of wilds or an autospin feature. Continuing in this direction, a recommendation engine would be the next step, but to do this we have to understand how other industries recommend content depending on what players have experienced and enjoyed before. We consider this concept in gaming similar to the way it works in the music industry: albums, for instance, could be software providers, and users can skip songs or choose not to play games. They can listen to a song many times or play a slot many times. 

If a user marks a slot as a favourite, this can obviously be taken as a declaration of their likes and helps us recommend similar things to them. This is known as the declarative method. However, there is another method, called behavioural, which takes each bet as a vote. The consumer pays over and over again to enjoy this service, while all other industries rely on a subscription paid once per month that offers unlimited access to content.

In our industry, every time players pay they scream that they like whatever slot they are playing, and the longer they play, the more they pay. This is what makes recommendation engines so important for our industry - they help us to provide a better experience for our players, which means more successful businesses.

As previously mentioned, SlotsMillion’s filter already works based on themes, software providers, paylines, features and type of slot. What we want to add, among elements such as the way the slot returns the money, the hit frequency and the max bet and win, is something we consider to be in the core of any gambler, ultimately the true essence of gambling: variance.

Risky Business

What we need to discover is the player’s appetite for risk. We want to know if the player wants frequent, smaller wins, or big, less frequent wins. Sometimes they themselves are not sure, and even when they are, they may not know which slot to choose, and this is where we run the risk of losing players. Imagine that a customer enters an Indian restaurant, asks for something mild and receives something that blows their head off; they will be unhappy. Similarly, if a customer enters with a desire to eat something very spicy, but receives something mild, they will be bored and disappointed - and both will walk away. It is very important that we give customers what they want - even when they may not even know themselves - and that means paying close attention.

We could do this by asking players directly, or we could do so by looking at their behaviour - offering them two types of bonuses, for example; one which multiplies the possible win on the next spin by ten, and another which would allow them to receive twice their potential win on the next five spins. The choice made here would help us understand if the player enjoys high or low risk games, which would help us direct them to the slots they want to play.

We want to reduce the rate of players we lose between their first and second deposit. We want to engage users with the perfect game for them from the start, and a recommendation engine will help us to provide a great experience right from the beginning.

This article has been authored by one of the members of the iGaming Industry Council, a group of 50+ industry professionals, who met in London in March to brainstorm, in an innovative Open Space format, key opportunities and challenges that the iGaming industry needs to address to ensure its success in the future. Recommendation engine was one of the topics discussed by the group. To find out about iGIC, please contact Ewa Bakun at

No alternative text provided

Clarion Gaming and Racing Post agree strategic partnership as ICE North America beckons

American Football

Nevada breaks sports betting record in 2018


Paddy Power Betfair takes majority stake in Adjarabet


Rhode Island sports handle hits $13m in first full month

Gaming Products & Services Directory

The essential directory for the gaming industry