BtoBet break the mould for recommendation engines
BtoBet break the mould for recommendation engines
BtoBet CEO Alessandro Fried has a recommendation or two for the gambling industry at this year’s ICE Totally Gaming Show. Here he explains how the b2b firm has built its recommendation engine.
Totally Gaming: Recommendation engines are starting to be seen in many places. What makes Btobet's so special?
Alessandro Fried: Recommendation engines (RE) are very popular on e-commerce industries, but unfortunately less widespread in our specific iGaming and Betting industry.
What our analysis showed is that nowadays there are few sportsbook operators that have developed sports betting RE for their own company use, and even fewer software providers are offering only casino’s B2B RE.
We are the first company in the industry providing both casino games and sports betting recommendation engine, which has results that can be combined and used to give unique experience to the players.
Based on the algorithm used, the recommendation engines mainly can be divided on the following 3 types:
- Collaborative Filtering
- Content-Based Filtering
- Hybrid Recommendation Systems
BtoBet’s B Neuron System is a hybrid recommendation engine, which can provide more accurate recommendations than pure approaches. It is based on both content and collaborative methods and it is extremely useful to overcome some of the common problems in recommendation systems such as cold start and the sparsity issues.
In fact, recent research has demonstrated that a hybrid approach, combining collaborative filtering and content-based filtering could be more effective. For example, Netflix Recommendation System is based on the hybrid recommendation engine.
How it works:
Casino RE: all the Games are recommended according to similarity, comparing the attributes – such as theme, volatility, category - of the games and player’s preferences. In case of new player, the recommendations are strictly based on the collaborative approach: it follows a filtered method, it collects and analyzes a large amount of information on players’ behaviors, activities or preferences and predicts what players will like giving social proof recommendations.
Sport betting RE: recommendation engine, which is available for online and mobile betting, can control and suggest bets or events, included virtual, live and eSports.
It will record and analyse the bettors’ behavior according to Sport, Category, Tournament, Competitors, Market Types preferences.
BtoBet’s recommendation engine is regularly re-trained on an extremely short time frame, with any new data and players behavior, fully updating the system and every single player’s profile precisely.
TG: How long have you been developing your recommendation engine? What were the biggest challenges?
AF: We have been working on it for a full year, but as per our company philosophy we will never stop updating his system to keep it always ahead in the market. Honestly, having our platform based on Artificial Intelligence, it was already acquiring, analysing, filtering and re-organising a huge quantity of player’s data with the objective of managing automatically fraud and risk and marketing campaign. So we adapted the algorithm for the recommendation engine purposes, following a similar logic, technology and process. It needed just time to be properly developed by our efficient IT team.
TG: How tricky is it to integrate sports betting into a recommendation engine given that events are far more fluid than game types?
AF: The tricky aspect for our development team has been to reach the precision of the algorithm that rules the process nowadays, no matter if it is for sports betting or games.
Despite that we managed to reach the highest engineering levels: the automated engine avoids manual intervention and is extremely accurate. BtoBet’s intelligent algorithm tracks players’ behavior and uses collaborative filtering to provide the perfect suggestions, for the most appreciated games, for each player or segment of players.
TG: Why should operators use a recommendation engine? What advantages does it bring?
AF: According to a Forrester study, 15% of visitors admit to buying recommended products.
Cross sales and upselling efforts influence the increase of revenues. Recent researches said also cross-selling techniques increase bets by 20% and profits by 30% and confirmed that more than 73% of users prefer personalised products experiences. Why this should be different in our industry? Any solution that provides a compelling experience to the player and makes him feeling special through personalised offers, that shows tailored products’ offer and give better options to players, deserve a secure success for all the operators. Products with better features encourage players and bettors to spend more and increase the average spending value.
The possibility to offer cross-sell and up-sell complementary games and bets to users, based not only on their behaviours but also trough Social proof widgets - like “Similar Players segments also point this or played this” - show recommendations based on the wisdom of peers crowd and use social proof to engage visitors, more than any campaign.
TG: Can it be used in a retail environment?
AF: In the countries where we are expanding, it can transform the betting shops’ consumer involvement in a totally enjoyable betting experience, combining the proximity to the shops with the possibility to play in the real shop. Statistics also confirmed that recommendation tools considering location/time to further personalise the gaming experience make product suggestions more relevant and contribute to higher conversions. At the moment, we are finalising a technology that is going to extend the recommendation engine also to the retail environment.