Promoted feature: From investor deck innovation to real-world problem solver

After countless years of enjoying the craze that is ICE, and four years of championing AI at the world’s biggest egaming tradeshow, something was different this year, and not least due to the first face masks heralding in the craziness the world, and gaming, is now in.

Wherever we went and whomever we met, we noticed a refreshing awareness about AI and the benefit it has on user experience and business, and the willingness of operators and institutions to try AI as a tool to help them with those challenges mounting in front of the industry. Gone are the days where most wanted AI as the ‘cool’ thing in their investor decks, but rarely put it into their roadmaps.

But, while operators are slowly warming up to the business benefit of an offering tailored to every user, and the ability to accurately predict acquired users’ future value, the industry is experiencing an even faster trend of maturing markets and tightening regulation, especially on wagering, marketing and promotional mechanisms like advertising and loyalty programmes.

Reactive initiatives like VIP programme overhauls may pose a danger pushing everything else down the roadmap again, although now is the most important time to use new mechanisms like AI.

Why

We argued that current promotional retention measures were wrong for players, and unsustainable for operators in the first place. On the one hand, operators are driven into rapidly escalating bonus bidding wars to keep users on the site. On the other, there are users, which they have acquired for too much money from acquisition sources in the first place, and – unlike e-commerce – without the metrics to measure conversions, and mostly with inaccurate capabilities to quantify and predict value.

Bidding wars are not only leading to an increasing number of accounts per user, but teaching players to shop around for bonuses instead of getting to like an offering for its brand and experience. It also drives users to play more often, and the patterns trend towards pathological gaming if pushed aggressively. The bonus becoming the primary driver to play, opposed to the offering, is exactly the reason why regulators across Europe (and soon the world) will limit money as a retention mechanism.

In a climate where operators are losing all or parts of bonusing, aggressive promotions and VIP programmes as monetary retention mechanisms, other tools are needed. The so often cited ‘innovation’ at ICE is a fluffy concept to tackle those challenges. AI is here now and can help address the precise issues driving operators and users into a malicious cycle.

Acquisition marketing

As more markets regulate across the globe, operators have access to new channels such as paid search and social media. As such, operators are nowadays able to acquire customers and not solely rely on affiliates to drive new traffic. However, regulation comes at a cost, with higher taxes and licence fees resulting in lower margins. Having the ability to calculate the ROI accurately, and quickly, becomes more important than ever before. You must ask yourself: can you wait three to six months to determine if a campaign is profitable? How much are you overpaying before you can determine a campaign is a good investment, and could that cash be spent elsewhere?

Using deep learning, Vaix has an all-round solution which predicts the future value of a campaign (or player) in seven days of that campaign running, with unparalleled accuracy compared with conventional BI & Analytics. This means the future value predictions are a reliable tool for operators to use to stop bad campaigns, renegotiate the close to breakeven deals, and – most importantly – invest more in the profitable deals.

As an example, one Vaix customer who put this into practice was able to stop over 25% of their campaigns within the first two month of them running, equating to 20% savings in the marketing budget in the first month. In addition, 18% of deals were taken back to the table and renegotiated. In summary, almost half the deals were either stopped or improved.

Retention marketing

One of the biggest challenges in today’s world is keeping a hold of a player you’ve just acquired. Punters have many options of where to play, and with all the lucrative sign-up bonuses out there, it’s very easy to ‘site hop’. The traditional approach in the egaming world is to be reactive; wait till the customer leaves (or you think he has left) and then offer him a bonus to come back and play. Not only does this drive up operators’ bonus costs, and eat into those low margins, but the likelihood a customer comes back is between 5%-10%. The solution here is churn prevention: reaching out to a customer when they are thinking of leaving an operator, increases the chances they stay to 15%-20%. As a side benefit, operators’ bonus money is spent on the right players, leading to savings in the CRM budget.

One US sportsbook leveraged Vaix’s AI to help identify users at risk of churning on their sports betting app. Using Vaix to identify this user cohort allowed the operator to devise a test to see how effectively at-risk users could be retained, as well as win back users who had already churned. Test results demonstrated that 16% of at-risk users returned when identified accurately, and engaged through proactive and targeted marketing tactics, versus 9.6% who returned on their own, a 67% improvement.

Using Vaix to identify at-risk users has enabled the company to be more proactive and strategic in devising effective targeting methods in improving retention.

Personalisation

The real reason why users love a company is the product: WHAT it is, and HOW it is. It’s ‘easy’ for companies like Apple to offer a product like the iPhone, which differentiates itself to a higher degree by WHAT it is: a phone doing phone calls, internet connection and its own map app. But it also differentiates HOW it does it: design, enabling other apps to thrive in the app store etc.

In egaming, the WHAT is often very interchangeable between operators: while some differentiate with features like real-time streaming, special market offering or exclusive games, the 1,000s of sport markets and slots games are mostly homogenous. This makes the HOW more important: brand, elegance of the UX and features around the offering. This includes how easy it is for a user to find their favourite markets and games amid a broad range of options.

Solving the information overload users are facing everywhere in the world, and stopping making sports and casino front pages look like Bloomberg terminals, is one of the biggest improvements users can get.

Starting with the most important use cases, Vaix.ai provides deeply personalised sports league, event, market or casino game recommendations, even to newly signed-up users. Even pre-login, recommendations can be tailored based on location or other parameters. A similarity AI model provides the most closely related events, leagues and sports a user is already enjoying.

And it works. Our experience with every one of our customers shows that deep learning-based personalisation by far outperforms any editorial, rules-based or conventional AI system in activity metrics like clickthrough, bets/wagers and turnover, but also profitability metrics like acca length and encouraging betting across the long tail: punters using the AI service did not only place twice the bets, they also placed those across a significantly higher number of leagues then the benchmark, during football season and even more after.

Additionally, in off-peak seasons users adding a bet into their bet slips, actually placed those bets 10% more often than in the benchmark.

And what proved to work in off-season is now holding true in the direst of times: the Covid-19 effect removing dozens of sports events and leagues of daily fixtures and betting offerings alike. Instead of showing a bet365-esque Bloomberg terminal with Russian hockey and Cuban volleyball to all users, we have seen AI align recommendations of what events are left, to the closest matches of users’ previous betting preferences, between sports or between leagues.

Integration does not have to take long, as nowadays platforms are service-oriented and can be connected. For example, Vaix’s integration with platforms like Kambi and SG’s Openbet reduces integration effort to a minimum on the operator side, with the APIs doing the hard work. The Vaix AI then does the rest without the operator having to hire a team of expensive data scientists who need to research, build and maintain a complex development setup to run and serve deep learning models and outputs.

As Amazon already showed us 10 years ago, user value grows the deeper personalisation is integrated into an offering. For example, in one customer project, adding truly personalised recommendations into emails provided a significant uplift vs conventional editorial selection.

In today’s world data is gold, and being able to mine it is just as hard as mining gold. Vaix is helping operators bridge that gap and offer services which enhance the user experience, as well as generate higher profits, through personalisation and better marketing and retention. At Vaix we believe egaming operators can achieve this, without spending the resources of an Amazon to get there.

Andreas Hartmann, CEO and co-founder of Vaix, has over 20 years of experience running product in gaming, fintech and search. As the product director at PartyGaming/bwin.party, he built what is now GVC’s core gaming platform and merged the world’s number two and number four poker networks. Before that, at Yahoo! in Silicon Valley, he built one of the world’s first contextual search query understanding algorithms. He holds a diploma in BA from LMU Munich and four patents.