Built for purpose: A new dawn for personalisation in sports betting

Aiming to solve the gambling industry’s personalisation conundrum is Graphyte, the brainchild of two former Ladbrokes Coral Group (LCG) executives who decided to leave the operator when the merger took place to set up their own business.

Ladbrokes Coral’s former director of sports brands analytics Damien Evans and ex-CTO Rob Davis teamed up last year to create a personalisation and recommendation platform that is built specifically for the betting and gaming sector by those within the industry.

According to CEO and co-founder Evans, Graphyte offers marketing modules with a decisioning engine powered by artificial intelligence (AI) on top of leading-edge real-time data technology. The system is built to scale to the unique demands of the betting and gaming industry, operating at low latency and built around the same core principles that power Amazon, Netflix and Facebook.

Evans explains that the ability to deliver personalised brand experiences is a top priority for operators, but it still remains one of the industry’s toughest nuts to crack. Graphyte’s CEO tells EGR Technology how the idea for Graphyte was born and how it is helping the industry to address its tech challenges.

EGR Technology: How did the idea for Graphyte come about? When was it officially launched?

Damien Evans (DE): It’s been a busy three years. When Ladbrokes merged with Coral – it was a huge merger and Rob and I were heavily involved on a number of levels. I was responsible for marketing analytics and tech, while Rob Davis, former CTO of Ladbrokes Coral, was responsible for all the tech, reliability and platforms for the sportsbooks. He and I both took the opportunity to exit as part of that merger.

We set up Graphyte in March 2018 and spent much of last year building the product. We’ve been working with clients since the back end of last year now.

EGR Technology: Why did you decide to set up your own company?

DE: Rob and I worked closely together for many years at Lads first and then Ladbrokes Coral. One of the things we tried to do was find a way of differentiating customer experience by personalising. Simple things like making recommendations to customers based on their profile.

We managed to execute some relatively simple versions of this, including at Ladbrokes back in 2015 when we launched what we think was the first ever recommendation engine for sports betting, certainly in the UK. But it was really basic, and the reason for that was that there wasn’t really any tech we could buy off the shelf that does this well in the gaming space.

There are lots of after-market solutions out there like Salesforce and Optimizely (among many others), platforms like that which do personalisation well for retail and ecommerce, but they just don’t fit the gambling model. A lot of that is down to the specifics of the product inventories, particularly sports betting, where a typical product inventory, ie. what a customer can bet on at any given time, is massive and highly dynamic. Typically in the region of 125,000 bets are available to each user, with constantly changing price and status, which makes betting a much more nuanced and technically challenging solution to build for. As a result, we found that the more established personalisation platforms can’t really cater for this dynamism.

Co-founder Damien Evans outlines Graphyte’s ability to offer marketing modules with a decisioning engine powered by artificial intelligence

EGR Technology: Can you tell me more about how you built the Graphyte solution and how it works?

DE: Graphyte is a personalisation and recommendation platform for both betting and gaming – a true multi-vertical solution. Graphyte is unique in that sense, in that we can optimise the customer experience for sports betting as well as casino, slots and bingo and lotto. We have got a dedicated solution for every vertical, and this was driven by us coming from a multi-vertical operator background. Why have separate solutions for different verticals when the customer experience should be seamless across all of them?

We’ve also built a triggered messaging capability, so if, for example, a user makes a pre-event bet, Graphyte has a triggering functionality that can send a message to that user with a highly-targeted in-play bet on that event. It’s a very strong proposition and given the level of targeting, very engaging for the customer.

Graphyte works by connecting to either browser events, so things like adding a bet to a bet slip, playing a casino game, depositing and so on from the front-end, or from the sports and gaming back-ends as well – and in an ideal world from both sources.

We are able to connect to major back-end platforms, pulling in what customers are betting on, alongside ingesting what they are doing on the front-end of the website. Are they abandoning bet slips, for example, or removing legs from multiples. We can then do automated messaging as a follow-up. From that data we build profiles of customers and we match that up with what’s available in the product inventory by running a suite of algorithms in real time within our cloud environment.

Which brings me to the AI component. AI is an integral part of Graphyte. We use some really sophisticated machine learning and AI to generate personalised content and manage the decisioning process, making sure that decisions that the engine makes are automatically updated with the results to` continually fine-tune performance. Graphyte is running multiple algorithms at any one time, so we have an additional AI layer that sits on top of the individual models to decide which algorithm, which bet and which game to use for that customer at that time.

The platform’s AI layer learns from what happens. For example, if we make a content recommendation for a user on a site, and they like it, we make sure we understand that and that goes into influencing future predictions and recommendations. It’s a self-learning engine in that respect.

That said, Graphyte is far from being a black box-type of solution. Operators can pick and choose which algorithm or content type to go with, as well as upload their own custom models like responsible gambling, harm-minimisation, churn prediction, cross sell etc and deploy on-site or via CRM in real time.

EGR Technology: What kind of skills/knowledge did you bring with you from your time at Ladbrokes to the new company?

DE: At LCG I was responsible for marketing tech, personalisation, analytics, business intelligence, data science, all that kind of stuff. It was a mixture of what kind of segmentation we would use, how we would target customers, customer relationship management (CRM) communications and personalising the on-site experience. A big part of my remit was personalising a site, mobile or desktop and trying to drive more personalised communications CRM across gaming and sports.

Rob was responsible for platform stability, new product integrations and managing the OpenBet relationship. He’s a well-known guy in the industry too.

So it’s a really nice mix. We’ve got the marketing, personalisation and tech component and the integral industry stability components that are essential for a gaming tech provider. We also understand how to work with major back-end providers and ensure that Graphyte doesn’t have any negative impact on operators’ platform stability and speed.

We’ve heavily focused on making sure Graphyte operates at the lowest possible latency. You can’t introduce a piece of software that slows down the load times of a betting site, and that’s a massive part of Graphyte’s value proposition. We manage all of our processing end-to-end away from the operator’s platform. We would never slow down a site and no sports betting operator would integrate a piece of software that slows down the site. These are the sorts of things we know inherently from being in the industry and that’s by having our fingers burnt by testing out software that did that. Anything you can do to negatively impact the customer experience when you have customers with three to five different betting accounts is fatal in gambling, so we have to make sure the hygiene is on point to eliminate any risk of that.

EGR Technology: How does Graphyte solve the personalisation conundrum in the gambling sector?

DE: When we’re talking to operators, they all understand that creating unique experiences for their users is a no-brainer. By personalising experiences on site and via CRM, we’re making it easier for customers to find the products they want to bet on and play.

Going back to the earlier point about loyalty and multiple accounts, If a betting customer can find what they’re looking for straight away on an operators’ site, it’s going to have a big impact on minimising churn as well as offering that crucial ability to differentiate a digital offering far beyond traditional marketing or bonusing approaches. That was a big part of what we were trying to achieve with our platform.

Graphyte, fundamentally, is about building solutions that are specific to the betting industry. A great example of this is in build/request a bet type products, which are enormously popular and now available on virtually every tier one and two site. They’re a fantastic proposition all around, enabling users to construct their own bets on certain types of markets and compound the odds. They tend to be a higher margin product but at the same time highly engaging for users. The bets themselves, however, are represented in a very difficult format to analyse and therefore personalise or recommend to customers. We’ve built specific, easily consumable algorithms that enable operators to recommend and cross-sell customers into build a bet-type bets based on their normal regular bet. It’s a solution that is completely bespoke to sports betting.

Operators have told us they’ve tried pilots with industry players in ecommerce that haven’t been successful. The typical ecommerce type solutions don’t really understand the betting and gaming product or customer. From simple things like bet slip versus shopping basket, or the high frequency recurring deposit to play model, high volume acquisition and marketing models, all of which are very different to ecommerce, travel etc.

That’s the conundrum. Everyone in the industry is saying they want to personalise, but there is no solution to buy, hence why we built Graphyte.

EGR Technology: How is Graphyte helping the industry to address its tech challenges?

DE: Graphyte is a very easy platform to integrate with. Our data structures are built to replicate major sports and gaming back-end providers. We can work with data directly from a sports or gaming platform. If that’s a problem, such as with legacy systems that don’t have nice data structures that can be streamed out easily, we are able to work with data that comes from the browser and deliver the same level of accuracy and functionality.

Integration is always difficult in B2B software, and there’s no such thing as an effortless integration. That said, we made sure that one of the core principles of how we designed and developed Graphyte was to make it as easy and as flexible as possible to integrate.

Ex-Ladbrokes Coral CTO Rob Davis founded Graphyte with Evans

EGR Technology: What has been the reaction from the industry so far to your solution?

DE: On the whole, unbelievably positive. The typical feedback we’ve received after a demo is “everything you’ve covered is everything we want to do in our 2019 roadmap”. It’s really nice getting that feedback. Even though we have an industry background and believe we’ve something relevant, going out and getting that positive feedback and that traction from operators, especially for a new product, is a nice confirmation that we are hitting the right spot.

EGR Technology: Where next for the product and Graphyte?

DE: We want to get as much penetration in the market as possible with our core offering which is on-site and below the line personalisation.

We are looking at integrating more qualitative sports stats and gaming data into those recommendations, particularly on the sports side. And a lot more on the triggering side. We have some nice triggering tech where we can send out automated messages below the line based on a whole host of events such as bet placement, settlement, balance etc. We definitely want to build that out more.

It’s a real upside for marketing teams if they can take away the manual component of messaging customers, or at least a significant part of it. So many operators still have manual processes integrally within their CRM workflows, with those kinds of overheads, personalisation at any truly individually targeted level is probably still a pipedream.