The iGaming sector inherently operates in a high-risk environment, making fraud protection essential, not only to safeguard the players but also to protect the operators themselves.
Striking the balance between fraud prevention – and maintaining a seamless player experience presents a significant challenge. It requires innovative technology, industry expertise, and a deep understanding of consumer experience technology.
Enter Sift, an AI-powered fraud decisioning platform specifically designed for high-risk online industries, with iGaming emerging as one of its largest verticals. The firm leverages a global data consortium based on geography and industry to evaluate more than one trillion transactions annually, delivering real-time risk assessments with precision.
“In the US alone, we have about a 90% market share of all the major gambling platforms, said Armen Najarian, Chief Marketing Officer at Sift. “That gives us relevance and a volume to provide real time risk scores with which to make accurate fraud decisions.
200 milliseconds to decisioning with Sift
In the iGaming context, Sift activates at the moment a player attempts to register on an operator’s platform. As soon as the player submits their information to create an account or make a transaction, the system initiates a real-time call to Sift.
In around 200 milliseconds, Sift returns a risk score indicating the relative threat level associated with both the transaction and the user behind it.
“We’re benefiting from that rich context within the industry of online gambling to make even more accurate risk decisions,” Najarian added.”
Founded over 14 years ago in San Francisco, California, Sift embodies the innovative technological spirit of Silicon Valley, claiming to be ‘part of the AI revolution from the very beginning’. The company emphasises its science-based approach, backed by 43 patents for technological innovations registered with the US Patent and Trademark Office.
These patented scientific methodologies enable the firm to carry out instantaneous risk assessments, helping operator clients avoid potentially criminal or high-risk financial transactions.
“The phrase I often say is ‘I don’t need to know your name to know who you are’,” Najarian explained. “We’re not brokering PII (personally identifiable information), and we are not a credit bureau. To make that point clear, what we do have is a longitudinal view of transactional history associated with certain people who are transacting on the digital web.
“Without knowing your name, what we are looking at are the devices associated with the transactions that you’re making, the geolocation composure and footprint and behavioural attributes as well. What has shifted over time?”
Reducing false positives with three layers of data modelling
While comprehensive risk assessments help operators maintain compliance and financial security, these processes come with inherent challenges. Primarily, even sophisticated risk assessments are not foolproof – legitimate players can sometimes be incorrectly flagged and blocked, resulting in lost revenue for the operator.
“You don’t want to leave money on the table. You don’t want to offend an otherwise great customer or member of your community,” Najarian explained. “That really gets down to the science and the models that we build.”
He went on to add that Sift is effectively building a model for every digital platform that they work with.
“Those models get refined continuously. We provide our customers with a model based on their own with everything we know about your player behaviour, and we augment that custom model with the industry consortium model for the broader gaming industry,” he said.
Sift leverages its enormous database of over one trillion annual events to creates powerful models which minimise errors. Its ability to filter this data by jurisdiction, locale and industry further distinguishes its offering from competitors.
“The Global Data Network learns from every transaction and event, so the feedback becomes important. So in the example where there is a false positive, that data makes it back to Sift and it makes our model smarter. We don’t make the same mistake twice,” asserted Najarian.
Sift refers to this utilisation of the three categories of data as “ensemble model tuning.” This process helps reduce user abandonment on operators’ platforms by creating a less intrusive experience with fewer false positives.
Finding the Optimal Block Rate
To further reduce both customer abandonment and false positives, Sift has introduced RiskWatch – an automation solution designed to identify the sweet spot for block rates based on its predictive data science models.
“Like any digital platform, it is trying to arrive at that magical point, where we know statistically that, for example, a half of a percentage point of all transactions will very likely be fraudulent or questionable,” Najarian said.
“Getting to what that calibration point is is the challenge, and RiskWatch does this for our customers automatically,
“It takes all the guesswork out of what in-house fraud teams have been doing for years.”
The results have been remarkable – one client, in another high-risk payment vertical, achieved an additional $19 million in bottom line revenue after implementing RiskWatch.
Democratising fraud prevention insights
Sift recently launched another solution aimed at enhancing its reputation for transparency and reliability within the industry. FIBR (Fraud Industry Benchmarking Resource) was introduced approximately 18 months ago as a public-facing data tool offering insights into fraud prevention trends globally across several industries.
Developed in response to significant client demand, FIBR allows users to compare their performance against broader market trends using Sift’s trusted data resources.
“In FIBR, you can select from different categories. In this case, I’ve chosen online gaming or online gambling. You can also select a geography. You either have to choose an industry or geography,” said Najarian. “So we’re looking at online gambling globally, and these are the different metrics that we expose, payment fraud data, charge back data, account takeover data.
“The number one question we get asked from every single Sift customer in the gaming space and beyond, when we’re doing a quarterly business review with them, is how that data compares to my peers in the industry, or other Sift customers. We heard that so many times, and reflected upon that.”
Sift takes pride in its reputation for trustworthiness and continually works to strengthen this perception. That’s why FIBR is accessible to both customers – who receive enhanced insights – and the general public.
“FIBR as a resource is not just for our customers in the gambling space but the entire industry, even if you’re not a Sift customer. This is a one of a kind, valuable resource that’s based on actual real-time data across the trillion events that we see per year,” Najarian concluded.
“We look at FIBR as a way to add value to the fraud fighter community and to help everyone answer the most important questions about their fraud programs.”
