DataVisor launches channel program around unsupervised machine learning fraud detection solution

DataVisor has a differentiated unsupervised machine learning solution, with the initial Go-to-Market being focused on systems integrators and strategic vendor partners.

David Cassady, DataVisor’s Vice President, Business Development, Partnerships, and Sales

DataVisor, which uses unsupervised machine learning in a differentiated fraud detection solution, has announced ExtenD, their first partner program. The startup is relatively new to the channel, and out of the gate is focusing on systems integrators and technology partners, with plans to expand to boutique partners going forward.

DataVisor was founded in 2013 by two female Microsoft veterans, Yinglian Xie, the CEO and Fang Yu, the CRO. Their head offices are in Mountain View CA, and most of their R&D is in the U.S., although some is in China.

“At the beginning of this year, we  began to develop our global channel strategy and are committed to that,” said David Cassady, DataVisor’s Vice President, Business Development, Partnerships, and Sales. “The big verticals for us are Internet social media, gaming, and finance. We “We think we have an opportunity to corner the market, but we can’t reach out to all that ourselves with just a direct sales force. So we want to augment it with a robust channel strategy. In 2020, more than half of our global revenue will be channel –  and that’s conservative.” The direct sales force now gets paid on a channel-neutral sales model, so they get paid just as if they sold it direct.

DataVisor has a differentiated technology where the Big Data and AI approach is done through unsupervised machine learning, which proactively identifies and prevents sophisticated online attacks, on any platform.

“We think we are a number of years ahead of our competitors,” said Tom Shell, Head of Alliances at DataVisor.

“The first and easiest technique for identifying fraud is creating a rule for a fraudulent transaction,” Shell stated. “It’s trickier where there are subtle changes to what a transaction looks like, which can trick the rules engine. Many organizations use supervised machine learning where a computer learns to identify variations of fraudulent transactions it sees, even though they change over time. However, with supervised machine learning, you have to know what fraud looks like and train it to recognize it.

“Unsupervised machine learning eliminates the need to train the recognition engine,” Shell added. “It lets us be more precise in how we find fraud because it can see fraud patterns without training and allow them to be detected quickly. This can be from seeing them three or four times – rather than taking months.

Shell said that while other players use unsupervised machine learning, DataVisor believes that from a commercial capability standpoint, they are unique in the fraud space.

“There are some other startups and a few established firms trying to use unsupervised learning, but they are very dependent on a generic open source approach,” he said. “We have customers, large financials, who were experimenting with unsupervised machine learning, but they found it was hard to build a solution that eliminates the noise. Instead they chose to buy it from us.”

The sweet spot for this is the enterprise, but it plays downmarket for some types of customers.

“It’s more about volume of transactions,” Shell said. “A credit union who would have a million transactions a day still makes sense. The bigger the data sets, the more value we have.

DataVisor will directly sell their DataVisor dVector managed services solution,  while their dCube and Feature Platform will be sold both direct and by channel partners.

DataVisor’s new ExtenD is designed to enable a select cadre of partners.

“Our strategy is to have the right partners, and that’s a short list,” Shell said. “That will include technology partners, and we already have Experian publicly reselling us.” The Experian partnership was formally announced in early November.

Systems integrator partners will be critical out of the gate.

“In terms of the global systems integrators, we need to have relationships with the big four and will be announcing one of those in the next few weeks,” Shell said. “The goal is to get two or three of those in the next year.”

Shell said these larger partners will be key to developing initial momentum.

“It’s all about creating scale and velocity,” Shell said. “I worked for small resellers in large tech ecosystems, and spent four years at a global SI. Our technology is well suited to be an enabler for something a large partner is doing, so it’s likely that channel will develop over time. We aren’t invent

ting the wheel. We want resellers who have strength and capabilities in different markets. We believe that  consulting and system integrator folks will look to build practices around us because we are complementary to something they are already doing very well. We will create and craft a complementary Go-to-Market with each of these partners.”

The plan is to add in traditional small boutique partners a little later.

“We have a philosophy of how to go to market with them, and we will sign up a few of those, but not this year,” Shell said. “We will work with a few SIs and technology partners first, and choose boutiques to augment them.”