Intellimize introduces a new way to test websites
The tool uses machine learning to give marketers an alternative to A/B testing.
What if I told you that instead of testing only two different versions of your website, you could test 94 million different versions? It may seem impossible, but that’s exactly what Intellimize can do.
Using machine learning to update its algorithm on the fly, Intellimize wants to shake up the way marketers test and optimize websites. The niche is currently dominated by one-on-one A/B testing.
CEO Guy Yalif told me that the company’s predictive personalization process allows companies to automatically discover segments of visitors and optimize their site visitor experience to drive more conversions.
“We’re making marketers heroes by driving great sales,” Yalif said.
The startup uses predictive personalization, a new form of machine learning website optimization, to enable marketers to test the performance of many different versions of their websites at the same time. In its first year of operation, the company found that their average customer tested 127 ideas, a total that would have taken 25 years using traditional testing methods.
In addition to the sheer number of permutations that can be tested, Yalif says predictive personalization yields conversions a lot sooner than traditional testing.
“For most websites, it takes a few weeks to get statistically significant results with a single test,” Yalif said to me. “And often (we’ve heard 80% of the time), the result is not actionable.”
Intellimize starts by brainstorming with their clients to determine what would make the biggest impact. Once the client delivers its website creative, coding is added. The system then begins automatically trafficking and optimizing the different versions of the client’s website. Finally, the company helps its clients to identify further personalization opportunities.
Current customers, which include Grammarly and Chime, have seen a 79 percent and higher lift and more conversions.
Opinions expressed in this article are those of the guest author and not necessarily MarTech. Staff authors are listed here.