We spoke to Omri Mendellevich, CTO, Dynamic Yield, who offered us insight into how they are currently using AI and how the new technology will enable them to revolutionise the marketing scene going forward.
Founded in 2011 by Liad Agmon and Omri Mendellevich, Dynamic Yield is heavily involved in the AI-space. According to their website, their “advanced machine learning engine builds actionable customer segments in real time, enabling marketers to increase revenue via personalization, recommendations, automatic optimization & 1:1 messaging.”
During our conversation with Dynamic Yield's CTO and co-founder, Omri Mendellevich, he expanded on how the New York-based company - who also have offices in Philadelphia, Berlin, Tel Aviv and Singapore - uses this exciting new technology and how they see it changing businesses in the future.
AI Changing the Marketing Game
We started off our conversation with Mr. Mendellevich by asking him to give us an overview of how Dynamic Yield is currently using AI. “Artificial intelligence is at the very core of our strategy, currently manifested in our machine learning capabilities,” he began. “Our abilities in AI are vital to our consumer and investor interest and have garnered serious attention from prominent media outlets.”
Mendellevich went on to highlight the three main applications involved in their machine learning platform: “Automated Optimization/Decision Making”, “Recommendations” and “Audience Segmentation”.
We then started to discuss which industries they were gaining the most traction from in relation to AI. “First and foremost, marketers across industries are increasingly looking for single sources of truth to house all of their user data and apply AI to activate that data,” he stated.
“Serving the most relevant and highest-performing recommendations and dynamically allocating traffic to the best variation of your website for each audience can deliver huge uplift in e-commerce, travel, media, gaming and direct to consumer brands,” he explained.
Mendellevich summed it all up in Layman's terms for us. “In a nutshell, we have customers across verticals and all are very interested in our AI capabilities,” he recapitulated.
Contextual Bandits in AI
With so many companies looking towards adopting AI, we were keen to get an insider’s opinion on which areas of business is AI having the biggest effect. “Since the turn of the century, marketers have focused heavily on running manual A/B tests, splitting traffic equally between two variations, determining which performs best,” started Mendellevich.
“Recently, new approaches such as multi-armed contextual bandits have caught on as they understand the notion that each person may react differently to different content,” he commented.
Mendellevich went on to explain how contextual bandits can change the website optimisation game. Let’s start with some ‘context’ around ‘contextual bandits’. This is in reference to a new approach certain researchers, including Dynamic Yield, are taking towards their testing methodologies. Instead of adopting the traditional A/B testing method, in which there’s a clear winner and loser, Dynamic Yield prefer to tackle multiple problems at once, which will produce faster results since there’s no need to wait for an outright ‘winner’.
However, Mendellevich explained it much more concisely than I ever could. “With minimal machine learning, marketers can dynamically allocate traffic to the highest performing variations. But it requires AI to actually understand which variation will be best for me as 25 year old male from New York that loves French bulldogs and Italian food,” he informed us.
“This has long been promised in the marketplace but AI will help optimization companies truly deliver,” pointed out Mendellevich.
The Difficulties to Adopting AI
Naturally, our conversation moved on to Dynamic Yield’s competition. We were interested to find out what sets them apart from the other people in the market using AI. "For Dynamic Yield, it all starts with data," he began.
“We provide a data activation platform where marketers can onboard data from across web, mobile apps, email, CRM, 3rd party data sources, beacons and customer support touch-points to form one cohesive dataset,” Mendellevich offered.
“Basically, we provide our machine learners the best textbooks,” he summarised. “As a result, our machine learning and artificial intelligence applications are juiced with the the most robust dataset in the industry,” he continued.
“Ultimately, your AI is only as strong as the information it can ingest and our AI can take in data from all elements of onsite personalization and optimization to serve the best experience to each user,” finished Mendellevich.
With the buzz surrounding this new technology, it’s no wonder that so many companies are looking to adopt AI. However, this is much easier said than done, as Mendellevich explained, “The biggest challenge for businesses looking to adopt AI is understanding exactly what they would like their AI to accomplish.”
He continued, “Much like personalization, AI has become such a hot item that every company wants to have AI but hasn’t clearly defined what they would like AI to accomplish.”
Another challenge to adopting AI is, according to Dynamic Yield’s Vice President of Marketing, is human nature and our overconfidence. “By nature, humans are often incredibly overconfident in their own abilities and believe themselves to be paragons of rationality when there are many times machines can make better decisions by overriding human bias. Truly trusting AI is a challenge that won’t go away soon,” he mentioned.
Five Years’ Time...
We concluded our discussion with Mendellevich by asking him to try predict where Dynamic Yield will be in five years’ time. “Five years ago, we were two guys trying to build our minimum viable product,” began Mendellevich. “Now, we are a 100+ person company with scores of customers. We kind of think it is a fool’s errand to predict that far out, particularly in regards to something evolving as rapidly as AI.”
“Our mission is to be the market leader in AI based personalization. For now, we’ll keep working steadfastly towards that end,” he finished. Judging by the way things are going for them at the moment, we’re confident that Dynamic Yield will achieve their goal. However, we don’t wish to jinx anything by saying so.