We had some coffee and talked about the future of machine learning.
Etgar Shpivak is the Co-founder and CEO of Fixel. He has a 15-year track record of hands-on digital marketing expertise, working with a myriad of companies – from small start-ups to large enterprises in Israel and globally. He is also the co-author of the best seller “Marketing in a Digital Age”.
What is the future of machine learning?
There are two major changes on the horizon for Machine learning. One is accessibility and the other is scale. Right now ML is a cutting-edge technical capability deployed and employed by scientist level experts, comparable in imagery to the room-sized vacuum tube computers of the 50’s run by doctoral students. The future will see ML deployed by far less sophisticated users at scale, monitoring, optimizing and engaging across interactions and systems. On a more immediate timescale, in the realms of e-commerce, we will see increased personalization as machine learning and AI play matchmaker, bringing customer and seller together in tailored, user-centric experiences, driving satisfaction up for users and expenses down for sellers as they compete for a more specific and better-targeted audience.
What are the criteria for selecting the most important customers?
Ultimately, lifetime value beats all, but this is an indicator that takes time to develop and even more time to understand. On a shorter scale, the engagement creed holds true. “An an engaged user is more likely to convert”.
How can AI affect the e-commerce industry?
This follows up wonderfully with the last question. Because though the creed holds true, deciphering its meaning from an obscure statement to a list of action items is not as simple as it seems. What is “engagement”? How does one measure it? How do I make sure the method I selected holds true over time? These are exactly the topics data scientists tackle by finding ways to analyze complex data sets to find actionable information at the service of e-commerce giants and agencies. AI and machine learning will democratize this, making the power of data science available at a price point accessible to a Shopify drop shipper.
In which direction will AI go?
Not going quite as far as a Kurzweilian ‘Singularity’ but still daring to venture beyond the limits of immediately visible technology, I believe AI will become increasingly contextual and API-less. I could see AI mediating between very different existing applications, old and new, creating a layer that generates API’s ad-hoc when needed, ushering in an age of unprecedented interconnectivity between anything and everything.
How to make the most of your website?
There are some great ways to demystify the concept of engagement, which if you boil it down, falls into three steps the user goes through. Awareness, Intent, and Action. The last part, “Action” – like an add to cart button click, is fairly easy to act on. But unlike it, “Awareness” and “Intent” remain slippery. Also, while only 10% of visitors perform a ‘Hard’ action like clicking add to cart or beginning a checkout process, the other 90% stay hidden under the umbrella term Engagement. Engagement becomes a single, simple, and super actionable metric from which you can easily generate audiences, allowing marketers to incorporate engagement into their existing marketing and remarketing efforts.