Algorithmic content and commerce

Podcasts · Startups, Technology

Created by Rishabh Srivastava, Founder of Loki.ai

Context

Source URL:

‎Pivot on Apple Podcasts

Why is it important: Most of the podcast is shit. But some parts were useful. AI-driven content and commerce is going to take over the world. First gradually, then suddenly. I want to both gain from this, and ensure that this is done in a way that adds value to the world

Keywords

Recommendations, Recommender Systems, Content, Commerce

Summary

...

Highlights

From 47:45

A. Why is TikTok so much more powerful compared to everything else?

Signal Liquidity: Netflix gets two points of signal about you when you watch a one hour show – a) you picked the program, and b) you watched it all the way through. That informs the algorithm and the recommendation engine

Bytedance gets between 800-1600 signals within the same period (watch length, likes/follows, patterns of interaction etc) because of how short each video it. The inputs that feed the algorithm give it calibration so that it can move to recommending what you want MUCH faster that any other media company. Mobile is key to this (alternative metrics: scroll-stopper index, scrolling vs clicking)

B. What differentiates media giants from also rans?

The media companies that have really gone to $100B+ valuations (YouTube/FB/Netflix/TikTok) leverage 2 things – 0 marginal cost of content creation (free content from the consumer base), and an incredible recommendation algorithm sitting on top of it

C. Algorithmic Commerce + Content?

If you can gather signal liquidity about specific consumer categories in e-commerce, you can begin to really leverage results and supercharge commerce. You can begin to get more and more signals, and head to a zero-click ordering world ⇒ Algorithm Commerce (AComm)

Algorithmic commerce can really supercharge commerce. If you marry it with algorithmically created content, the impact will be far greater than “influencer marketing” and “social commerce”