Suggestions
Michael Mignano
Dynamic capital markets professional
Michael Mignano is a Partner at Lightspeed Venture Partners, a prominent venture capital firm that invests in technology and innovation. He has a strong background in product management and development, having previously served as Vice President of Product Management for Aviary, a company that was acquired by Adobe. Additionally, he was the Director of Product Development for Atlantic Records, showcasing his experience in both technology and media sectors.1
Mignano is recognized for his expertise in advising companies on product strategy and development, making him a valuable asset in the venture capital space. His work at Lightspeed involves identifying and nurturing startups with high growth potential, particularly in the tech industry.1
Highlights
Meet the Supertake Bot for X. Tag @supertake on any post containing alpha, data/charts, or even a strong opinion, and you get a take back as a post.
If you're already a Supertake user, it'll put the take in your drafts. Otherwise it'll reply with a public take anyone can fork. https://t.co/87Enmxzpio

Traditional product metrics are dead. How do we measure the performance of products in the age of agents? As users shift their behaviors from UI to agentic interactions (such as via text, voice, etc.), we'll need to rethink how we analyze engagement.
For example, standard retention measurements that track whether a user comes back each week are meaningless when the agent is working on their behalf whether they return or not.
Here are a few specific metrics that will need to be rethought from scratch.
Active users: What is an active user in the age of agents? Is it someone who proactively engages with an agent? Is it someone who has a task completed for them in a given day? Is it someone who reactively responds to an agent that reaches out to them?
Frequency of engagement: Traditional consumer products look at metrics like DAU/MAU to measure how often somebody interacts with a product. What should it be for agents? You could argue that with some of the best agents, the user shouldn't have to interact with it at all while the agent is off doing work on the user's behalf.
Time spent: Historically, the more time spent with a product, the more valuable the product is for that user. With agents, we may decide that it's actually the opposite. The best experience may be spending as little time as possible instructing an agent, then getting out of the way.
Revenue retention: Normally you want to see revenue expanding over time. Will powerful agents actually result in shrinking revenue as inference gets cheaper and they need less of it to get the job done? Or will users simply hand off more and more tasks, so revenue grows even as each task gets cheaper?
There are many more implications to consider. What did I get wrong? And what metrics am I missing?
