Evidence receipt / belief
Published · transcript-backedDavid Rosenthal: belief
28 Oct 2024 Acquired Meta
“If yes, then it’s process power. In this case I think that’s potentially true. Yeah, you can bring in new engineers, you can bring in new product people, and transfer the advantage to them.”
Source trail
Everything needed to verify it.
- Speaker
- David Rosenthal
- Attribution
- Verified speaker
- Claim type
- belief
- Recorded
- 28 Oct 2024
- Publisher
- Acquired
- Episode
- Meta
Transcript context
…Yup. Now fortunately, in the land of creators, you don’t actually need to switch. It’s just an and. It’s not like you’re ripping out one vendor and putting another vendor in. I do have a sunk cost in building a following on a given platform, but that doesn’t actually prevent me from also launching on another platform if you have the time to do that. I think there might actually be process power. Normally there’s not, but I keep going back to this, like there’s something in the water. Their products grow in engagement over time. Their growth team does 10,000 little things to open up every step of the funnel as wide as it can be, and make the most frictionless fluid experience for users. Interesting. I was actually having this discussion with Hamilton recently about the difference between process power and cornered resource. A good test for this is when you bring somebody brand new into the organization and plug them in, does the power transfer to them as well? If no, then it’s a cornered resource. If yes, then it’s process power. In this case I think that’s potentially true. Yeah, you can bring in new engineers, you can bring in new product people, and transfer the advantage to them. For a while they had legitimate process power in how they shipped. Statsig has started on the premise of this. It blew my mind when someone from Facebook came and gave a talk at Microsoft in 2014, on how they ship product and how it rolls out. Not only the feature flagging elements of it, but they can watch performance auto roll things back if they’re causing negative performance metrics. How there’s just deployed code all the time with a whole bunch of experiments turned off. How there are these experiments that are running in different ways in different markets, and they can statistically significantly disentangle which results are from which experiment when they have multiple experiments that are concurrently running with the same user base. That was real voodoo that only Facebook did for feels like a decade.…
Stored transcript either side of the excerpt. The highlighted words are the published quote; the surrounding text is unedited source, never generated.