Does @ChatGPTapp have $300B worth of network effects? Not right now - the network effects are quite weak. But they could become a lot stronger if @OpenAI
turns @ChatGPTapp into a platform similar to iOS/App Store.
https://t.co/PA9cnWqXD9
@ErikHovenkamp@skominers Thank you @ErikHovenkamp for the nice summary and @skominers for the related papers. The Dixit paper is indeed closely related though we go quite a bit further because we have a platform setting prices to lay the trap (no price-setting in Dixit)
New paper “AI and Competition Policy” forthcoming in IJIO.
Will AI be dominated by a few firms? How will AI affect the market structure for existing sectors? What “traditional” competition policy issues arise with AI and what novel issues does AI raise?
https://t.co/9oa7sg1eNG
My new @HarvardBiz article with Julian Wright on how to evaluate defensibility and upside potential for marketplaces. Available online now, will be in the July-August magazine issue. Please share with your marketplace entrepreneur and investor friends!
https://t.co/lGBg8gVOY0
Miss our group chat with @theplatformguy earlier this week?
This was a great chat where we did a deep dive into topics like network effects, defensibility, the impact AI will have on marketplaces, & more.
The recording is now live!
https://t.co/iX0h2xRZ0D
@jsndr@falsedvmaster@thumbtack That seems like the most reasonable option indeed. The "no thanks" button... more generally, make the mechanism as simple and transparent as possible to both sides.
Used @Thumbtack to hire a plumber and realized there is a design/pricing issue that seems unfair to pros. Here is my understanding of how it works (please correct me if I'm wrong):
@falsedvmaster@thumbtack The question then is whether there is any serious downside in clearly communicating the constraint to clients? Sthg along the lines "if you want to negotiate, go ahead, but only do so if you are seriously considering the pros, otherwise you are costing them for no reason"
@falsedvmaster@thumbtack Right, so I guess my behavior isn't very common. Interesting to consider the "right" mechanism design here (assuming the platform is interested in that and not in a quick $25 on the sly). Agreed there should be some friction for pros to get refunds in order to keep them honest.
@falsedvmaster@thumbtack Still, why should the burden be placed on the pro to request a refund? Why not make it clear to the client that there is no need to reply "no thanks" as that triggers a fee to the pro?
Then a pro I didn't hire let me know that my reply cost him $25. @Thumbtack never told me that my replies would cost the pros, but did encourage me to reach out to multiple pros. That seems like blatantly misaligned incentives on their part and not fair to the pros.
Article with Julian Wright "Data-enabled learning, network effects and competitive advantage" has been nominated for the #AntitrustWritingAwards2024
If you think it worthy, please don't hesitate to vote for it (one click and no login required)
#AWA2024
https://t.co/H3wWAq1hbE
This is how @Shopify differentiates from @amazon - NOT implementing full marketplace discovery
Also: @owner (Shopify for restaurants, no discovery) vs. @DoorDash (full discovery), @Teachable (Shopify for online teachers, tried and abandoned discovery) vs. @Udemy (full discovery)
New paper in Management Science on how much discovery #platforms should enable. Making it easier for buyers to discover sellers leads to more transactions, but it also commoditizes sellers. So LESS discovery may sometimes be more attractive to sellers.
https://t.co/AlZeDy2R7C