4AAVC101 - week 5

« Back to 4AAVC101

These are my notes from October 24 for 4AAVC101 at King's College London for the 2017-2018 school year. The lecturer, Nick Srnicek, is the author of two excellent books at the intersection of technology and leftist politics: Inventing the Future (with Alex Williams), and Platform Capitalism.

The usual disclaimer: all notes are my personal impressions and do not necessarily reflect the view of the lecturer.

« Previous week

Next week »

The Rise of Platforms

Readings

Matchmakers by David S. Evans and Richard Schmalensee (chapter 1)

No notes for this yet.

Bridging differing perspectives on technological platforms by Annabelle Gawer

The economics perspective (two-sided markets, which lets us understand competition) and the engineering perspective (the technical architecture, which lets us understand innovation).

Lecture

last week, we looked at peer production & its limits
- the model has historically had trouble expanding past a certain point
- simultaneously, we’ve seen the rise of large corporations
we can see platforms as the new business model
- contrast with the Fordist model of mass production and consumption, vertical integration; corporations expanded by buying up suppliers/dealers (1920s-1970s)
- also contrast with the post-Fordist model, which arise in the 1980s: individualised manufacturing consumption (think designing your own Nike shoe)
  - tied to the lean production model, where the corporation tries to own only the high-profit-margin aspects of the business
  - so Nike will own the brand but outsource the less profitable elements (like manufacturing) to corporations based in less economically privileged areas (i.e., they have no choice but to accept lower profit margins)
platforms can be intermediaries or infrastructure (or both)
- they can be intermediaries for passengers/drivers, users/advertisers etc
- they can also be the infrastructure, giving you tools to build apps/pages/profiles/social connections/etc
- for the purposes of this analysis, as in Srnicek’s book, we don’t treat Apple as a platform company (since most of their revenue is from selling hardware)
  - (though I’d personally argue that the hardware only has value because of the platform)
- platforms can also be physical, e.g. shopping malls (which bring together stores and potential customers)
three important phenomena that arise in the study of platforms
- network effects: basically the more people on the platform, the better it becomes (incidentally, I thought this was a fairly common term, but as it turns out I was one of the only people who had heard of it before … I guess I’ve just been living in startupland for too long)
- cross-subsidisation within large platforms, where some arms of the business are offered below cost (subsidised by others)
  - freemium model: those who pay subsidise those who don’t (enticed by better features, usually)
  - direct model: something is available for free but it’s subsidised by something you do pay for (e.g., Amazon’s free shipping)
  - ad-based model: your usage is subsidised by a third-party advertiser (though the subsidisation is only on the individual level, not on the societal level—unless the advertiser is doing something horribly wrong, they will be making enough money off of product sales to offset the advertising costs, and thus society is really subsidising the existence of the advertiser)
- designed core architecture: platform intermediation is political, not neutral
- some behaviours are encouraged or discouraged simply by the design of the interface (whether deliberately or not)
- Uber examples: surge pricing is due to predicted demand at any given point; phantom cars designed to make you think you’ll get a ride faster than you actually will
- or Facebook A/B testing your newsfeed to get you to engage more
Srnicek’s hypothesis: platforms are the most appropriate business model for the digital age because they allow you to capture & control data
- 4 main types of platforms, though of course any corporation can own multiple types among its different arms, and any individual arm can also be more than one (from his book)
  - advertising, where most users don’t pay for the product, and instead “pay” via their data & attention (Facebook)
  - cloud, which provides infrastructure services that you pay for (Amazon Web Services)
  - product, which provides a consumer-facing product that you pay for (Spotify/Netflix)
  - lean, which tries to not own too many assets and instead claims to function as a marketplace (Airbnb/Uber)
on advertising platforms & privacy concerns:
- there are always going to be such concerns just because of the economic structural forces (the need to make profit)
- the economic motives will always butt up against users’ expectations of privacy
- hence the constant privacy battles with Google/Facebook (it’s not just bad actors—it’s the whole business model)
- Apple is an interesting somewhat-exception to this—since their business model is mostly hardware, and not advertising-based, they can (and do) make privacy a selling point
on cloud platforms: Amazon Web Services (which, incredibly, most people in the class hadn’t heard of) runs most of the Internet these days, including many government services
- profit comes from renting out the platform (including developer tools)
- AWS is incredibly profitable and in fact kinda cross-subsidises the e-commerce business (though Amazon does sometimes make a loss)
- an interesting subcategory of this: industrial cloud platforms (Internet of Things, factories); GE and Siemens are both working on versions of this
on product platforms: goods are transformed into services available for rent
- examples: Spotify, Zipcar (though of course the economics are different since Zipcar’s goods are hardware, not software, and thus margins might be lower, depending I guess on Spotify’s negotiations with record labels)
- a really unexpected example: Rolls-Royce, which rents “thrust”—they make jet engines which are provided to airlines, but their business model is a subscription service for these engines; Rolls-Royce owns the data generated, and offers the airline maintenance in exchange
on lean platforms:
- also called the “sharing economy”, which is really a misleading moniker as it’s about making money, not about sharing; any attempt at disguising this fact is basically propaganda to hide the exploitation of workers
- or the “gig economy” which highlights the transition of the economy from careers -> jobs -> tasks (labour becomes even more modularised and flexible, which is of course the ultimate desire of capital)
- currently, the online gig economy is only a small proportion of the workforce: estimated to be 1% in the US and 3% in the UK (mostly Uber); for perspective, the agricultural sector is 2% (not sure which country)
- Uber, the largest taxi company, owns no vehicles; Airbnb, the largest accomodation provider, owns no property (yet)
- these companies are thus virtually assetless—they don’t even own servers most of the time (AWS, often); their only assets are intellectual property (software, brand)
open question: are advertising platforms sustainable?
- recent controversy over Facebook ads (though I recently came across a fairly convincing take that it has less to do with Facebook and more to do with the undermining of the idea that capitalism & democracy are compatible)
- some research has surfaced showing that highly personalised and targeted ads are not that effective
- my own anecdotal data from using Facebook: people are posting less and less, and so my news feed is increasingly being composed of indirect activity (“X is attending an event”, “X liked this post”), which suggests content on FB has peaked and is in decline
- there’s also limited room for growth in this industry, since growth in advertising can’t really exceed growth in the economy indefinitely (and economic growth has been sluggish lately, partly because of the market-shrinking powers of the tech industry …) & the market is being close to being maxed out already (how many more ads can we bear to watch before we just unplug forever)
- Google’s chief economist has said that ads may not be sustainable and they may have to switch to a subscription-based model eventually
- Facebook has said that it’ll charge publishers fees to have content show up in the news feed as sponsored posts (which is frightening because it raises the question of who they think will pay in the long run? aren’t they just shifting the costs to advertisers on a different platform?)
another open Q: are lean platforms sustainable?
- many are sustained by VC funding right now—not profitable
- also, and this is a fairly important point that I will be writing a longer post about eventually, the fact that so many people are willing to work for them despite shit wages and awful conditions is a historically limited phenomenon that, you could argue, is the result of several decades of neoliberal policy in much of the developed world
  - change the political landscape by implementing pro-worker policies that upend neoliberal doctrine & Uber et al could find it much, much harder to recruit workers
  - they’re already facing severe resistance and backlash (attempts to unionise, etc which I am all about)
- so many lean platforms have gone bankrupt already (esp in the food & cleaning spaces)
- Srnicek’s prediction: the sharing economy will disappear in a few years
- on Uber’s investment in self-driving cars: they’re shifting from a lean platform model toward a product model because the lean platform model isn’t sustainable
  - he predicts that Uber will either nail the self-driving car aspect, or it will go bankrupt
- I asked about Airbnb’s recent announcement re: branded apartments (i.e., “hotels”) which seems to indicate that Airbnb is also discovering the limits of the lean platform model, and is trying to become more of a product platform
on Google’s original dilemma: they could have either gone with the ads-based model, or they could have become a subscription-funded product platform
- they chose the former because it results in quicker growth, but now that they’re hegemonic, they could switch to a fee-based model
- similarly, if Facebook switched to a subscription-based model, they would feel less economic pressure to sell/use their users’ data (though ofc that doesn’t rule out greed—they could still try to sell/use user data just to increase revenue)
- incidentally, there’s a great Twitter exchange between sociologist Zeynep Tufekci and the head of Facebook’s News Feed that is quite illuminating