Developer platform economics

· May 13, 2018

In this post, developer platforms means multi-sided markets where one of the sides consists of software developers. Platform businesses are immensely powerful drivers of the modern tech landscape— pretty much every notable tech company has at least some aspect of a platform business in their DNA.

The core job of a platform is to perform matching. A simple marketplace like eBay focuses on increasing the number of good matches (bids or purchases) between seller and buyers. Similarly, the Apple App Store wants to connect iOS users to app developers, and Google’s advertising platform attempts to match publishers to advertisers.

Platforms generate value through both the volume and quality of these matches — lots of matches that don’t result in positive interactions (e.g. ads people don’t click on, or products people don’t buy) aren’t valuable.

Capturing value

Platforms make their money by capturing some of the value from the interactions their matches generate. There are a handful of ways of doing this:

  • Take a slice of a transaction. When the match involves a purchase taking a percentage works well. The classic % charge from a payments platform like Visa is an example here.

  • Charge a fee for access. This works when there is a valuable network which isn’t available elsewhere.

  • Charge for enhanced access or service. For example, selling ads on keywords in the App Store allows Apple to capture value from developers who get the most out of being on your platform (and therefore are willing to pay for more downloads).

Platform strategy involves balancing how much value is captured, how the value is captured, and how it is split between the participants. Segmenting audiences allows differentiated value capture — you can incentive different types of customers by reducing the value capture, or even deliver extra value to a key segment in order increase the overall worth of your platform through increased participation.

This is because often one side is more valuable than another: for example, Google wants as many Android users as possible, so there is no charge to use the Play store, but with developers they want to set a minimum bar: so developers must pay a (small) validation fee, and have their apps reviewed.

Most notable platforms are more than a two-sided marketplace — they have several constituencies they are attempting to serve, which adds another level to each of the factors discussed above.

There is an interesting case study of these principles at work, which has taken a new turn recently: the opening up of Amazon’s Alexa to paid products.

Amazon has played the game very well with Alexa so far, so it is worth looking at how the platform has been managed.

A platform for voice

Amazon had a more expensive first step than many app based platforms, as they had to create Echo and Alexa. The hardware and software capabilities were good enough to make voice-based interaction possible, and practical, in a noisy home environment. The first device started shipping back in 2014, and it’s clear that Amazon knew it was a platform play — there was an SDK from the get-go (though it was invite-only at first).

A new platform always faces a chicken and egg problem: users want the device for capabilities, developers want to build for the device because of user reach. The first release of the Echo leveraged Amazon’s size: you could shop, listen to Amazon music, and get answers to questions thanks to Amazon themselves pulling data from the web. But it also involved third-party developers, notably audio services iHeartRadio, TuneIn, and Pandora. Finally, the Echo also functioned as a Bluetooth speaker, allowing you to play music from phones.

This was a beautiful bit of product positioning. Consumers were familiar with bluetooth speakers, and the price of the Echo was comparable (at $100) to many of the better units. Amazon marketed Echo very much as a speaker: it started out only available to Prime customers, who by default had access to Amazon Music. This meant that the initial experience could be “Alexa, play ” and the user would likely get a result.

This differentiated the speaker experience from conventional Bluetooth with its pairing problems. By focusing on music, Amazon both framed the product in familiar terms and started to differentiate it, while introducing the truly interesting capabilities of the device: the voice based interface.

The next step towards developing Alexa as a platform was threefold:

  • Offer an SDK for building Alexa integrations (skills).

  • Offer Alex Voice Service to allow hardware manufacturers to integrate Alexa into their own devices.

  • Create a $100m fund to support developers using 1 and 2.

Releasing the skills SDK was a natural next step once Amazon were confident in the product — through having partners Amazon had already had invest a lot of effort to externalize its SDK, and had already acknowledged that consumers had their own preferred music streaming services they wanted to use.

Voice Services was a more interesting move. Offering hardware manufacturers the ability to integrate Alexa runs the risk of a really superior product cannibalizing Echo sales. As it turns out, this didn’t happen: the most notable integrator, Sonos, has a much more high-end offering. What the move does do is show Amazon’s confidence in platforms: if someone ends up selling millions of Alexa-enabled device, Amazon is confident it will be good for them.

The final move, the fund, was both a direct investment to spur the creation of new integrations and a very strong signal that this market was significant to Amazon. They were putting their money where their mouth was and saying “bet on us, and we won’t abandon you”.

That is a huge value in a platform — if someone invests the time to build an integration with you, pulling the rug out from under them can be devastating to them, and to your reputation. As a very pleasant side benefit, the fund seems to have paid off at least in terms of outsourcing R&D — Ring were funded, and then later acquired, by Amazon.

The sum of the moves helped Alexa to become the default integration point for voice control of smart home devices, host to the broadest available set of music services, and be the entry point for numerous other integrations and tools.

The question then goes to value capture. As it stood, the value of Alexa is pretty strongly weighted towards Amazon and the consumer. While they take a hit on the device cost, the access to the consumer is significantly valuable to the company. Having your service present on the device is valuable to the likes of Spotify or WeMo as well, but it isn’t necessarily delivering that much to smaller developers.

The recent announcement about monetization is a nod towards unlocking new streams of value, both for developers and for Amazon. The first attempt at sharing value was Alexa Rewards, which allocated some cash to popular skills. The new option is In Skill Purchases, which follow the App Store/Play Store In-App Purchase model. This is targeted at digital goods, with the platform taking a 30% cut. The size of the cut means you’re limited to very low marginal-cost products, which tends to lean the model towards games and other digital goods. Amazon has also offered Amazon Pay for Alexa Skills which enables physical and other higher marginal-cost purchases without the huge cut, as long as you use Amazon’s payments system.

Whether the In-Skill Purchases model translates well into voice interactions is to be seen, but it does offer an option to developers who are looking to get some more value out of their investments. The Amazon Pay integration is even more interesting — Pay itself is a platform, with Amazon intermediating banks and merchants. Amazon is segmenting different types of app developer, without forcing them to go outside the platform to do their purchases, as happens on App Store, for example. This type of cross-platform leverage is a classic Amazon move: drive value to their platforms by increasing interactions while capturing more customer touch points.

There are a few takeaways from the Alexa example:

  1. Focus on increasing the number of valuable interactions on your platform, even if you’re not clear precisely how the value capture will work. Amazon did a great job of getting Alexa into homes and giving developers an easy route into building skills.

  2. Learn from other platforms and identify opportunities to enable forms of exchange that have previously been a challenge. Amazon’s support of a low-overhead payment system is great for retail and restaurant businesses and ensures Amazon gets a cut even when consumers go outside Amazon’s own services.

  3. Frame your offering in terms familiar to platform participants, and ensure they can start getting value right away. By focusing on the speaker aspect Amazon was able to deliver value to consumers and to a set of developers right away.