App Disaggregation’s Physical Limit: 30 apps per device

eBay Marketplaces: 3 apps (Core, Fashion, Motors)
Facebook: 5 apps (Core, Messenger, Whatsapp, Instagram, Paper)
LinkedIn: 3 apps (Core, Contacts, Recruiter, and until a few days ago, CardMunch)
Google: 36 apps (ridiculous)

Across mobile, the companies most serious about mobile are disaggregating their features sets and building new, more focused, apps. There are compelling reasons to split up your apps:

  • Apps get one chance to get key permissions from the user — push and geo, in particular. With multiple apps, you can ask multiple times, and users might have a clearer understanding of why the permissions are necessary. And if one of a company’s apps annoys users, the users are less likely to totally cut out that company (by revoking permissions in all apps).
  • Smartphone screens are small — even on the largest Android devices. Focused apps can have simpler user interfaces, without hidden features. If a button isn’t visible, most users will never find it. This is why Facebook removed their slide out navigation panel and why Venmo’s right nav surprised me. RedLaser has a left navigation, but we find that users are orders of magnitude less likely to use any features there.
  • Updates take time on iOS. If you’re testing new features which might anger customers or be unstable, it’s better to test in less important apps and eventually roll the best and most stable features into the core experience.

However, most users don’t have many apps on their devices: Surveys suggest users have only gone from 31 to 33 apps on their devices over the past five years. (See my data here.) Moreover, most of these apps are likely games.

Average Apps per Device

If users don’t have more apps, who is using these more focused apps? A few hypotheses:

  • Power-users use anything other than the core apps (and maybe Whatsapp). App manufacturers can create power tools for some users without distracting most of their user base.
  • Apps from powerhouse developers are beating out smaller companies. eBay/Facebook/Google/etc. have a larger share of the install base than they did 4-5 years ago.
  • None of the ancillary apps have significant install bases. They’re purely used for experimentation, and all the successful features will be rolled into the core apps.
  • The cohort of users with smartphones in 2010 might have many more apps installed in 2014 — but the average is weighted down by new late-adopting smartphone users.

I suspect all these explanations hold a bit of truth. It’ll be an interesting place to watch — particularly because discoverability of installed apps becomes a huge issue once users have 100+ apps. I’m already at the point where any apps off my home screen are accessed via Apple’s built in search feature or Siri.


Top PMs are top trusted advisors

I’m a huge fan of Quora. I have no idea if it is sustainable as a business (and get a bit flabbergasted by its regular, and enormous, rounds of funding), but there are few other places where you can find reflections on the height of the author of Dinosaur Comics author Ryan North and the valuations of small, private, startups on the same website.

One of my favorite answers is from a general manager at Amazon, who ponders what distinguishes a top 1% product manager from a top 10% product manager. I’m particularly struck that the respondent (Ian) goes beyond the typical trite description of the role — “Product owner,” “mini-CEO” — and lays out very detailed and particular activities he thinks sets the greats apart:

  • Think big
  • Communicate
  • Simplify
  • Prioritize
  • Forecast and measure
  • Execute
  • Understand technical trade-offs
  • Understand good design
  • Write effective copy

(He goes into more detail in his answer.)

What struck me about this list is that the defining characteristics of a great product manager are also the defining characteristics of the greatest consultants I’ve known — and the greatest lawyers, and marketers. At its heart, answer suggests that a great product manager is a great trusted advisor.



Uber’s business model: What does last month’s SF price decrease mean?

In January,  Uber announced it was dramatically slashing rates for UberX cars. With the new rates, UberX cars are the lowest priced transportation option I can find in San Francisco.

Techcrunch and other venues posited that this was a purely competitive move: Consumers were trying out alternative services like Lyft, and Uber needed to be cheaper to compete. While I’m sure that’s part of the story, I think there’s something more going on.

The ride share service is brutally competitive — and not just on the consumer side. It puts enormous power into the hands of drivers, because people don’t just want low rates but also supply. One impact of this is that the economic surplus will probably accrue to drivers — services will compete to pay them a larger and larger percentage of the total charge just to get them driving for their service, rather than a competitor.

Moreover, since the people using services like Lyft and UberX are extremely price sensitive, the best way to get new customers is to lower rates. Overtime, you’ll probably see movement like this:

Profit curve for ride share businesses


Prior to January, it looked like profit for ride share businesses would slowly decline as prices dropped and drivers took a larger and larger percentage of revenue.

Uber’s January move changed that — they jumped from A straight to B on my chart, instead of slowly bringing down prices. Since they didn’t reduce the absolute dollar value of driver revenue, the drivers effectively take a higher percentage of revenue.

So what? Uber just kicked off a price war … because it is one they can win. 

See, the above curve is  only true for community drivers. Black car services are a totally different market — you differentiate on service and convenience, not  price. This is a market Uber owns.

Uber’s main concern is about winning marketshare & press in the broad taxi service (possibly because they see a future with driverless cars, where capital is more important than labor, and they’ll win back their profit margin), and about making money via Uber black car. 

Which gets to another reason for the price decrease: To better differentiate its services: Uber vs. UberX.

Uber Business Model

Prior to this price decrease, Uber had better service but otherwise wasn’t much different — a bit more expensive.

By lowering prices (and forcing its competitors to do the same), Uber actually might have reduced the number of cars on the road. Suddenly there is a huge competitive advantage for the black car service — they’re simply more available, all the time.


Why freemium’s only going to get more common

Localytics, via Techcrunch
Evernote blog, via Markitecture

These two charts, taken together, are what have me convinced that freemium is going to be the dominant business model for any serious application in the future.

What these charts tell me:

  1. Mobile applications either win or lose; users try lots of apps, decide which they like, and then use those few heavily
  2. Users are willing to pay for apps they use heavily

The key, then, is how to convert users into heavy users, and thus convince them to pay. Evernote, Dropbox, Pandora, and others find the freemium model well-suited to this–get lots of free-users, offer a compelling product, and convine the heaviest users to shoulder the majority of the cost.

There are other options. Microsoft and Adobe are able to jump immediately into a relationship via their reputation and business lock-in. But I suspect the days of such uniformity are drawing to an end. Google Docs, Evernote, and a whole suite of applications threaten the Office lock-in. Adobe and PDFs, another must-have software set, is also on the decline.

I can imagine the freemium model expanding to other industries. Amazon’s announcement of free book rentals is an interesting stab at a physical-world freemium program–users still need to buy (subsidized) Kindles, but it seems like their goal is to lower the trial costs as much as possible, and encourage the heaviest users to come in with gusto.

There’s real money to be made in figuring out how to apply this pay-gradation to other industries. I could imagine it would be effective in video, in particular.