When Dropbox got big, Apple and Google giggled and smacked their lips and said ‘fuck them we’re going to in-house this’. Drew Houston says Steve Jobs told him you’re a feature, not a product and we’re gonna have to come after you in an M&A meeting.

Google drive only launched in 2012. Dropbox had been around for years by then, although Google docs launched in 06 after they bought Writely. There was speculation even back in 06 about a ‘G:Drive’ product. (The colon didn’t make it).

Every platform that Dropbox targets has a Dropbox-clone storage product, but they survived. We’re living in the platform era, and the removal of the still beating heart of apps is the ritual by which platforms sustain their power. Understanding how Dropbox dodged the ax is relevant to anyone who wants to do business online.

Sherlocking vs cross-platform networks

Jobs wasn’t wrong to think that featurizing was an option. Developers of productivity apps for Mac have long feared getting ‘sherlocked’, which means you release a tool for Apple and then they turn it into a feature.

The original sherlocking story isn’t totally compelling – a company released a tool called Watson that improved on Apple’s built-in Sherlock 2, and then Apple released Sherlock 3 which also improved on Sherlock 2. But products who launch on platforms aren’t wrong to fear the platform.

Dropbox survived, despite competition from every platform they support, because their network transcended any single platform and truly enabled collaboration. If you clicked this headline, I’m not telling you anything you didn’t know when I say that networks are a useful moat. In this case, cross-platform networks in particular are tricky to sherlock and a good defense.

Apple + Google’s storage offerings are, on paper, cross-platform, but it’s hard to offer something that’s simultaneously a ‘platform feature’ and ‘works everywhere’. In part because things that are seamless on your own platform are harder on other platforms (login, for example). But also there’s cognitive dissonance at the strategy level; like a wild animal kept as a pet that periodically attacks its owner, you’re at constant risk of slipping up and shipping a lock-in feature.

This is a reality of launching anything in the platform era. Are they showing their teeth because they’re smiling? Or just hungry. (If you have 54 minutes, watch all of Jon Zittrain’s early + prescient 2008 sendup of the US v Microsoft litigation at the dawn of the platform era).

For examples of the ‘DNA dissonance’ at platform companies, look at Apple’s de-cross-platformization of TestFlight and Dark Skies. Because of Apple’s desire and ability to not have a TestFlight competitor on its platform, no good cross-platform competitor has popped up.

Apple definitely knows that there’s an undercurrent of hatred for its tools by its devs, and any method for distributing apps that they don’t control will quickly attain dominance just by being marginally less shitty. So sheltering TestFlight may be strategic. But Dark Skies is a weather app. There’s no mind at work here. This is Apple succumbing to the beast within™️.

Feature cannibalism as a model

Dropbox wasn’t a cannibal startup. In an early pitch deck, they say:

Online storage is unclaimed, unmonetized territory, much like search pre-Google

So it wasn’t like they entered life with a strategy to repel featurization by platforms, other maybe than ‘suck less than they do’. But the forces that allowed them to survive would be true for companies founded intentionally to steal share from the platforms. (Other than momentum, of course, which dropbox had and startups don’t).

But now in 2020 we’re at peak platform in a bunch of ways:

  • Users are trying to get off platforms, or at least new york times columnists are (note that the nokia feature phone in this article runs YunOS, an Alibaba fork of Android, and shouldn’t represent ‘escaping the tech giants’)

  • Developers are starting to push back against tax-like fees (hey.com / epic games)

  • Government pushback, mostly in Europe but also in California. Rumor has it Facebook stopped fighting the CCPA because of news about Cambridge Analytica and some other political data sharing stuff. I’ve mostly ignored Facebook in this article because while you can have a competitive strategy against G + A, nobody should ever launch anything on FB and we should salt the earth when they’re gone.

  • It seems like users are willing to install multiple apps for X for most values of X. Chat is a primary example, which includes being in many different slack orgs at once.

Rather than just using cross-platform networks to resist getting featurized, we may see a swarm of new ‘de-featurization’ and ‘cross-platformization’ startups who are parasitizing the frayed edges of the platform companies.

What meta-Dropboxes are out there – things that platforms have already featurized, thereby establishing a value prop, that are bigger than a feature, and maybe have a network attached to them that the platforms can’t grow because they would accidentally crushing it.

‘Bigger than a feature’ is a point here too. Platform features are inherently low-attention; they can’t be hard to use, they’re build to be ignored, and they have good reasons not to become complex. When on the other hand you have a thing someone installed on purpose, you have more leeway to innovate. A cloud rando on twitter twists the Jobs line above and says ‘you need to be building a product, not a feature’.

Role of antitrust

All of the above comes with a hitch:

Platforms can use their control to break or ban apps that compete. Apple constantly says stuff like ‘we already have five of these, we’re not approving any more’. They have saturated categories where new apps have to provide a ‘unique, high quality experience’.

The end of the platform era depends on a new era of antitrust to prevent these new products from being smothered.

Courts are already learning from the past few decades of platform behavior. You’d be surprised, or maybe you wouldn’t, that supreme court decisions contain pretty deep discussions of two-sided marketplaces and relative contribution of value between platform and software publishers.

A world where there’s a cause of action for ‘featurizing this thing I invented’, or which lets a company competing with their platform subpoena emails for use of unfair information, looks very different from this one. Every time a court reveals internal emails at platform companies, it’s bad PR. Nobody’s hands are clean.

Should platforms be allowed to sherlock somebody? Currently, under the law, you can’t patent most businesses, and so the answer is yes. Imagine if it weren’t.

Did Dropbox win?

Last point – did Dropbox’s competitive strategy succeed? I found stats on wikipedia and this random site and have no idea if they’re right, or current, but using numbers between 2018 and 2020:

  Dropbox Google drive
Total users 600 million 1 billion
Paying users 10 million 1 million?
Teams 400k 1 million?

It seems like they’re within 2x. There are two big questions I don’t have enough information to answer:

  1. Is Google drive operating at a loss?
  2. In a new antitrust regime when Google isn’t allowed to bundle Drive, and consumers are forced to research ‘best storage’ and install something, which way do these numbers bend?