There's this part in the seminal Alan Moore/Brian Bolland Batman story The Killing Joke where the Joker, his plan thwarted, shares a joke with the caped crusader: Two convicts are escaping from prison and make a rooftop escape. They reach a ledge between buildings.
One makes the jump easily, but the other is too scared to fall. It's OK, says the first convict, I'll turn on my flashlight and you can walk across the beam. The second convict goes "Are you crazy? You'll turn it off when I'm halfway across!" (This is better in its original form, I promise.)
The modern world of technology platforms finds itself in a very similar dilemma. Look at Slack, the increasingly popular chat collaboration tool that's become Silicon Valley's latest darling: Its secret sauce lies in the fact that you can pull in data and work in functionality from a growing number of external sources, including Twitter, Twilio, Github and (sigh) Yo.
When CITEworld first pointed this out earlier this week, an interesting Twitter conversation ensued: It was ZDNet's Ed Bott pointed out that this approach only works as long as the other services play nice -- a really easy way to strangle Slack in the womb would be for, say, Dropbox to discontinue support, turning off that figurative flashlight.
Without integration with a super popular platform like Dropbox, users would have to go outside Slack's carefully cultivated garden, with a possible net result of more users spending less time in Slack, which is the kiss of death for any social tool. This might be good news for Dropbox, especially if it's working on its own social collaboration platform.
But if you take the long view, Dropbox would be only hurting itself. If users really like using Slack, and Dropbox integrates with Slack, then users will want to use the two together. This seems obvious, but I'm a veteran of enough video game console marketing wars to know that sometimes, platform providers confuse the immediate drive for profit with the long-term needs of its users.
A good service that works when users want, how users want, in the ways users want will always prevail in the long run. This is the philosophy that's driven Box and Dropbox to their current levels of success, after all.
If you need another strong example, look at Salesforce. Not only does Salesforce have its popular AppExchange business app store to bolster its offerings, it also offers extensibility via APIs to a whole mess of partner applications and third-party tools. This is good for Salesforce, because it wants you to keep pumping more and more data into its platform.
But it's also bad for Salesforce, because Salesforce wants you to rely primarily on its first-party sales and marketing tools, the ones it has bolstered with millions in acquisitions and R&D. More choice for customers keeps those customers happy, especially when partners offer vertical-specific solutions. But sometimes, the suits that run big companies can look at the ecosystem and see money that they're not making.
This is why you see companies like Amazon Web Services eating their own ecosystem by offering services (like mobile backend) that compete directly with the partner developers they claim to love. If you liked competitor X's mobile backend, you'll love AWS Mobile Backend Services!
And it's not like there isn't a history of service providers closing down or severely limiting their APIs, hurting or killing nascent ecosystems, in the name of profit: Twitter famously shut down access to its API for developing third-party clients, killing the business of every single third party developer in the name of promoting a more advertising-filled experience.
Maintaining a platform for external developers is a very tricky line to toe. Look at how skittish Apple gets just adding features to every subsequent release of iOS. It's trying very hard to integrate the features and services that people really want without losing its iron grip over a smartphone platform that basically defined the market category. Meanwhile, Microsoft is building itself out as a developer platform once more, trying to convince the world that it's shaken its proprietary-first ways.
So basically, every single cloud service you love lives on top of an interconnected network of flashlight beams. All it takes, sometimes, is a single beam being turned off or even pointed a different way, and the entire structure shifts. But, at the end of the day, they're just flashlight beams. Tread with caution, or else not at all.
(Note: This doesn't describe the classic "prisoner's dilemma" problem in game theory. Please forgive the authorial license.)