App.net's freemium gamble

Twitter's transformation into a media and advertising company was a "tragic mistake", says Silicon Valley entrepreneur Dalton Caldwell. His start-up App.net aims to become the advert-free platform for real-time social networks that Twitter should have been -- and today it launched a freemium business model, signalling the start of a make-or-break expansion phase.

When Twitter first appeared, Caldwell was "blown away" by its application programming interface (API). "This is the best plumbing I've ever seen. You can build anything with it," he told the Patch Monday podcast.

But as Twitter grew, two factions emerged within the company. One wanted to build a business around the real-time platform, the other around advertising. After a hugely divisive debate, the advertising faction won. Twitter became a fundamentally different company, where tweets are "content" and access must be controlled.

Twitter's new API, which comes into force this Friday, is one of many advertising-friendly changes. LinkedIn and others failing to follow Twitter's display guidelines have been blocked. Apps can bring external content into users' Twitter streams, but not export it elsewhere. Apps must provide functionality different from Twitter's own offerings while simultaneously looking and behaving exactly the same. Eat your heart out, Kafka.

Caldwell contends that Twitter was a communications company, and historically we pay for the infrastructure such companies provide. We don't expect telcos to run adverts in our phone calls, and we certainly don't expect them to data-mine our usage patterns for profit. Yet that's precisely what Twitter, Facebook, Instagram and so many other internet start-ups do.

"Twitter and Facebook are buying data centres. You have to sell a lot of ten-cent CPM ads to pay for a data centre in Oregon," Caldwell said.

Free services also create what Caldwell calls "disintermediation of responsibility". It impacts the corporate culture of these places," he said. "Sorry we screwed you over, but hey it's a free service" justifies all manner of user-hostile decisions.

On 1 July 2012, Caldwell blogged a lament for what Twitter might have been. The response encouraged him to launch a Kickstarter-style fundraising campaign, and a month later he'd raised US$803,000 from more than 12,000 backers to create App.net. Most had paid a $50 annual service fee, or $100 for a developer account.

By the start of 2013, App.net had around 30,000 users and 100-odd third-party apps in development, and the price had been dropped to $36. The initially-basic API had been extended to include a framework to add arbitrary data to messages (think machine-to-machine communication or the moves in a chess game), group and non-public messaging, photo sharing, and file management for the 10GB of storage that now came with each account.

But to produce all this working code, App.net had been burning through the cash. "In terms of us having the employees that we do and prioritise the way we have, it's because we're trying to move as fast as we can, so that this isn't just a science project," Caldwell said.

App.net now hopes to mirror the freemium-based success of other ad-free services such as Dropbox, Evernote, Flickr and programmers' code-charing site Github. The free accounts announced today are limited to following 40 people, 500MB of file storage, and a maximum file size of 10MB. Free users must be invited by an App.net paid member, who earn more storage space for every invite.

So will it work?

Leslie Nassar, head of technology for Amnesia Razorfish, thinks limiting the number of users a free account can follow is smart. "It limits the spread of bots and spammers without stifling the discovery of interesting topics and people," he told Technology Spectator.

"File storage is a little trickier. Every man and his dog offers file storage. I can get gigabytes for free. I can get gigabytes as a paid, premium service. As a paid service, it's a market that is exceptionally well catered-for with the likes of Box and Dropbox.

"That said, opening up the user base to free accounts and combining it with a more flexible (or less inhibited) file storage model could expand App.net beyond a 'Twitter clone' and into spaces held by Yammer and 37signals' Campfire; namely it moves from a broadcast-only utility to a collaborative service."

Consultant Kate Carruthers, a supporter of App.net since its early days, agreed that the freemium move makes sense, but shared Nassar's doubts.

"Space is relatively cheap now, so that doesn't seem like much of a temptation. The viral model, similar to Dropbox, where you and friends get benefits from additional users might be attractive," she told Technology Spectator. But getting people to switch their primary social network will still be a problem.

Caldwell says App.net's file storage is about control. Think of it as your personal cloud storage. "When you upload a photo, it doesn't get stored in the app developer's account, it stays in your own App.net account," he told Technology Spectator.

"It's a different take on how social data is stored... We're used to signing in for each site, and we never pay for the hosting on that site. We just kind of hope they make enough money for us that they can make their business work. You never think about or get exposed to the fact that those files are being stored somewhere on someone's servers and that they're no longer yours, more or less."

Carruthers isn't so sure. "To me it looks like App.net is focusing on rewards that might not be so important to people that they'll sign up for the free service and then convert to the paid version," she said.

App.net's entire model has been controversial and counter-cultural for tech start-ups from the start, says Carruthers. "It's more like a telco than other social networks, who have adopted advertising driven approaches," she said.

Indeed, Caldwell has received threats from people who believe App.net's user-pays model is designed to exclude people of colour from the internet. But he's unfazed.

"Truly better products are going to be built this way. I don't think people should use App.net like medicine because it's good for you or something. I actually think that this platform is going to yield better software, and be better for developers and better for users at scale."