The Internet of Tweets


Everybody’s got advice on what Twitter needs to do at its current crossroads. The answer might lie in revisiting the moment they first broke geeks’ hearts.

I took it as a bad sign that Will.I.Am was hanging around backstage at a conference for software developers.

It was hard to understand why nobody’s favorite rapper was in the building, since he had expressed no particular interest in software engineering prior to that moment. To be fair, Will.I.Am’s interest in technology has become clearer in recent years; he recently put out a smartwatch called PULS, which is a good alternative to the Apple Watch for people who prefer a smartwatch that doesn’t work. But back in 2010, Am’s inexplicable presence at a technology conference was a good indicator that things were changing in a way that didn’t necessarily favor the folks who were there to hear stories about Apple apps, not apl.de.ap.

We were at Twitter’s first developer conference, an event called Chirp. And indeed, Will.I.Am’s armchair interview was just one of a few clear indications that Twitter really wasn’t just for geeks anymore. Instead, Twitter had suddenly blossomed into a company that aspired to be a player in media and entertainment and advertising, with its eye focused on becoming the giant, publicly-traded company it is today. Twitter began to take its first tentative steps away from its geeky roots, which set the stage for a nerd backlash that still hasn’t fully abated.

But the business types whom Twitter had been courting as its new BFFs turned out to be just as mercurial as the early geeks once the company finally did go public. Investors and pundits have taken issue with everything from the company’s CEO to its often-desultory product direction to its lackluster growth in the number of active users logging into the service each month. This led to a precipitous recent drop in its share price, the delivery of a 360-tweet-long diagnosis of the company’s ailments from shareholder/cheerleader Chris Sacca, and culminated in today’s announcement that Twitter CEO Dick Costolo is stepping down from his position.


While the Internet has never shied away from offering unsolicited advice to Twitter’s leadership, it has perhaps never faced such impassioned arguments about how the company can get its mojo back from some of its closest supporters. While many businesses would envy the fact that Twitter has pretty huge revenues and a user base that is enormous in comparison to almost anything except Facebook, it’s become conventional wisdom that the company needs to work to find its bearings again.

And oddly, getting Twitter back on track may mean taking another look at one small bit of news that got overshadowed by a Black Eyed Pea on that day five years ago.

Let’s Get It Started

In the spring of 2010, Twitter was just graduating from being the hottest startup in tech to being a genuine cultural phenomenon. Chirp was the company’s first big conference, a glitzy coming-out for the company and its leaders. Advertised as its “first annual” event (there have been none since), Chirp showed that Twitter was a big enough institution to be able to command the attention of a theater full of geeks and press. I found myself there as a sort of legacy admission, thanks to the fact that I’d been working on building some apps on top of Twitter’s platform, and had known the founders of the company back when they were just ordinary geeks.

Twitter’s spate of announcements at Chirp included big news like Dick Costolo announcing the company’s first real foray into offering advertising. But within the tech industry, the real headlines started the night before with a low-key announcement that Twitter would be buying or making its own Twitter-branded apps for smartphones. (As absurd as it seems now, until that point, searching for “Twitter” on most devices wouldn’t yield a Twitter app that you could just use to read or post tweets. Instead, a confusing array of Twitter “clients” offered a selection of blue bird icons and terrible portmanteau names beginning with “tw-”.) At Chirp, Twitter made clear it would be offering its own, official app, just like Facebook and all the other popular services.

Then-CEO of Twitter Evan Williams at Chirp, explaining the biggest challenge facing Twitter as a company. He since founded this site, Medium, which is also too hard. (Photo by Scott Beale/Laughing Squid)

The announcement of Twitter’s first-party mobile apps, coming just on the eve of Chirp’s start, rankled some of the most influential developers in the crowd, despite the obvious logic of the company’s rationale. These developers were independent coders who now faced the challenge of competing with the very company whose service made their products possible. From the standpoint of the market, or of regular users, Twitter was right to dismiss the complaints of these developers at the time. Twitter’s decision to make its own client apps did yield a significant simplification of the experience for normal users, and set the stage for the company to be able to display ads on the service, which might have been impossible through third-party client apps.

But unfortunately, this dismissal of developer concerns, however legitimate, would leave a lasting mark on the company’s relationship with the geeks who had first embraced the service. Even though the vast majority of programmers who built tools on top of Twitter weren’t making client apps, the mood of the entire developer community was led by those who were. And once the developers got upset, their disgruntlement informed the attitudes of the entire early-adopter tech community. Later announcements to these same geek audiences would be plagued by vague messages, unclear goals, or accurate communication of actual bad decisions on the company’s part. All of this merely confirmed the suspicion of many developers that Twitter didn’t love them anymore. For nerds, Chirp could be seen in retrospect as marking the official end of Twitter’s Good Old Days.

Where is the love?

Twitter’s struggle to win over investors is mostly due to the fact that most folks on Wall Street don’t really have a nuanced view of how social networking and social media platforms work. The market wants metrics, but picks and chooses which metrics they care about pretty arbitrarily, based on the whims of not-exactly-Zuckerberg pundits like Jim Cramer, or by looking at the numbers of other services like Facebook’s much-vaunted Monthly Active User (MAU) count — the number of people who log in to use the service each month.

The thing is, MAU numbers are a pretty arbitrary measure of the utility of a media-based service like Twitter. Google’s investor briefings almost never ask about how many YouTube users are logged in each month — they care about how many videos are watched, and how many ads are seen, and maybe they care a little about the fact that YouTube is the second most popular search engine on the Internet after Google itself.

By contrast, relatively few users go to Facebook to do a search on a news story or to find a particular video. And being measured against Facebook really only makes sense if there’s a zero-sum game where people are replacing one service with another. For example, in the photo-sharing realm, Flickr’s lead a decade ago has been almost completely supplanted by Instagram’s dominance today. In that context, a user-for-user comparison makes sense.

But when Anita Baker and Cheryl Lynn have beef on Twitter, it’s only on Twitter. And a ton of people are going to read about it by seeing their tweets embedded on a blog or news site, just like they discover YouTube videos that are embedded on those same sites. Whether they’re logged in to the service or not is irrelevant. (The fact that no ads are displayed along with those embedded tweets certainly is relevant.)

So what about some non-MAU metrics? Would that appease? What about shifting the conversation entirely? Other companies have been able to broaden focus from their ad-based businesses by getting into nascent markets that hold a lot of promise. This is where we encounter things like wearable 3D goggles and smartwatches, but R&D-intensive hardware folly is generally the domain of much bigger (and richer) tech titans like Apple and Microsoft and Google. And, uh, Will.I.Am.

When Will.I.Am DJed the afterparty for Twitter’s Chirp conference, he literally followed Don’t Stop Believin’ with Sweet Child o’ Mine, which is hard to read as anything but contempt for the conference’s attendees.

I Gotta Feeling

The funny thing about many developers being convinced that Twitter doesn’t care about them is that it’s pretty clear that Twitter is better to its developers than almost any other social networking or social media company. I was at Chirp because I’d helped built a tool that relied on Twitter’s data, not as a client app but for analyzing and understanding a user’s activity on the service.

Lots of other companies went that route, too — with a number of them selling for hundreds of millions of dollars as “enterprise analytics services”, and many more getting lots of funding for the apps they’d built on top of Twitter. That’s a striking contrast to the ecosystem around, say, Instagram or Pinterest, which have almost no similar success stories. In fact, if we look at the landscape of major social networks or messaging applications that developers could have relied upon in the years since Chirp, a striking pattern emerges:

  • Google: They launched Google Buzz (hey, remember that one?) and then mercy killed it a few years later. They’re about to do the same thing with Google+. There has never been a meaningful client app for Google+ from any developers outside of Google.
  • Instagram: There are a few third-party tools that do specific tasks like regramming images or making collages, and some reading apps for platforms like iPad where Instagram was late. Still, Instagram hasn’t really spawned many successful full-featured client apps. A handful of analytics apps exist, but none of them have gotten that huge.
  • Tumblr: The few attempts at building client apps or services have generally been stymied by the company. Almost none of the analytics services built for Tumblr have been very successful.
  • LinkedIn: There used to be a few ways to build apps and services around LinkedIn, but almost all of them were shut down over the past year.
  • Pinterest: The company promised an interface for developers for years, but has only shipped a limited release of some tools for brands to be able to interact with the service. There are no Pinterest client apps.
  • Snapchat: Pretty much nothing.
  • WhatsApp: Same deal, you take what they give you.
  • Facebook: They’re not too bad, providing access to a lot of useful data, but recent privacy improvements for users have meant fewer features for developers that were relying on previously-accessible data.
The truth of it is, when compared to other social networking companies, Twitter ends up looking like one of the most developer-friendly big platforms.

As a developer trying to build on top of all of these services, most of the ones that were of the same vintage as Twitter screwed us. The newer ones don’t care about third-party developers at all. None of these companies have ever cared about enabling developers to make a client app. Yet none of them has earned the same scorn or derision from developers as Twitter has.

This may be changing a bit, as we see a generational split amongst developers, with newer coders unaware of Twitter’s polarizing past. Indeed, Twitter’s recently made available a whole new suite of developer tools to simplify everything from displaying ads to keeping track of what’s making a mobile application crash on your phone. It’s still early, but every indication is that these new developer tools are already becoming popular, and that the developer services that Twitter acquired haven’t lost any credibility by becoming part of the company’s portfolio.

Enough years have passed that some of the older disgruntled developers may either have mellowed over time, or simply may not be relevant in shaping community opinion anymore. It’s actually possible that Twitter has another chance to make its core platform appealing to developers again, if it can find the right tools to offer them.

Scream & Shout

If we go back to that day in 2010, we can find out exactly what kind of things excited Twitter’s first wave of developers.

The start of the Chirp conference was a flurry of speeches from top execs, and of course the headlining Will.I.Am appearance. But by afternoon, the splashier news had passed, and it was time to get down to nuts-and-bolts developer conversations. The highlight for coders was a preview of an upcoming developer feature that seemed incredibly promising: Annotations. This nerdy new capability promised that developers would be able to pack more information into each tweet.

Annotations promised to upgrade tweets from being a 140-character postcard to being a 140-character message written on the outside of an envelope. What was inside the envelope? Whatever a developer could imagine.

At the time, this idea for Annotations was radical. Twitter’s core service had only just begun to stop displaying the fail whale, a whimsical and all-too-common illustration that popped up whenever the service was too overloaded to respond. To imaging going from not just being able to deliver tweets in realtime to anyone in the world, but to delivering almost any message in realtime to anyone in the world was a shocking leap forward.

Reaction from developers was immediate, and ranged from enthusiastic to downright rhapsodic.

Even some of the more sober analyses of Annotations saw the potential for Twitter to be transformed as a service. And the excitement over the feature was clearly justified — many of the ideas that developers immediately suggested for Annotations, such as including photos or videos in tweets, have become indispensable parts of the platform’s current success.

Even Twitter’s most interesting newer features, like the “cards” in tweets that show excerpts and thumbnails of linked stories, or the experimental “buy” button that the company has been testing to enable e-commerce, are all things that Annotations had promised to make possible years ago. These capabilities were hinted at in the first reviews of the feature written back in 2010.

But more exciting than the ability to add new little buttons or links within Twitter is what Annotations could enable outside of Twitter. But that potential was never discovered, because the company soon stopped talking about the upcoming feature, and eventually abandoned the effort altogether without even so much as a tweet marking its demise.

#thatPower

One of the fastest growing categories in consumer technology these days are smart devices, which take ordinary household items and connect them to the internet. Each one usually comes with its own app. One app will tell you when your smart lightbulb is going to burn out, and another app for your smart toaster will tell you when it’s going to burn your toast. Each of the individual gadgets is pretty cool, but all of this noise on top of the nonstop barrage of notifications that light up your phone during the day seems like a recipe for message fatigue. It’s not difficult at all to make the argument that these notifications would be a better experience as delivered in Twitter’s signature stream.

Delivering these notifications is already done regularly by developers today, but each has to reinvent the wheel when it comes to handling tons of users or building the necessary code to support features on each new phone or device platform that comes along. If a smart device is fortunate enough to succeed, its makers get to face the same kind of fail whales that plagued Twitter for years. Piggybacking on Twitter’s infrastructure and could solve a lot of these issues, especially if they relied on an updated version of the features that Annotations promised.

This concept already exists for coders — it goes by name like a “messaging bus” or a “message queue”, though these services tend to be far more complex and expensive than it is to simply send a tweet. More to the point for developers, Twitter’s also a lot more fun than using an “enterprise message bus” service.

There’s a more subtle point here, too: Twitter enables connections between accounts. What exists today as a social network between people who follow and reply to each other could tomorrow expand to be an information network between devices that could follow and reply to each other. Telling your smart smoke detector not to set off the alarm when the smart toaster has said it’s about to burn the toast is currently a task for only the most stalwart geeks. There’s no reason that kind of connection couldn’t be a new use for the “follow” button.

But the point here isn’t to outline every possible interesting feature that can come from a fully-networked Internet of Things. Hell, I don’t even know what the most interesting capability would be. The key here is that Twitter is the best service to offer these features. It’s simple, fun, familiar, and available already on every device that matters. None of those things are true about the other “Internet of Things” platforms that have been announced.

Twitter could charge apps or devices (or humans!) that send high volumes of messages with Annotations, diversifying its revenue stream from simply being advertising-based, and building trust with any developers who still had lingering concerns about Twitter’s developer strategy by making these new features sustainable.

Chris Sacca speaking at Twitter’s Chirp conference in 2010.

Boom Boom Pow

The thing is, it’s easy to write fan fiction imagining all kinds of features being invented by Twitter. I want a pony! And a new CEO who’s from an underrepresented community!

But most of the smart suggestions are probably already on Twitter’s roadmap, including those on Chris Sacca’s list. These suggestions tend to assume, though, that the only audience that matters are Twitter’s mainstream, non-technical users. This is an odd assumption given that many of Twitter’s most engaging signature features were pioneered by its early adopter nerds.

And it’s essential for Twitter to get out of the MAU metrics game with investors. We can imagine a hundred different ways to juke the stats and try to distract fickle financiers, but revisiting Annotations may be one of the few ways to do so in a way that’s based on substance — actually providing the kind of meaningful platform innovation that developers actually want.

There’s something unique and distinct in being an information network instead of just a social network. Twitter has been that network, and can be again.

Revisiting Annotations wouldn’t be a magic bullet. But it could be a way to reconcile Twitter’s past and its future, to appease both developers and investors. And it might be a chance to rekindle some of the unqualified enthusiasm that so many saw for the company at its coming out party half a decade ago, before it’s 2000-and-late.

Sure, you’ve got a better idea for how to fix Twitter, or a funnier Will.I.Am joke. That’s what the responses option is for — let us hear it!