Notice I said “wedding” and not “honeymoon.” Sarver has for some time been Twitter’s first line of defense between the company and a occasionally disgruntled army of third-party developers. The relationship between those two entities has been testy at best and an out-and-out war at worst for the better part of two years. Unfortunately, mid-2011 was no time for a honeymoon.
[aditude-amp id="flyingcarpet" targeting='{"env":"staging","page_type":"article","post_id":345763,"post_type":"story","post_chan":"none","tags":null,"ai":false,"category":"none","all_categories":"dev,","session":"A"}']I met with Sarver in Twitter’s San Francisco headquarters, where bikes stood in racks against the walls and neon plastic deer and framed cross stitch decorated the foyer. It’s a hipster paradise, and the clean-cut, blazer-wearing Sarver seemed at once infinitely at home and slightly out of place in it.
But where Sarver is unquestionably beleaguered is in the relatively public forum of Twitter’s ecosystem developers: the men and women who develop apps that use Twitter’s APIs. They try to build businesses that complement Twitter, and when anything goes wrong with that setup (for example, Twitter makes a minor change to an API), Sarver is the de facto shock absorber. And that explains why he skipped his honeymoon to deal with his unruly constituency.
AI Weekly
The must-read newsletter for AI and Big Data industry written by Khari Johnson, Kyle Wiggers, and Seth Colaner.
Included with VentureBeat Insider and VentureBeat VIP memberships.
For such a young man, he handles the role of “the guy developers love to hate” with a large measure of grace and resilience.
“The ecosystem is over 750,000 developers,” he told me. “It’s a massive thing. Any community of that size, you’re never going to keep everyone happy at all times. … Some people might be mad at me, but I try to hear what they’re trying to say and fold that back into what we’re trying to do.
“The most critical people are usually the most valuable ones. They tell you where the points of friction are and how you can improve.”
Unlike Sarver, Twitter did get a honeymoon. There were a few years when it was the much-beloved social platform of tech insiders and hipsters everywhere. Developers flocked to it, drawn by its openness, flexibility and simplicity, building a wealth of Twitter apps and integrating it into their own web services.
Now the honeymoon is over. For Twitter to survive challenges from rivals like Facebook and Google+, much depends on whether it can continue to hold onto the love of its developer community.
“The most critical people”
[aditude-amp id="medium1" targeting='{"env":"staging","page_type":"article","post_id":345763,"post_type":"story","post_chan":"none","tags":null,"ai":false,"category":"none","all_categories":"dev,","session":"A"}']
In 2007, the infant microblogging service had yet to establish any real foothold outside San Francisco techies, yet it was trumpeting the wild success of its APIs. Co-founder Biz Stone said at the time, “The API has been arguably the most important, or maybe even inarguably, the most important thing we’ve done with Twitter. It has allowed us, first of all, to keep the service very simple and create a simple API so that developers can build on top of our infrastructure and come up with ideas that are way better than our ideas.”
Yet for many developers, it seemed that as soon as an outside party came up with an innovative idea, Twitter jumped onto the same idea too, stomping on the little guys. A long list of Twitter’s features, including mentions and lists, first appeared as features of third-party Twitter clients, only to be incorporated into the main service.
Of course, since such features are often dictated by user behavior and requests, just how much Twitter was scamming off the ideas of others and how much it was simply responding to user needs are unknown.
What we do know is that the company began to seem opaque to the developers who used its API. By the spring of 2010, when it hosted its first developer conference, developer angst had reached a fever pitch.
[aditude-amp id="medium2" targeting='{"env":"staging","page_type":"article","post_id":345763,"post_type":"story","post_chan":"none","tags":null,"ai":false,"category":"none","all_categories":"dev,","session":"A"}']
As the conference opened, the world learned that Twitter had acquired Tweetie and crowned it the company’s official iPhone app. Other makers of Twitter iPhone apps felt blindsided and betrayed. Businesses went down not with a whimper but a bang. Development shops started wondering aloud what areas of the map were safe territory and what Twitter might decide to annex next for itself. No one seemed to be getting or giving any clear answers.
Twitter had built an ecosystem before it decided on a business model, and while it tried out different avenues and experimented with new features, it wasn’t disclosing all of its roadmap to all of its presumed partners. While some folks benefited from a close relationship with the company and its founders, others suffered from a full-on communication breakdown.
It was a situation that was bound to create bad feelings.
“The change in attitude Twitter had since Chirp conference … is just the feeling that they no longer care about developers or their ecosystem,” said third-party developer Arik Fraimovich in a recent interview. “They seem to look at it from the narrow perspective of ‘what’s in it for Twitter?,’ failing to understand that in an ecosystem, all parts need to contribute and communicate. They also fail to understand that the next creative use of Twitter won’t be developed by their developers, but by some API developer — just as it was since their inception.”
[aditude-amp id="medium3" targeting='{"env":"staging","page_type":"article","post_id":345763,"post_type":"story","post_chan":"none","tags":null,"ai":false,"category":"none","all_categories":"dev,","session":"A"}']
A big part of the problem was what plagues many relationships: Lack of clear communication.
Twitter only gradually worked to clear things up. A year after the Tweetie acquisition, after everyone had adjusted to Twitter’s official, in-house suite of mobile apps, the company made an attempt to communicate part of its roadmap, the upcoming acquisition of TweetDeck and a renewed focus on consumer features.
In a controversial memo to a Google Groups list in March, Sarver laid down the law: Don’t make a Twitter client using Twitter’s API. Don’t duplicate Twitter’s features. Don’t rename them. Don’t channel users away from Twitter’s ads and experience.
[aditude-amp id="medium4" targeting='{"env":"staging","page_type":"article","post_id":345763,"post_type":"story","post_chan":"none","tags":null,"ai":false,"category":"none","all_categories":"dev,","session":"A"}']
It was clear, but it was also surprisingly heavy-handed for a company that until then had been known for its open and easy-going approach.
On one hand, Twitter as a business had every right to control how its APIs were being used. On the other hand, it had willingly and knowingly cultivated a thriving sub-economy of Twitter apps and was now telling many of them to kiss off. It’s a communique Sarver regrets, and it started quite the war of words.
Laura Fitton, the perky, audacious founder of Twitter-related startup Oneforty (and an unofficial team mom for the Twitter app community), told me back in March, “Twitter is full of genuinely earnestly awesome people who want to do the right thing, but it has resolutely failed to create the conditions for real business success on their platform.”
I also spoke to Loic LeMeur, the tall, eccentric French entrepreneur behind the rapidly pivoting Seesmic, who told me, “There are two types of Twitter apps, the ones Twitter likes and the ones that are competitive and don’t have good communication with them.”
[aditude-amp id="medium5" targeting='{"env":"staging","page_type":"article","post_id":345763,"post_type":"story","post_chan":"none","tags":null,"ai":false,"category":"none","all_categories":"dev,","session":"A"}']
Of Sarver’s missive, which put Seesmic in the direct line of fire, LeMeur said, “I wasn’t expecting such dramatic changes, but instead of complaining, I have to adapt to it. … Competing with your number one partner is a little dangerous.”
Jesse Stay is another entrepreneur/engineer who has built a business around a collection of social APIs, including Twitter’s. Stay is the picture of a mild-mannered Mormon, a truly kind man who would think long and hard before uttering a negative word about anyone. Still, during yet another minor but uncommunicated technical change at Twitter, he told me, “This move by Twitter is no surprise, as I’ve experienced over and over through the last three years. It’s why we’ve moved to a $29-per-month plan for all the services we offer. … That’s the only way to afford the risk of Twitter.”
Perhaps no one understands the risk of Twitter better than Noah Everett, a man whose name came up more than once in my talk with Sarver. Everett built TwitPic, a highly popular and successful app for uploading photos and sharing them through Twitter. Out of the blue earlier this year, Twitter announced it would be implementing media-sharing capabilities into its web and mobile apps, first for photos, then for videos, cutting a number of photo-sharing apps — including TwitPic — off at the roots.
“TwitPic will continue to still live on,” Everett told me when the feature rolled out to users, espousing the kind of grim optimism I’ve heard from many entrepreneurs facing hopeless odds. “We’ve got a huge base of loyal users and we still want to continue to provide them with the best service possible … regardless of this feature announcement from Twitter.”
But there was a bitter twist. Everett had the previous year started working on a stealth project called Heello, a project attempting to “make email suck less,” as he put it to me. But out of spite or pique or frustration, he decided to turn Heello into a Twitter clone.
“If Twitter can compete with its developers without fair notice, then why can’t we?” he quipped. “Twitter rolling out their photo option to everyone [the day before the Heello site launched] was a complete coincidence, but I’m glad the timing happened that way.”
These voices and many others ring in my ears – and likely, to some extent, in Sarver’s, too – as we face one another over a small table in a small conference room at Twitter.
I ask him if he thinks the company made any mistakes.
“You can always do things better,” he told me. “If you go back to the guidance we gave last March, we could have executed that message much better. The tone and format could have been improved. But the intent is still important: The most important thing for us as a platform is to give people a clear idea of where we’re going.”
And that clarity of communication is all most of these critical voices were asking for.
“Investing in a conversation”
About six months ago, Twitter hired Jason Costa, a specialist in developer relations and a former Googler. Around the same time, the company also saw the return of co-founder Jack Dorsey, who has something of a heroic stature among Twitter third-party developers.
“Jack – he’s a developer,” said Sarver, with perhaps an inkling of good-humored chagrin. “He came from that world and built two awesome businesses out of it. He shows the core part of the company understands and is committed to developers.”
With Costa and a handful of others on hand to help Sarver iron out the community kinks, “We were able to be a lot more proactive about setting up events and investing in that conversation,” Sarver said.
In fact, Twitter is gearing up for a developer tour that will take it outside of its comfort zone — and outside of San Francisco — and bring it into contact with a much wider audience. “We’ve done so many events here, but there’s community everywhere,” said Sarver. “Developers, investors – all those markets.”
But Sarver’s daily bread as its platform chief is all about monitoring ripples in the developer ecosystem. “Every day, that’s what our team is focused on: How what we do will affect the ecosystem and all of our partners.”
Once the developer advocacy team was in place, Sarver said the real fence-mending work started happening.
“Once we were able to actually solicit feedback, Jason started calling developers on a weekly basis,” he said. “We got tons of information. Jack started a thread to get feedback. We aggregated that into five top themes, listed them and put an action plan in place for how to address them.”
As I listen to Sarver, I notice a new specificity that hasn’t been present in our past communications over the past couple years. “We committed to all the API changes having a minimum of 30 days between a notification and making the change,” he told me.
“It’s all about getting the right people to support this stuff – community management tools, better documentation. … We offer a lot of channels, we have a dedicated forum for answering questions, regardless of who is asking it. There’s some limit in the number of people we can talk to on a daily basis, but we try to support everyone as much as possible.”
Of course, some developers have long been asking for the impossible: A complete disrobing of Twitter’s roadmap. And as much as I sense Sarver is willing to help and communicate with external developers, that item on the wishlist ain’t gonna happen.
“That’s the constant back and forth – we just need to give the best guidance we can,” he said. “We can’t put out a complete roadmap for competitive reasons and because roadmaps change. The best way to go about it is to give directional guidance, to give some boundaries that people can build up to.”
“Investment in the ecosystem”
Therein lies the million-dollar question: With LeMeur, Everett and scores of others standing as examples of how not to use the company’s APIs, how, then, can one safely, profitably build a business on Twitter? How is a developer to know that she isn’t suddenly going to be jilted when Twitter decides to go in a new direction?
If that question could be answered in one word, that word is “enterprise.”
Dick Costolo is the company’s relatively new CEO, and he brings a seasoned, trustworthy voice to Twitter’s leadership. When he said over breakfast at Twitter HQ last month that Twitter wasn’t getting into enterprise solutions, I mostly believed him.
“I get a million emails a day about what Twitter could be doing to make more money,” Costolo said. “But … the advertising business will sustain us. We have no intention to scale the data licensing business. It’s important to us that third parties have access to the data. We’re just going to focus on scaling the advertising business.”
Also, Sarver gave a broad hint when he told me, “We try to optimize for the most benefit for our end users.”
In other words, hearkening back to the infamous March memo, the advice for developers is this: Don’t build Twitter clients, but more specifically, don’t build features for the average consumer; instead, focus on business uses that Twitter is not going to pursue itself.
Twitter is still very much experimenting and optimizing its services for ordinary consumers, and the company doesn’t want to have to crush you when it launches a feature that duplicates the entire functionality of your app. But if it has to, it will.
For clarity’s sake, I asked Sarver to outline some areas where developers are absolutely, 100 percent safe in creating Twitter API-powered businesses.
“Analytics is a huge sector for us, obviously,” he began. “Finding those insights you can bring to brands using this real-time corpus of data.” He also mentioned curation (“finding the key tweets to talk about events that are happening, be that the elections, the Olympics, etc.”) and publishing (“figuring out the right content that will resonate with the audience”) as possible areas for investment and development.
Sarver also added that Twitter’s ecosystem is still financially robust, despite the online brouhaha. “During December 2010 through May 2011, we added up all the investment in the ecosystem, and it was around a half billion dollars, and during the same period, there was around a billion dollars in exits,” he said.
Twitter already offers some analytics, but sources say these might end up centering largely around the ads brands place on Twitter itself, leaving the company’s metrics around unpaid tweets and engagement sparse and open for opportunity.
“Analytics is not a zero-sum game,” said Sarver. “It’s way leveled up from what we offer. We hear from brands all the time that they’ve been looking for more ways to get data out of Twitter. Even though there are a bunch of solutions out there, they still want more.”
Finally, he reiterates the kindest-worded warning Twitter could ever communicate: “We’re going to invest heavily in the core consumer experience, and we want to see the ecosystem focus on those other areas of opportunity.”
That might just be the clearest directive the company has ever given to its third-party ecosystem. Stay away from the core consumer experience, and Twitter will continue to show you the love. Cross the line, and you have only yourself to blame if it rolls over yol.
All Sarver, Costa, Dorsey, Costolo and the rest of the team have to do now is hold to those promises, and the hatchet just might stay buried. And then maybe Twitter might get a second developer honeymoon — and Sarver might just get to take his first.