An Algorithmic Feed May Be Twitter's Last Remaining Card To Play

Twitter is out of options — except for one big one.

In the nearly two years since its November 2013 initial public offering, Twitter has tried seemingly everything to grow without alienating its loyal-until-death group of core users. It's made itself more visual by expanding images by default. It's quashed the "RT" in favor of the simpler quoted tweet. It's introduced autoplay video. It's added an e-commerce layer by giving brands the option to sell products straight from tweets with a "Buy" button. It killed a confusing rule preventing people from sending direct messages to folks who don't follow them. And it's extended the character limit for those messages to 1,000 characters from 140.

In other words, Twitter has done nearly everything it can to refine and enhance the way tweets present information, without touching the order in which they're published. That order -- chronological, with no algorithm elevating the "best" tweets -- has long been sacred ground for the company. The order is cherished by Twitter's core users, with many arguing its organic surfacing of news and conversation is what makes Twitter work. But with every other card on the table and Twitter's Wall Street masters still unhappy, the company's last remaining move may be the one it's long avoided: the application of an algorithm, or code, that decides which tweets get shown and which don't. The end of the order.

"It's not an easy task, but I think it's a necessary task," Nate Elliott, vice president and principal analyst at Forrester Research, told BuzzFeed News. "If they implement an algorithm properly, it could be the savior of their platform."

The deluge of tweets flooding into Twitter each day (some 500 million daily) can bury the best among them, undermining the conversations and engagement on which Twitter built its success. "I think it was very good and clutter caught up with it," Elliott said. "It's a shadow of its former self."

That complaint is fast becoming the main knock against Twitter as it attempts to become a platform used widely to spread and consume news and information. "The world's very best content is already inside of Twitter," Twitter investor Chris Sacca recently said in a lengthy blog post. "Yet, for most people, using Twitter to see that great stuff is too hard."

An algorithmic feed could solve that problem by automatically elevating tweets when they hit certain metrics that indicate they are high quality. Facebook, for instance, uses an algorithm to filter its feed, creating less work for users trying to get value out of the platform. But an algorithm would also be risky to introduce. A bad one could surface irrelevant tweets and quickly turn users off. It could also bury good tweets that don't hit certain superficial metrics.

Twitter in its current form also has the power to elevate ideas very quickly, something that could be lost if a filter is applied to its feed. During the protests in Ferguson, Missouri, University of North Carolina assistant Professor Zeynep Tufekci pointed out that awareness of the events in the city exploded on Twitter's unfiltered feed while remaining relatively buried on Facebook's algorithmically curated one until they became national news.

"I wonder: what if Ferguson had started to bubble, but there was no Twitter to catch on nationally?" Tufekci wrote in a Medium post. "Would it ever make it through the algorithmic filtering on Facebook?"

In an interview with BuzzFeed News, Tufekci said she's dubious that an algorithmic feed would solve Twitter's problems. "Wall Street's overloaded focus on the idea that heaping on a serving of algorithmically selected 'oh, look, already popular tweets' is all that needs to be done seems to misunderstand the value of differentiating this product and this platform, and the reason why the current base uses it," she said.

But those hoping for a change do have ammo. According to Sacca, almost 1 billion people have tried Twitter and left -- likely put off by the daunting task of following and unfollowing accounts until they develop the right mix. "It can be an incredibly important tool for reinvigorating the platform," said Elliott of an algorithm, citing Sacca's numbers.

In fairness, Twitter has made an effort to surface the best content posted to its platform. It recently released a "While you were away" feature that shows high-interest Tweets you missed while the app was closed. It experimented with inserting favorites, not just tweets and retweets, into people's timelines (though that experiment has concluded for the time being). It developed a notification system called "Magic Recs" to tell you when something is bubbling up in your feed. And last week, BuzzFeed News reported that Twitter is working on "Project Lightning," an effort to curate tweets generated around live events.

But Twitter's experiments haven't been enough to rev its business to the point where it's living up to the expectations that come along with its $23 billion market cap. Last year, the company brought in just $1.4 billion in revenue.

An algorithmic feed could help Twitter from a business standpoint, said Jason Stein, CEO of social media agency Laundry Service. "It would make the company even more appealing to advertisers and, in turn, much more valuable," Stein said. Most ad buying agencies are not set up to buy ads on the fly, Stein said, meaning they miss out on the most valuable part of Twitter, its real time nature, and can't make campaigns that fit in. (Try running an ad for Intel on Twitter in the middle of the NBA playoffs.)

Stein said Twitter's Project Lightning, which could conceivably someday allow agencies to plan and buy ads to coincide with live events, is a good start. He believes the platform will eventually blend algorithmic, curated, and raw elements. "It's clearly what it has to do to grow," he said. "And if they do this it could be hugely successful."

Twitter's current leadership shake-up will likely bring the algorithmic feed question to the fore. The company's departing CEO, Dick Costolo, was dogged with questions about his ability to grow the platform, and, for the interim at least, his replacement, Jack Dorsey, will be tasked with picking up the slack. The introduction of Project Lightning showed Twitter is willing to bring curation to its product, though its main feed, for now, remains relatively untouched. Deciding whether to take that next big step will be the most significant question facing Dorsey.

In a worst-case scenario, an algorithmic feed could turn Twitter into an inferior version of Facebook, which might, in turn, alienate its core users. But Twitter is a company motivated by profit. And if that worst-case scenario juices revenue, it could prove to be one its investors accept, even as core users decry it. Said Forrester's Elliott: "If they become a worse version of Facebook, at least it will be for a more profitable future than they have now."