It’s been a rough decade for Facebook and Twitter. The two entered the 2010s with so much promise. On Twitter, you could talk to just about anyone. On Facebook, you could treat yourself to endless amusement, scrolling through photos and statuses your friends probably should not have posted.

Then, all hell broke loose. Social media was no longer fun. It was toxic. And it got this way for a reason.

Ultimately, Facebook and Twitter descended into chaos by their own doing. Over the course of 10 years, they made a series of misguided product decisions that transformed them from online amusement parks into hellscapes. Here’s how it happened.

2010

Time magazine named Facebook CEO Mark Zuckerberg its Person of the Year in 2010, in what was likely the apex of social media’s popularity. The site, back then, had a mere 550 million users. “It started out as a lark, a diversion, but it has turned into something real, something that has changed the way human beings relate to one another on a species-wide scale,” the Time article said.

Time approached Zuckerberg with an optimistic tone -even reverence -aware of his accomplishments and power. Throughout the profile, however, there was a sense of uncertainty as to where that power would lead.

As Time was interviewing Zuckerberg, then-FBI director Robert Mueller walked into the room. A Facebook spokesperson tried to put that off record, but the reporter wouldn’t have it. “They shook hands and chatted about nothing for a couple of minutes, and then Mueller left,” the Time profile said. “There was a giddy silence while everybody just looked at one another as if to say, What the hell just happened?”

Facebook looked like this at the time:

The social network’s News Feed was up and running, but Facebook was primarily a directory where people could share their lives with friends and families. Public content -from celebrities, news sites, and politicians -was barely present.

Facebook had grand designs. It wanted third-party developers to build applications, and it wanted to entrench itself more deeply into services around the web. To do this, it loosened restrictions on how much data developers could hold onto. “We’ve had this policy where you can’t store and cache any data for more than 24 hours, and we’re going to go ahead and we’re going to get rid of that policy,” Zuckerberg said at Facebook’s 2010 F8 developer conference. According to a CNET report, the audience cheered.

Twitter, meanwhile, looked like this:

Twitter cofounder and then-CEO Evan Williams was sending updates about his plans to watch Apple CEO Steve Jobs present at Apple’s Worldwide Developers Conference. He was also tweeting about a new ride-hailing service called “Ubercab.” Quite a day.

2011

In 2011, Facebook introduced the Subscribe button, a new feature meant to make the service bigger and more influential. The button “lets you hear from interesting people you're not friends with -like journalists, artists and political figures,” Facebook said in its announcement. “Just click the button to get their public updates right in your News Feed.”

The Subscribe button was addictive, both for Facebook’s users and for public figures. But by creating the Subscribe button, Facebook sacrificed some of its friends-and-family feel for influence and size.

Twitter in 2011 introduced photo sharing, taking a major step toward becoming a more visual, less text-heavy network. “140 characters, now worth 1,000 words,” then-executive chair Jack Dorsey wrote.

2012

Facebook rolled into 2012 with more than 800 million users. It went public that May, increasing the pressure to grow its user base and revenue each quarter. “Facebook aspires to build the services that give people the power to share and help them once again transform many of our core institutions and industries,” Zuckerberg wrote in the company’s IPO filing.

Twitter, meanwhile, was the defining political platform of the 2012 US presidential election. A post-victory tweet with a picture of Barack and Michelle Obama became the platform’s most retweeted post ever.

Eight days after the election, Facebook tried to catch up. It introduced a mobile share button -a retweet clone -that would add speed to its service, making it more attractive for news organizations and politicians.

2013

With 230 million users, Twitter went public in 2013, increasing the pressure to grow its user base and revenue.

Facebook spent 2013 copying more Twitter features. It added hashtags in March and a trending column in August. The signal was clear: Facebook badly wanted more public, real-time content.

Traditional news publishers answered Facebook’s call, working hard to build operations that would translate their work from something people would pay for into something they would click. Nontraditional publishers crashed the party as well, including digital-only news and entertainment sites (BuzzFeed included), along with made-for-Facebook fake news.

“I was seeing those sorts of sites all over the place with large followings and they were getting good traffic and I just thought to myself, Well I could do that,” Jestin Coler, who set up the fake news site National Review in 2013, told BuzzFeed News.

Facebook’s decision to increase public content and add a mobile share button ignited fake news on its service. People spread all manner of lies without much thought -the share button removed almost all hesitation from the act of sharing -as long as those lies confirmed their worldview.

2014

Twitter began experimenting with a “retweet with comment” feature -now known as the quote tweet -in 2014. The feature would introduce a new practice to the service: dunking. This would add a level of public viciousness to Twitter. “The biggest problem is the quote retweet,” former Twitter product head Jason Goldman told BuzzFeed News earlier this year. “It’s the dunk mechanism.”

In 2014, Twitter users showed a new capacity to do harm using the service’s tools. That year, Gamergate erupted. Gamergate was a harassment campaign against women in the game industry, and it weaponized the retweet button. Chris Wetherell, the project lead on the retweet button project back in 2009, watched with horror as Gamergate unfolded, understanding the damage his creation had wrought. “Ask any of the people who were targets at that time, retweeting helped them get a false picture of a person out there faster than they could respond,” he told BuzzFeed News. “We didn't build a defense for that. We only built an offensive conduit.”

Facebook had an eventful year in 2014 too. It revealed it had manipulated users’ emotions via a News Feed experiment, it bought the virtual reality startup Oculus for $2 billion and the messaging app WhatsApp for $19 billion. Facebook was now massive -with more than a billion users on mobile alone -and it wasn’t dedicating sufficient resources to monitor its products.

2015

Facebook’s desire for public content caused its service to flood with posts from public sources, making normal sharing between friends and family feel intimidating. By 2015, people were sharing fewer original posts on Facebook, leaving a void that fake and sensationalized news purveyors happily filled. That year, Facebook knew it had a problem on its hands, and pledged to show fewer hoaxes.

In 2015, Ted Cruz’s US presidential campaign began working with the data firm Cambridge Analytica to target messaging to voters. Cambridge Analytica, according to a Guardian report that year, was using “psychographic profiles” of US citizens built with Facebook data that researcher Aleksandr Kogan collected and kept under the site’s loose developer policies.

Twitter -focused on making money and growing its user base to please the public markets -largely ignored a festering harassment problem that worsened after Gamergate. “We suck at dealing with abuse,” Twitter CEO Dick Costolo said in early 2015, in a rare moment of self-reflection. Costolo stepped down, leaving the mess to Dorsey, who returned as Twitter’s CEO.

2016

Twitter introduced an algorithm to its feed in 2016, and the trolls went wild. One summer evening, a mob sparked by Breitbart editor Milo Yiannopoulos acted so brutally that their target, Ghostbusters actor Leslie Jones, quit the service. Twitter had become a honeypot for assholes.

Facebook, meanwhile, became the defining platform for the US presidential election in 2016, but perhaps not for the reasons it hoped. The top fake news stories went more viral on Facebook than real news, hyperpartisan pages spread misleading information about the candidates, and teens in Macedonia ran scores of pro-Trump, fake news websites to cash in on the frenzy.

A Kremlin-linked troll farm, the Internet Research Agency, seeded posts on Facebook and Twitter meant to sow discord and tear American society apart.

And after Ted Cruz pulled out of the race, Donald Trump’s presidential campaign, which would go on to victory, began working with Cambridge Analytica.

2017

After nearing saturation in the US market, Facebook had for years worked hard to expand in international markets. The company succeeded, reaching 2 billion monthly users by mid-2017, but it was woefully unprepared to keep its expansion safe. In 2017, lawmakers in Myanmar posted hate speech and lies about the country’s Rohingya Muslim minority. For these lawmakers, Facebook became a key tool in a violent campaign that displaced 700,000 Rohingya in what the UN called a genocide.

At Twitter, a contractor on their last day of work deactivated President Trump’s account.

2018

In March 2018, the New York Times and the Guardian published damning stories revealing the full extent of Cambridge Analytica’s work on the 2016 election. Kogan gave the firm data on up to 87 million Facebook users, yet only 270,000 people consented to hand their data over. The data identified people’s personality traits, helping tailor messaging based on whether they were agreeable, for instance, or religious. Christopher Wylie, the whistleblower, declared: “I made Steve Bannon’s psychological warfare tool.” A month later, Zuckerberg was explaining himself on Capitol Hill.

Having been called out and dragged before Congress for its failings, Facebook began to patch up some of the vulnerabilities in its core service. But it still struggled to contain the damage being done in its satellite apps. Like the retweet and share, WhatsApp had a “forward” button that allowed news, memes, and rumors to spread rapidly, passed on without thought or hesitation. In July 2018, a mob in the Rainpada village in India beat five strangers to death after viewing untraceable videos on WhatsApp that warned -baselessly -of child abductors. Facebook would limit message forwards in 2019.

Twitter spent a good chunk of 2018 defending its decision not to ban the conspiracy theorist Alex Jones, giving the impression that it was still unsure where to draw the line on its speech rules. Then, in September, after Apple banned him, Twitter followed suit.

2019

As the end of the decade neared, Facebook and Twitter were scrambling to clean up their services. Twitter has banned political ads ahead of the election, and Facebook is considering restrictions of its own. Twitter said it will label politicians’ tweets when they break its rules, and Facebook is trying to reemphasize friends-and-family content.

Over a decade, the two services prioritized growth and influence over safety, creating a mess they’ll spend the next decade cleaning up. They’ve reached their goals, achieving great wealth and power as a result, but at a cost to society that won’t be fully calculable for some years to come.