The inside story of how Tumblr lost its way

If you can believe it, Tumblr turned 15 this month. The microblogging site that launched a thousand ships was created by software consultant David Karp in 2007. During a time when longer-form blogging on platforms like Blogger and WordPress dominated, Tumblr offered a short-form, multimedia-rich experience that drew in millions. By 2013, the site was acquired by Yahoo for more than $1 billion, in what some saw as a gamble. “Tumblr is redefining creative expression online,” then Yahoo CEO Marissa Mayer said, calling it “the internet’s fastest-growing media frenzy.” A press release promised “not to screw it up.”

Somehow, they did, and Tumblr has never recovered.

As someone who drank in Tumblr’s ambrosia from 2010 to 2015, it’s pained me to watch its prospects dry up. A recent New Yorker piece called it “popular for being obsolete.” And while I know Tumblr users don’t give a damn about relevance, I still believe in Tumblr and the possibility it holds for a new generation. I spoke with former employees and users to understand how a whirlwind of acquisitions, integrations, and wasted energy left Tumblr wading in a pool of uncertainty — and how the site so many loved might find itself again.

David Karp at Tumblr HQ in February 2012.
Credit: DON EMMERT/AFP VIA GETTY IMAGES

A brief history of Tumblr’s acquisitions

Jack Christian (his name has been changed at his request for anonymity) joined Tumblr’s engineering team just after the Yahoo acquisition. Internal sentiment around the acquisition was initially positive and “culturally, it wasn’t a bad fit,” but Karp’s nimble fingers had built Tumblr on a foundation too small and shaky to support millions of blogs. 

As the team began rebuilding for scale, Yahoo leadership “were really pushing us to get on a lot of their systems and integrate more tightly without really understanding why we built things a certain way [to be] Tumblr-specific around our legacy model. We spent a lot of time working on [integration] instead of thinking about what could really benefit the users,” says Christian. And while Tumblr tinkered, competitors like Twitter, Instagram, and Facebook developed new front-end features to draw users in. 

The Tumblr team began to see the first instances of what would soon become a pattern: leadership’s push for flashy gains that betrayed a crucial misunderstanding that what makes Tumblr unique are the online communities it fosters. 

Tumblr.com homepage circa March 2007 that reads "Tumblelogs are like blogs with less fuss. Tumblr is your friendly and free tool for creating tumblelogs."

Tumblr.com homepage on March 6, 2007, less than a month after the site’s launch.
Credit: WayBack Machine

By 2016, Yahoo slashed Tumblr’s value by more than half, citing its failure to meet sales goals. In 2017, Verizon Media acquired Yahoo, Meyer stepped down, and COO Jeff D’Onofrio replaced Karp as Tumblr’s CEO. A year later, Isabella Kahle joined Tumblr as a product marketing manager on a marketing team that was weathering the storm of all that change. “We were severely under-appreciated, understaffed, underfunded,” she sighs, “and lacked a lot of tools that could have made our work great and potentially grown the business.” 

In August 2019, Tumblr was acquired again, this time by WordPress.com owner Automattic for a paltry $3 million.


We were severely under-appreciated, understaffed, underfunded, and lacked a lot of tools that could have made our work great and potentially grown the business.

– Isabella Kahle, former product marketing lead

“It seemed like a really good move in terms of a fit for Tumblr at the time,” says Christian, “although very quickly, I felt quite the opposite.” Automattic “didn’t really understand the value of Tumblr,” seeing it as just “another space where people go to blog.” Like Yahoo, Automattic clumsily shoehorned Tumblr into its existing infrastructure. “Early on, they were talking about ‘why can’t we just run Tumblr on WordPress? It’s less work,’ without understanding the nuances [of the platform]. There were a lot of opinionated people very high up in leadership guiding a lot of our decisions and strategy without understanding the user journey and impact that we could have in certain areas of growth. It was kind of just all whimsical. [We] wasted engineering time trying to integrate for no benefit to the user whatsoever.” 

Both Christian and Kahle left Tumblr in 2021 and, by that time, Kahle said she had a “gut feeling…that Automattic had kind of given up on Tumblr. The vibe that I got from listening to executives speak about Tumblr [was that] I think they have a mental deadline of when they want Tumblr to prove itself. And that timeline is not feasible based on the team’s bandwidth and support and funding.” Tumblr wouldn’t shut down, but would be put “on autopilot” she predicts, to be kept running without any future development.

Automattic CEO Matt Mullenweg calls the claim “untrue,” citing his own full-time involvement in the brand as further proof of Tumblr’s value to the company. Since the sudden departure of D’Onofrio in January 2022, Mullenweg has stepped in to oversee Tumblr. “We haven’t gotten the results in the past two years that we [had] hoped for,” he tells Mashable. His first priority is growth, to “make the product great so that people use it more [and] tell their friends.”

Matt Mullenweg, CEO of Automattic, who is now managing Tumblr full-time.

Matt Mullenweg, CEO of Automattic, is now managing Tumblr full-time.
Credit: Sportsfile/Corbis/Sportsfile via Getty Images

A product in need of direction 

When Christian joined Tumblr in 2013, a formal product team was just beginning to take shape. Plagued by constant churn, a phenomenon that permeated the entire company, the team never reached its final form. He eventually moved from engineering to product, losing track of how many product leaders he saw come and go. “They all came in with their own opinions of how they wanted to run things without any context of Tumblr’s history or background or data.” 

Leadership busied themselves with “trying to figure out what the next thing for Tumblr could be,” but they “never really took a critical look at what it would take for that feature to improve the community and the platform as a whole,” Christian says. They missed opportunities to pull in the marketing and community teams, people who were close to the users, who might not only identify how to evolve the product, but also “help the public understand why Tumblr is great, why it should be here, and why it’s cool.” Long-time employees and folks like Christian, who had been using Tumblr for five years by the time he was hired, “felt like we knew what we needed to do to help the community and help the product grow. But it was difficult to have a voice in a place where the loudest people had the least amount of context.”

A screenshot of Tumblr.com featuring mock ups of the dashboard and log in fields.


Credit: WayBack Machine

It was not for lack of trying. After observing that users would purposely delete posts after a set number of hours, Christian suggested the site offer an “ephemeral post” that could be scheduled to disappear. As he watched Giphy scrape the site’s API to populate its own database with GIFs ripped directly from Tumblr users, he suggested the company lean into the format as one of platform’s undeniable “strategic advantages.” He also floated the idea of bringing back an early feature called Tumblr crushes that showed a user the nine blogs they’d liked the most posts from and fostered a sense of community and closeness.


It was difficult to have a voice in a place where the loudest people had the least amount of context.

– Jack Christian, former product manager

None of these ideas saw the light of day. It didn’t matter how much support he had from the marketing, data, or engineering teams, “the product direction was essentially led by whoever was in charge that day.” And many of those people were not native Tumblr users or long-time Tumblr employees, which meant they pushed to fulfill flashier, short-term priorities that “shifted every three months” and weren’t rooted in what the platform truly needed.

What they did pursue in the name of remaining competitive felt like spinning their wheels, resulting in “really talented and passionate people doing a great job, but just not working on the right thing.” 

Tumblr’s revolving door

By the time Tumblr arrived at Automattic, Christian says “all the changes from Yahoo and Verizon kind of broke down the culture and made people less excited about being there.” It seemed the most constant element of working at Tumblr was the perpetually revolving door of hires.

Christian and Kahle attribute these departures to a toxic cocktail of frustrations: lack of upward movement for anyone director-level and above, shifts in leadership across acquisitions that made people feel professionally untethered, and the constant poaching of talent by companies with better compensation and more interesting projects. Tumblr was like a snake eating itself: It was common for employees to experience burnout after taking on the additional workloads of former teammates. 

When asked about the lack of upward movement, Mullenweg says, “Nothing else matters for Tumblr unless we get people using it more. If it’s growing, there’s lots of opportunities for upward movement and new roles [within the company].”


Tumblr was like a snake eating itself: It was common for employees to experience burnout after taking on the additional workloads of former teammates.

Amanda Brennan worked on Tumblr’s marketing team for seven years, and left her role as head of editorial in March 2021, 19 months into the Automattic acquisition. “I felt like I couldn’t fight anymore because no one would listen. I have been saying this until I’m blue in the face and I’m not seeing the feedback from leadership that I was hoping for… If I couldn’t move the needle, who could? What was next for me involved more investment than leadership at the time was willing to commit. And that [was] investment in Tumblr as a culture platform.”

Christian explains that, across acquisitions, “There was quite a bit of top-down direction…focusing on features and business. We didn’t spend a lot of time doing deep analysis of why people come to Tumblr and why they stay. It wasn’t a core part of our strategy. There was just a lack of understanding.”

Kahle agrees, “It really comes down to leadership not understanding the platform and its needs.” She believes D’Onofrio “really did care about the users.” While he tried to steer Tumblr as a business, he also worked to “maintain David Karp’s original vision” and nurtured a relationship with the founder. “When we were acquired by Automattic, Karp [visited] the office and was excited for the future, to move things forward and see what we could accomplish,” Kahle says, adding glumly, “I don’t think we accomplished much.” 

According to both Mullenweg and Kahle, Tumblr is roughly a 200-person organization. In 2021, 53 employees who had been with Tumblr pre-acquisition left the company. Mullenweg estimates that 75 percent of the current team have been hired post-acquisition, with an uptick in hires in just the past few months. “I hope it’s really clear that I’m incredibly excited about Tumblr,” he tells Mashable. “I care about it deeply. And that’s reflected throughout the entire current team.”

The point of no return

“Our users are very opinionated, vocal, and passionate,” explains Kahle. During her tenure, users simultaneously adored the platform while being resistant to changes that could save it, and “there were very few product updates that we rolled out that were not received poorly.” Though she sympathizes — “any time Instagram redesigns their homepage I’m annoyed, too” — Tumblr users wouldn’t let go. “A year later, we’d still get comments from users being like, ‘literally no one asked for this stop changing things’ or ‘we want porn back.'”

The porn ban of 2018 was a defining event for Tumblr that led to a 30 percent drop in traffic and a mass exodus of users that blindsided the company. 

“Tumblr has always been extremely vigilant about abusers online,” says Christian. A trust and safety team of 20 maintained a physical wall in the office where they’d pin the information of users they had reported to authorities. Kahle calls them “the best team at Tumblr. They work incredibly hard, really closely with the FBI… We had a great team that was also working with human rights groups on the Hill.” Both Christian and Kahle say that the teams’ work inspired similar approaches at Twitter, Facebook, and TikTok. Though, “to be fair,” Kahle notes “some of our employees left and went to those platforms.”

SEE ALSO:

The death of Tumblr porn left a void no other site can fill

In November 2018, when Apple removed Tumblr from its App Store citing the presence of child pornography on the app, Tumblr panicked. Christian had “created a proposal that would allow [Tumblr] to be viable with [Apple and] Verizon’s policies without having to do a complete ban on anything controversial.” Instead, the fix they moved forward with was “a huge sledge hammer on a problem that was kind of specific. It was larger than a porn ban, it was [a ban on] explicit content based on whatever the computer algorithm decides.”

Tumblr has historically struggled to monetize effectively, and Christian recalls that in addition to being a legal and public relations nightmare, porn on the platform was a financial death knell. “We were told that we can’t make money if we have any explicit content on our site, that specifically advertisers that will give us a substantial amount of money to keep things running will not come here.” 

It’s a head-scratching sentiment considering that, at the time of the ban, company leadership had failed to grasp what made the platform valuable for more than five years. Wooing advertisers is putting lipstick on a pig. The heart of the problem is “how do you foster a platform with a solid business perspective, that’s this place where people figure out who they are and pour out all of these deep, vulnerable emotions?” asks Brennan. “You’re not going to monetize it in the same way that you monetize [WordPress.com, which is] a blank canvas for people to build a website for their thing that they already know about. Tumblr is the exploration, Tumblr is where you do the deep digging, and that’s hard to monetize.” 

The Creatr economy 

The porn ban was a major blow to Tumblr, which had already been bleeding out in the battle against user churn. In 2014, at least 84 million posts were published every day. By 2018, that number had dropped to 30 million. The platform was especially hard for new users to navigate and, crucially, lacked a “stickiness” Christian says would have unlocked retention. When a new user posts, “it doesn’t show up on [other] people’s dashboards or explore pages. They’re not getting any sort of engagement,” Kahle explains. [That leads them] to churn more quickly, within their first 10 days.” Christian says, “I think we failed to support the user journey [and dig into] what it means to be happy on the platform, what it means to have a great experience when you join for the first time.”

Tumblr has also failed to modernize, even in the simplest ways. “There’s a lot of complexity to and a history of weird features — like asks and submissions — that haven’t been really updated in a long, long time.” For example, he points to Tumblr’s Answer Time, where fans submit questions for a celebrity to answer. The delay between submission and reply can be days, even weeks. Christian says making it live and interactive, and then potentially letting users host their own answer times, are no-brainer improvements.


Tumblr is the exploration, Tumblr is where you do the deep digging, and that’s hard to monetize.

– Amanda Brennan, former head of brand advocacy and editorial

The platform also needs to nurture its creators. Christian says research found that only a “small segment of people were highly engaged in posting and replying.” The majority of users engaged in “passive consumption.” That means that if the blogs you follow stop posting, your dashboard is dead. Having engaged creators is vital to Tumblr’s survival, but there are very few financial incentives for them, especially compared to YouTube’s industry-leading revenue split and Instagram’s native shopping tools.  

Post+, a recent first step toward paywalled content to the platform, was met with disdain and frustration. Although the feature was completely optional, users protested that monetizing fan-created works negated the fair use of copyrighted characters and stories and put creators at risk of legal action. Post+ was a beta test, but “I think that users thought it meant that the entire platform would eventually become sort of like subscription-based,” says Kahle.

SEE ALSO:

9 ways Tumblr changed the internet

That reaction highlighted a fundamental imbalance in expectations. “Across Tumblr there is still this widespread mentality that content should be free,” says Kahle. “My challenge as a marketer was trying to convince our users that our creators deserve to be recognized and compensated for the work that they’re putting out there. “The platform recently launched a tip jar feature that Kahle says “many employees really fought for,” but the platform still has strides to make towards satisfying creator needs for on-platform monetization.

Where does Tumblr go from here?

For Tumblr’s 15th anniversary, Mashable has been publishing a series of pieces about the platform’s legacy. “Every article is saying how Tumblr’s dead or dying and it’s not,” Mullenweg assures me, adding, “It’s kind of getting better than ever.” Then he reads me a positive comment from a user, and later sends me an email with 15 more. They convince me that the user experience on Tumblr is getting better, with fewer bugs, bots, and broken features. But to be better than ever, Tumblr would need to re-cultivate its singular charm.

I spoke to a handful of former users, all active from between 2010 and 2016. For them, Tumblr was a glorious archive of their growth and discovery. In its heyday, the platform was a haven for teenagers craving information and meaning, and people searching for themselves. 

Dimitra Zuccarelli says her Tumblr is a record of her search for her identity: “It’s basically just research [on who you are]. You’re picking and choosing what you like and forming this aesthetic. You could literally craft an identity every time you started a new blog.” Zuccarelli never went by her real name, and created blogs dedicated to the skater lifestyle aesthetic, fashion (“a lot of pictures of the Olsen twins”), street photography, Christmas content, and one solely filled with “really depressing GIFs.”

That discovery was almost entirely manual, made possible by the all-consuming, boundless energy of adolescence without any algorithmic design. “We are in this place in culture, where so many people expect algorithmic understanding,” says Brennan. Tumblr might be the place for you to go out and find yourself, but “TikTok basically comes to you,” says Zuccarelli, because the algorithm “knows everything about you.” To compete in the recommendation algorithm arena, Tumblr would need to ask its users for more information about who they are and what they like, something it historically has rebuked in favor of anonymity. 


The key is to look at Tumblr as a creative culture platform versus a tech platform.

– Amanda Brennan

At the time of publishing, Tumblr reports that there are 9.4 million daily posts on the platform, compared to 84 million in 2014. But with the rise of ‘90s and early 2000s nostalgia among Gen Z, a revival of the platform seems possible. In January, D’Onofrio said that half of the platform’s active users and 71 percent of its new users are Gen Z. Kahle who, at 25, is a zillennial, says attracting Gen Z and getting them to stay “may come down to the product.” For example, Kahle says “if Gen Z’s attention span is whatever percent shorter than millennials’ then the recommendation algorithm needs to be used to churn out content way faster.” 

But the core of the Tumblr experience, for any age group, is community.

According to Brennan, “the key is to look at Tumblr as a creative culture platform versus a tech platform. People don’t come to Tumblr because the features are groundbreaking.” Automattic can’t treat Tumblr like WordPress.com. “WordPress is a utility, it’s easy to use, it gets you up and running,” she notes. If Facebook is a phonebook where “you go to connect with people you like, find a handyman, pick an event,” and Twitter is a call-in radio show where “you hear all these different opinions, but you can shoot back,” then Tumblr is a high school cafeteria. Aesthetics, normies, weird art kids, goths, and musicians all have their place, Brennan says, “and you can be anything you want to be. Sometimes you don’t fit into one so you try another. There’s eavesdropping. You’re always going to learn something, no matter where you are.” 

“That is so much magic,” Brennan gushes. “Some of the stuff that happens on Tumblr is not happening anywhere else on the internet… It is never going to be for everyone, and I think that’s OK.” In the end, she says, “Tumblr thrives in spite of itself.”

Source link

Leave a Comment