Automattic chief executive and WordPress co-founder Matt Mullenweg has claimed the company’s purchase of Tumblr is his worst so far, a rare public acknowledgment from one of open source’s most enthusiastic advocates. Talking to the WordPress blogosphere at WordCamp Canada, Mullenweg stated that Tumblr is just too expensive to operate compared to what it generates, and that Tumblr’s backend will no longer be based on the WordPress infrastructure as planned. “I’m not giving up,” Mullenweg added, but he also characterized the huge marsh between Tumblr’s potential and the current situation: the platform is based on a totally separate tech stack, hosts more than 500 million blogs, and continues to draw a large portion of the assets that other Automattic products provide profitably.
Automattic purchased Tumblr from Verizon Media in 2019 in a deal that people reported to Axios as being for less than $3 million. On paper, acquiring an internet-favorite brand with a great cultural and devoted following seemed like a great deal.
In reality, the platform was plagued by structural issues:
- Poor monetization
- A hefty moderation presence that became a barrier more than anything else
- A codebase completely outside of Automattic’s field of expertise
The platform’s growth trajectory was derailed by a catastrophe in 2018, followed by an adult content ban. Third-party analytics companies like Similarweb and Sensor Tower showed a noticeable downturn in traffic and mobile installs. Despite regular resurgences and heavy audience participation, ad revenues are not so good for Tumblr compared to those of its better-established adversaries. Analysts at Insider Intelligence have said many times that social platforms seeking a more focused audience tend to have lower income values per customer compared to completely massified rivals.
Automattic experimented with multiple levers: subscription features (initially Post+, later tipping and supporter tools), ad partnerships, and community experiments. The company also redeployed talent and cut costs. Mullenweg publicly said in 2023 that Tumblr was losing tens of millions of dollars a year, which just drives home how hard it is to make the economics work for even a high-traffic social property.
The biggest promise was to move Tumblr over to WordPress infrastructure. Consolidating stacks would eliminate duplicative hosting and tooling, reduce engineering load, and make it possible for Tumblr to benefit from WordPress’s more mature moderation, publishing, and commerce systems. For a site with hundreds of millions of blogs and petabytes of media, however, migrating safely and without interrupting users is a monumental engineering project.
Mullenweg said the migration was a bigger lift and costlier than anticipated. And any migration needs to keep identity, posts, reblogs, notes, and the social graph—while keeping performance for a community that expects immediacy. The hiatus reflects a practical calculus: until Tumblr can staunch the bleeding of money, Automattic is focusing on work that pays invoices today across WordPress.com, Jetpack, and WooCommerce.
Federation and the open social web remain a long-term goal
One strategic “why” behind the migration to WordPress was federation. Automattic has been a vocal proponent of the IndieWeb/open social web; it sponsored development of the ActivityPub plugin for WordPress and is currently looking at ways to allow Tumblr to work with decentralized networks like Mastodon. The larger market is moving in that direction: Meta’s Threads has started testing ActivityPub, and independent networks like Bluesky and Mastodon have seen slow but steady growth.
If Tumblr were natively on WordPress’s stack, it could plug more easily into these protocols, sharing moderation tooling and allowing creators distribution beyond a single walled garden. That remains the most compelling long-term vision for the idea, but federation doesn’t solve for the near-term unit economics, and Automattic must balance its ideals with its sustainability. The most viable path likely involves a mix of targeted monetization and careful scope management.
- Focus on high-intent communities where commerce and patronage already work
- Expand lightweight creator tools that match Tumblr’s culture
- Judiciously prune features that are expensive to run but don’t drive revenue
There are relevant precedents. Reddit’s gradual shift toward paid API access and storefront monetization has yielded significant revenue—and a good deal of community backlash—and X has leaned heavily into subscriptions with mixed results. Tumblr’s home-court advantage is an energetic, creative user base that values expression more than virality. Strategies that reflect that ethos—transparent paid support for artists, community-funded features, opt-in ads that are relevant—are more likely to succeed than playbooks copied from other networks.
Mullenweg’s admission that Tumblr was his biggest mistake is a big deal. In an industry in which founders and executives almost never acknowledge their mistakes—it undermines their credibility and, often, their stock price—his decision to go on the record signals to his employees, creators, and open-source contributors that he recognizes the stakes, and that he means to make deliberate, if painful, decisions in service to their long-term health.
It also puts Automattic’s wider priorities into perspective. The company remains a steady and considerate investor in WordPress innovation, from the browser-native development of Playground to the development work spanning Jetpack and WooCommerce to communication products such as Beeper. Simultaneously, however, its very public claims about contribution and stewardship in the wider WordPress ecosystem suggest an emerging commitment to rewarding those who give to the commons instead of just taking from it.
The path forward for Tumblr remains steep but not doomed. Should Automattic be able to get the business into better shape, slash some of its technical debt, and pivot business models to revenue models that better match what the platform’s creator community could hope for, it might just get the open web outcome that many generations of Tumblr users have hoped for, finally. For now, it is a place of both confrontation and confirmation.