It was only after it sold that Yahoo realised how many of them were using it for porn, which in retrospect should have been obvious: it's an image hosting site.
No, the percentage of porn content was actually relatively small at the time of the sale. Low to mid single digits if my memory is correct.
During Tumblr's peak heyday (~9 months before the sale iirc), Tumblr's daily posting volume was around 1/5th the size of Twitter's at the time. Keeping in mind that Tumblr had no character limit and much greater multimedia support, the amount of content being posted on Tumblr was quite significant.
It was a widely-used general-purpose social network / social publishing platform. Most major brands had a presence on Tumblr, and for a while it was extremely common to see a Tumblr icon on their sites right next to Facebook and Twitter icons.
From my POV as an early engineer there, what went wrong was poor execution in response to many simultaneous existential threats:
* Instagram succeeded at peeling off all the "just want to post photos on mobile" users, and also became the hit new thing with the teenage users
* Pinterest succeeded at peeling off all the "just want to digital scrapbook" users
* Although the porn percentage was relatively low at the time, it still scared advertisers
* Difficulty fundraising / running low on money at particularly bad timing. Facebook's stock performed poorly for its first 15 months post-IPO, and this temporarily made fundraising difficult for existing social platforms... especially when growth was relatively flat (due to the first two bullets) and revenue was disappointing (due to third bullet).
* Ever-growing tech debt. The early team was extremely solid, but very small relative to the amount of traffic, so we often had to creatively cut some corners. Later on, management didn't want to allocate staffing to fix existing things, and a lot of time was wasted on unrealistic big-bang rewrites that were abandoned. Meanwhile the hiring bar kept being lowered over time once the company was no longer a rocket ship.
I left just before the Yahoo sale, as working at Tumblr had gone from my dream job to a stagnant grind during my tenure there. I later returned for another year in the late-stage Yahoo into early Verizon days, with idealistic hopes of helping to prevent a Vine-like fate, but let's just say that was a poor personal decision and leave it at that :)
The thing that drove me crazy about Tumblr is how you would find a post with some interesting content and scroll down to see if there was any more information about it.
The comment section would be thousands of lines long, but every line was something like:
User blahblah reblogged this!
User whatever likes this post.
User someone reblogged this!
You can't hold a discussion when your own system spams the comment section with irrelevant crap like this.
There was also the meme of the "shitty Tumblr gif" thanks to their fairly low image size limits.
If I understand correctly, were you just consuming Tumblr content directly from individual blogs?
The user experience was generally better in the dashboard (activity stream / feed for logged-in users). The dashboard was the vast majority of traffic; blog UI didn't get much staff/engineering attention as a result.
The "notes" feature on Tumblr was absolutely never intended to be a "comment section", btw. There intentionally was no comment section at all originally. This notion was part of the conceptual difference between tumblelogging and traditional long-form blogging.
Contextually, this mechanism came about at a time when comments (on blog posts, YouTube, etc) were complete cesspools of negativity from anonymous jerks. The Tumblr idea was that to respond to something, you would have to reblog it in order to add commentary. This way, if someone wants to spew nasty comments, their blog would be terrible and no one would choose to follow their blog.
No, again, it was not meant to be a traditional comment section. It was not a discussion thread for casual readers, especially via the blog view. If you wanted a normal comment system, you could use Disqus embeds, or just use a traditional blogging software.
The reblog system was a way for Tumblr users to publicly respond to things, and allow that response to be visible to the responder's followers. The original poster would also get a notification, and they could reblog the responder's reblog to add more if they wish.
The dashboard made this slightly easier to read. Again, if you only ever experienced Tumblr on blog views and without an account, you literally never saw the core functionality that made Tumblr popular.
Also, most Tumblr posts weren't long-form text to begin with. That wasn't the point of the platform.
That explains it then. I never felt the need to make an account since I was just reading. They didn't make it clear that the service was incomplete without an account.
Well, the alternative is an aggressive full-page signup nag that intentionally breaks product functionality when logged out -- an approach used to various degrees by Facebook, Instagram, LinkedIn, Twitter, Reddit, and others. Is that really preferable?
It was only after it sold that Yahoo realised how many of them were using it for porn, which in retrospect should have been obvious: it's an image hosting site.