Tumblr’s great. It’s clean, simple, and the userbase is all about great content. There are some fundamentals missing that I haven’t seen addressed, though. Now might be a good time.
1. Too many things are hidden. A lot of the basic functionality is masked by 12px dog-ears and hidden syntax. Take for example, the question mark. In order for readers to reply to your post, you need to insert a question mark (?) in your post somewhere. Anywhere. I haven’t been able to uncover why this is the case but I’d really like to know if I’m missing out on an important piece of user psychology here. I did, eventually, come across the page with the hidden, “tips” but these should generally be more visible/discoverable on the site. Mainstream users shouldn’t need to go looking for basic social functionality. Only power-users go searching, remember syntax, and where the hidden page elements are. Normal users bail. This could be alleviated with something as basic as a pop-up tutorial or toggled overlay but, in any case, Tumblr shouldn’t be hiding such an important feature behind obscure actions. Save easter-egg features for the functions that most people aren’t looking for, like recovering old custom themes. Add a bullet or button to enable post replies.
2. The dashboard sucks. That’s harsh, but I’m not going back to re-word it because there’s a part of me that truly believes it. You can’t customize the dashboard without hacking together various Greasemonkey tricks. This isn’t going to pan out (though, I’m not in a position to say what Tumblr’s resources, critical path, or view are). You have to remember that a lot of us are coming from an age of overly-stylized pages, from Angelfire and Geocities to Xanga and Myspace. I don’t want to make my dashboard an eyesore but something beyond the repetitive blue and stout white letters would be great. I want a Death Star in my background dammit! I want the option to view my dashboard in the same style as my blog. Those are my preferences, that I took the time to set up, and I believe they should be carried through the entire Tumblr experience. Simply, though, I want to change things up every now and then. Maybe this restriction has to do with the promoted post placement and how well it monetizes but, again, that should be the last concern (and one I’ll address lastly). Start with the user experience. End with…the user experience. Rinse. Repeat. From the simplicity of the view and the genuine concern with user feedback, I truly believe you’re going for a clean experience. My sentiment, though, is that clean and robust aren’t mutually exclusive. Medium is consistent: the entire interface is uniform. Tumblr doesn’t have this advantage so some people don’t want clean. Let them be dirty—maybe they’ll pay for it.
3. The promoted posts aren’t actionable. It’s no secret that Tumblr is a heavily visual medium—and I’m not just talking about porn. The promoted posts are a perfect example of this. You won’t see a promoted post that’s textually based, which means most of the promoted content is art, infographics, pictures, etc. That’s great and all (for me, because I happen to love art) but, considering I can admire the picture from the thumbnail, I see no reason to click it.
I also want to see other types of content. Why not have promoted posts that are comprised of great text, audio, or video? If the signal from the promoted thumbnail is meant to be, “this content is captivating,” then I’d probably be more likely to click it, regardless of the thumbnail. It could be an alluring title to a text post. It could be a quote that embodies the ethos of the post. Hell, it could even just be the number of notes that a post got. I’d at least have a relative idea of how compelling that click is to me. “3000 notes? That must be something juicy because the max I’ve ever seen has been 1500.”
Sponsored posts between Tumblr users, transacted through Tumblr for a nominal percentage. Users get the added benefit of related content. Publishers can just decline if they don’t want to advertise. Just a thought :)