Storyboard

Peter Vidani on the Evolution of the Tumblr Dashboard

Ministry of Design senior minister Peter Vidani lays hands on every aspect of Tumblr’s visual and usage aesthetic. Initially contracted to work on theme creation, he came aboard full-time in 2009 and has steadily advanced his design philosophy of utility and simplicity. One of his primary, perpetual obsessions could be considered the real face of Tumblr: the Dashboard.

How did you feel about the Dashboard when you started working on it for Tumblr?

I was a Tumblr user for a couple of years before I started working here, and I loved watching every little change. Tumblr was the site I frequented most. When I started, we were on version 4 of the Dashboard. Unlike other social networks, the dashboard’s primary focus is on the user and their content. On other sites, there are navigational elements everywhere, a column full of stuff you never asked for — some relevant, some not — folders on the other side, chat windows, and so on. When you scroll down the Dashboard, you’re only looking at content you’ve asked for, from people you follow.

You once said, “If we can get rid of anything, we will.” How does that apply to the Dashboard?

We love short copy, not labeling things, not serving the fringe. We serve the majority.

Have you observed cases where new users seemed to be over-challenged by the interface?

Onboarding is really hard to get right. We want you to use the Dashboard in a certain way, but how we introduce and explain it, and how we guide you into using it correctly can be a challenge. We want you to follow really great people, to post something you love, and customize your theme. How do we do this without an installation wizard process or some other kind of hand-holding? If we can get that right for new users, it’s going to be right for people who have been using the site for years. If we can throw you into something so obvious you know immediately how to use it, we’re still serving the power users because the Dashboard will be smaller and have less clutter.

To a lot of people, Tumblr is a website-maker with beautiful themes. If you don’t want that, and you don’t know about the Dashboard, you’re going to miss out on what makes Tumblr really great.

Design has a very close relationship with the product team at Tumblr, which is not always the case with tech companies. How has that relationship evolved as the community has gotten so large and the technology side so complex?

It hasn’t changed much at all. I haven’t felt my job getting harder in proportion to growth. We’re still solving the same problems, still iterating the same ideas. In some ways it’s easier because we know immediately when we do something wrong, or when something’s not working. But the workload should ideally get smaller if the overall goal is to make the site simpler. We’re different from other companies in the way we’re set up. The design team sits in the company like engineering, community, business — we serve everyone. We’re like a little studio. We work on flyers, apps, internal tools, merchandise, marketing, all coming from the same people, hopefully with consistency.

That’s what’s changed, actually. Tumblr is doing more than making a website now. There are more mediums to work in.

How do you personally evaluate and internalize reactions to Dashboard design, from inside the Tumblr community or from the media, other designers, and so on?

Most of the feedback comes from everyone in the company. I hope that doesn’t change. I feel like even when we were five people, we all knew when something was right or wrong because we use it so much. We still get feedback from the Support team. If Support’s getting thousands of emails about a design or functional piece, we can react to that.

The advantage of this system is we’re making all the decisions ourselves — we’re recognizing the problems and solving them ourselves — so when something doesn’t work, we know exactly why. When you’re A/B testing or solving problems for other people, and you ask for someone’s opinion, you’re not going to get an honest answer. You’ll get an answer because you asked a question. Also, you’re not going to recognize why you’re fixing something if you didn’t yourself recognize that it was wrong. You’re solving someone else’s problem.

For example, there is no longer a follower count displayed on the Dashboard. We moved that to the user’s blog page for two reasons. First, we wanted that column on the Dashboard to only relate to things you subscribe to — who you follow, who you like, tags you’ve subscribed to. Second, we wanted to take the focus off follower counts. It can be an intimidating number, and something to obsess over, and ultimately a huge distraction from why you’re on Tumblr. A high follower count is not a good reason to share something, and posting something purely as follower-bait is not ideal. You should post something that you like, to attract the audience that’s kindest and most similar to you.

But when we moved the follower count to another page, it bothered a lot of people. Data would show that the number of visits to the page dropped off dramatically. Both of those facts would indicate that we should move the page back up front, but we made a conscious decision: We just don’t want to show the number so prominently.

How would you characterize the difference between the design philosophy for the Tumblr Dashboard versus conceptually similar user experiences elsewhere, like the Twitter homepage or the Facebook feed, or versus traditional blog-reading experiences?

We have fewer things you can click, and they’re bigger. We use 14-point font. Instead of just making a link clickable, they’re all blocks like you would see on a touch-device. When I look at other sites, they have a lot more things asking for your attention. As a result, everything’s smaller — small text, small icons.

Also our site is dark. Everything’s blue. If you look at it with a soft eye, the only thing that pops out is the post column. Posts are bright on that blue background and lifted up with shadows. I love that from a distance the site would look like it was only made up of posts. Another thing I love that we did recently was get rid of a highlight state in the right column. The selected tab used to be a big, green highlight, but now it’s a darker blue. We did that when we launched the new post icons. We should get rid of the Tumblr logo too. A table or a chair doesn’t have a logo on it. But when you drive a car, there’s a logo right on the steering wheel. Why is that?

Who is doing things in user experience design that excite you?

Apple, of course. And I like the new Windows Metro design interface. It’s just these colored boxes on black. It sort of makes Apple’s stuff look like children’s toys. I do love Apple’s design — the idea of skeuomorphism, when interfaces look like real-life objects — when it’s done well. Like when they put a big poofy shadow on the windows — and those shadows have gotten bigger and bigger on every release of OSX. A shadow sounds ridiculous because you’re trying to create the illusion that it’s floating off the screen. But really you’re just trying to distinguish it from the objects around it. In that sense, the shadow could be considered a border that fades out. Not a shadow but a gradient, which slowly divides the edge of that object from the things behind it.

I’m a big fan of old car dashboards, like Volkswagen’s Mark I Golf. I love seeing dashboards in old concept cars. Car dashboards are fascinating because they’re supposed to be usable instantly. And a lot of it needs to be usable without even looking at it. Turning on a blinker, using the radio. Checking speed, fuel, hitting the horn, even steering — all usable at a glance or less. You have hundred-year-old technology that makes sense to anyone as soon as they sit in a car. These dashboards deal with colors, they deal with touch, they deal with language, they deal with ergonomics. The result when it’s done really well — when someone can use it without being told how to use it — is really beautiful. And yet it doesn’t need to be beautiful because no one’s really looking at it.

Many current users may not be aware that a form of the Dashboard now appears when new users come to tumblr.com for the first time. What motivated that change?

Before, the Dashboard represented the side of Tumblr you only knew about when you were already inside. There was always a chance you signed up just because you knew someone that’s already on Tumblr, or because you wanted a website. But I sit on the Dashboard most of the day, so it’s hard to imagine Tumblr without it. By having a Dashboard visible before you even join, we give you a glimpse of what it’s like, which entices you to create your own version of it by following people. And we’ve got you.

How do you conceive of the Dashboard evolving over the next year?

It should get bigger. Our photos and videos are too small. We’ll still be stripping stuff away, still whittling posts down to something simpler. On a photo post for example — it’s a white box. First there’s a line of blue stuff, then a photo, then words, then maybe more blue stuff if it’s tagged. We could bleed the photo to the edges of the white box. We could hide things you don’t need immediately, like tags and source and username and post icons — the blue parts — until you hover over them. What you’re left with is an avatar, a photo without a border, and a caption (if that). If we can get rid of the right column, we can center these things on the page. That bothers me now — that once you get down the page, everything is off center. We can add more distinction to the post types, give them a bit more personality — so a link with a description doesn’t have to look the same as a text post with a title and body, a quote doesn’t have to look like a text post, links can stand out more.

— Chris Mohney (Photo by Joey Pfeifer)