Results tagged “streams”

All Dashboards Should be Feeds

January 8, 2013

Last week, we announced the new beta release of ThinkUp (if you're a geek or developer, try it out!) and one of the reasons I was most excited to talk about the new release is because it has a whole new user experience which exemplifies a belief about analytics that I've become pretty adamant about.

Every time an app provides a dashboard full of charts and graphs, it should be replaced with a news feed offering a stream of insights instead.

There are lots and lots of apps that provide a dashboard of analytics these days. From Google Analytics to Chartbeat to Facebook's Insights tool, there are all kinds of dashboard displays that we end up staring at while trying to manage a presence online. And they all share a consistent problem: It's hard to tell what the hell is going on.

chartbeat-charts

Beautiful but Empty

When I talk about "dashboards" here, I don't mean ones that are already a news feed or stream of posts, like Tumblr, but the old-fashioned kind where you see a bunch of meters and dials and line charts that are supposed to communicate the current status of whatever you're tracking. You can see Chartbeat's beautiful set of graphs and charts above, and Facebook's got a fancy one for insights on their platform; It looks like this:

Facebook InsightsCharts

And you know what? I have no idea whether my numbers on those services are good or not. I don't know what I'm supposed to do about them. In fact, though I love Chartbeat, the information that I get from them that means the most is their push notifications on my phone which tell me when my site is over its maximum monthly number of visitors. That is meaningful.

Insights like exceeding my usual level of visitors, or achieving some threshold I'd never crossed before, or doing some task particularly efficiently would be meaningful markers that I could respond to intelligently.

Worse, trying to make sense of a gauge on a dashboard essentially requires me to keep three bits of data in my mind:

  • What metric or measure a particular meter is reporting
  • The last time I looked at that meter
  • What the value of the meter was the last time I looked at it

That's a lot! For more esoteric points, it's downright impossible, so I'm left squinting at a little chart, trying to deduce its meaning. Or, on the other extreme, I get something like the line chart that shows my number of Twitter followers. That line only ever goes up and to the right. Sometimes it goes up more than others, but even that's generally impossible to discern.

A Better Meter

We had precisely these issues with ThinkUp in version 1. Lots of little line graphs and pie charts that either rarely changed, or that changed regularly but with no explanation of the meaning of those changes or recommendations of what to do as a result.

And just getting to that stage was hard! The community put a ton of effort into collecting useful data, and presenting it appropriately. But all of that hard work still left an average user of the app squinting at some inscrutable charts, ultimately unsatisfied.

So we killed the whole dashboard. And replaced it with a simple stream. Here's a live example of the White House's social data, from a very early development version of ThinkUp 2.0. Now, not all of the data here are presented in a very compelling way yet, and of course we're still working to shed our old implementation of basic charts and graphs to move into a stream that does much more to coach you about what you should be doing online.

Take the idea of your follower count. ThinkUp used to offer a pretty standard inscrutable little line chart showing your number of followers going up over time. But now, there's a stream with a couple of items in it that look like this (these aren't the final UI, just a work in progress):

White House Upcoming Milestone

That offers a bit more analysis, showing a forward-looking extrapolation of when the @whitehouse account will reach a certain number of followers. If someone sees that as a useful goal, they now have much more info than they would have had from a chart showing their past history of growth.

Then we can break down that data even further and tease out the meaning by determining which of these followers are notable for being popular or discerning enough that they should be called out or paid attention to. That looks like this:

White House Interesting Users

And this isn't to say that traditional charts or graphs don't have a place in communicating information in this stream. But what we've done is put them behind disclosure buttons (again these are just a first prototype of the UI for such a thing) and made it possible to reveal the details behind an item, whether that's a detailed chart or just a full list of people to pay attention to.

whitehouse-upcoming-milestone-expanded.png

Similarly, you can see an expanded version of the "interesting new followers" insight in the ThinkUp demo for the White House as well.

Even with just a rough version of our new stream built out, I immediately realized that this was a fundamentally better way to quickly consume this analytics data and be able to make decisions or act on it. There were also many other benefits to radically simplifying the user interface — I only see data now when it changes in a significant way, so I don't have to go digging around a bunch of different screens trying to deduce if something has changed and whether the change is meaningful.

This has really quickly ruined me for every other stats app that I use. Chartbeat is awesome for being real-time, but it'd be so much more compelling if it were a real-time stream or Twitter-style feed of information about how my site and my content's doing, with the ability to drill down into individual insights about my site. Google Analytics has always been totally inscrutable to me, but if it just bubbled up particularly meaningful tidbits about my site or its stats I can imagine being able to actually make educated decisions about what I do here. As it stands, I haven't had any reason to go into Google Analytics in ages.

So, a big but sincere request to everybody who's making analytics or stats apps, either standalone or as part of a larger app: Please throw away the dashboard. I know they demo well and look great in investor pitch decks or screencast videos. But they don't actually help me make decisions, or get better at what I'm doing. And that's the only reason I'm measuring something in the first place.

Related Reading

If you liked this (or hated it), you should also read Stop Publishing Web Pages, about the move from static web pages to streams that are part of app-style experiences. It's in the same vein.

More on Streams vs. Pages

August 15, 2012

Lots of nice responses to my plea to stop publishing web pages yesterday! Here's some highlights:

  • Dave Winer said I should credit him for a lot of the ideas in the post. I'm eager to grant his point, so please do know: Dave articulated a lot of the ideas like the river of news and other concepts that I mentioned well before anyone else. I care more about the ideas succeeding than whose name is attached to them, but I never want to deny people the credit they desire.

(For reference, it's easier for me to give you credit if you don't block me on Twitter while simultaneously insulting my friends.)

  • Hacker News had a skeptical but overall pretty thoughtful conversation about the piece. I really liked it!
  • Lots of people had more questions about advertising; I deliberately didn't try to be comprehensive in the piece, but there's certainly a big reckoning coming with regards to the transition away from old display ads to in-stream ads. The key transition here is not a technical one, but in getting the business-side infrastructure of ad sales to move to new units, new metrics and new sales processes, as Om Malik pointed out.
  • What about permalinks?! I love permalinks, and the more geeky sort were worried about what happens to permalinks in an all-stream world. The way I see it, there are two possible evolutions to the permalink: They can either be a link to a particular point of information within a stream, or (more interestingly) they can be a link to a representation of how a stream looked at a particular moment in the past. I'll be excited to see how those things evolve.

Now one beloved troll, I mean, VISIONARY (totally same difference, no?), is calling for the end of web pages. This is an appealing notion! "Most users on the web spend most of their time in apps," begins our pal Anil Dash. (He promised a citation on that stat later. I'm sure it's true, if you count Facebook!) And: "Most media companies on the web spend all of their effort putting content into content management systems which publish pages." Anil's sort of right, but he's also boosting an idea about business who act like—and design like—they have no interest whatsoever in being businesses. Producing a "feed" subsumed in the apps of our time is not a business. It might (SORRY) own the means of its production but it won't own the means of its revenue.

I disagree! But you should read it for yourself. I'll be keeping my eyes peeled for more examples of media sites that are transitioning from pages to streams; If you find good ones, let me know!

Stop Publishing Web Pages

August 14, 2012

Most users on the web spend most of their time in apps. The most popular of those apps, like Facebook, Twitter, Gmail, Tumblr and others, are primarily focused on a single, simple stream that offers a river of news which users can easily scroll through, skim over, and click on to read in more depth.

Most media companies on the web spend all of their effort putting content into content management systems which publish pages. These pages work essentially the same way that pages have worked since the beginning of the web, with a single article or post living at a particular address, and then tons of navigation and cruft (and, usually, advertisements) surrounding that article.

Users have decided they want streams, but most media companies are insisting on publishing more and more pages. And the systems which publish the web are designed to keep making pages, not to make customized streams.

It's time to stop publishing web pages.

But I'm Reading This On A Web Page Right Now!

Obviously, I've written this in an old-style content publishing system, and this piece lives on my website as an old-fashioned HTML page. But if I had my preference, I'd write up an article like this, and it'd seamlessly glide into a clean, simple stream of my writing, organized by topic and sorted with the newest stuff on top. Blogs have always worked this way, but they were shoehorning this stream-like behavior into the best representation possible under the old page model.

I don't have a tool I can use to run my website which will output a stream that works the right way. "What about using Tumblr to publish your blog?" you ask. Well, besides the fact that my site would have to run on their infrastructure, individual tumblr-style blogs don't allow you as a reader to personalize or customize the types of content in the stream, the way you would be choosing people to follow on Tumblr, Facebook or Twitter. You can't choose to follow just the music-related posts on my blog, ignoring the ones about technology.

This isn't just about how the content looks, it's also about how it works. The smart, responsive, dynamic apps most of us use on the web everyday have all kinds of subtle but powerful bits of functionality which appear as we hover or click on items in a stream. Meanwhile, our pages are still piling a row of awkward-looking share-button cruft at the bottom.

Also: Dollars

The vast majority of advertising online is dependent on a page-view model that users have overwhelmingly decided to abandon. Facebook, Twitter, Tumblr and others will succeed by making in-stream advertisements that fit in with the native content of their networks. Meanwhile, page-based sites are cramming every corner and bit of white space on their sites with ads that only ever decrease in effectiveness until they are made even larger and more intrusive every few years.

Stream-based content naturally flows across different devices and media, from tiny phones to tablets to giant desktop monitors, with an adaptivity that works naturally hand-in-hand with responsive design. Page based ads basically have to be reimagined on each platform, and fundamentally don't work in mobile form factors.

Streams of content can easily be read in friendly native apps on mobile platforms with the content flowing through simple APIs. Pages get squeezed into faux-mobile app experiences that look just enough like native apps to be frustrating and annoying when they don't perform correctly. Pages tell users there's no mobile version of this story available, or accidentally redirect an interested user to the site's homepage, from where they quickly depart. Pages stop your flow.

Pages vs. Streams

Let's Fix This

So: Start publishing streams. Start moving your content management system towards a future where it outputs content to simple APIs, which are consumed by stream-based apps that are either HTML5 in the browser and/or native clients on mobile devices. Insert your advertising into those streams using the same formats and considerations that you use for your own content. Trust your readers to know how to scroll down and skim across a simple stream, since that's what they're already doing all day on the web. Give them the chance to customize those streams to include (or exclude!) just the content they want.

Pay attention to the fact that all the links you click on Twitter, on Facebook, on Pinterest, all take you to out of the simple flow of those apps and into a jarring, cluttered experience where the most appealing option is the back button. Stop being one of those dead-end experiences and start being more like what users have repeatedly demonstrated they prefer.

And if you're smugly thinking "oh, we're an app — he's only talking about publishing content, so we don't have to pay attention", then you should get to work, too. Except for power tools which need to make use of the screen in a particular way, most of our other apps are going to be rearranged into streams, too.

Further Reading

  • From ten years ago, Stories and Tools (Michael Sippey, now Twitter's head of consumer product, liked this piece so much back then that he republished it.)
  • At Activate, we created a presentation called "What Matters" at the end of last year; It starts by offering some data about use of page-based sites vs. stream-based sites by web users.
1