Twistori’s Numbers

Twistori, as implemented, was a whim—I woke up that morning with the desire to ship something. I picked something I’d been thinking about a while, but had never committed pen to paper or pixels to Photoshop, and stripped it down to its bare essentials, something that could be achieved in a day.

The challenge was to find the tiniest atom of design that could stand on its own and do that atom up to the fullest.

I personally feel that we did a really good job.

While we were very satisfied with the finished project, and knew people would just love it, it has exceeded even our own expectations.

Before I write about the design premise and what led us to make Twistori the way it is (which I will do, but in another essay), I’d really, really like to talk about the numbers.

The “Launch,” as you call it

As you read the following, keep in mind that Thomas’ & my sole promotional efforts were two tweets.

Indulge me in my repetition for effect: We each tweeted once about it when we launched it. That was it.

We may have mentioned a bug fix here or there afterwards, but that was only once the traffic already had started rolling in like crazy.

NB: If you look at our Twitter accounts, you’ll note that we have over 1,000 followers between us and so that makes the “just two tweets” story less compelling. But that wasn’t the way it was before. Our followers are an effect, not a cause. Our numbers were more like a tenth of that before the launch of Twistori.

Pretty, pretty numbers

As of writing, we’ve had just under 150,000 visitors in 18 days. That’s good, but not terribly remarkable. However. Those visitors have performed a total of just under 970,000 actions (equivalent to “page views” for each emotion)—a mean average of 6.7 actions per visitor. That adds up to every single visitor checking out every single emotion (there are 6) at least once, and then almost one whole extra click, every single time (on average, of course). That’s unheard of.

But wait, there’s more.

The average (mean) visitor spends 25 minutes and 42 seconds viewing Twistori. That number’s from averaging visit length over all the statistics gathered so far, including all the notoriously picky digg traffic.[1] On the slow days—say, no front-page diggs or links on Business Week—the number creeps up to between 33 and 45 minutes. On average.

And, while digg did drive a fair amount of traffic to us, our power law distribution—aka long tail—is an incredibly long tail indeed. The percentage of digg referrers to all traffic is something like 25%. That means the other 75% comes from sources not populated entirely by 14-year-old boys and those who are psychologically indistinguishable from 14-year-old boys. In fact, over 31% of all our traffic has been direct / bookmark / untrackable sources such as IMs.

We’ve also been #1 “all time” on the Intwition leaderboard tracking links mentioned on Twitter (site down at the moment) since about a day or two after we “launched.”

Numbers tell us things

From all of this we can extrapolate:

• Lots of people find Twistori compelling enough to watch for minutes at a time (even Ritalined/Bawlsed out digg kids). And click all the emotions or (even more impressive) wait for them to cycle through automatically.

• Lots of people are directly telling their friends about Twistori, and people are coming back to it again and again.

• Much, much more of our traffic comes from individual’s personal blogs, and from mentions on design/art/psychology blogs/sites, than from digg, or even tech-centric media.

In short: there is really no interactivity in our design. But it’s stickier than flypaper coated in molasses and spilt Dr Pepper on a warm day, and just as viral as SXSWSars.

People also tell us things

We’ve also checked out just about every single referrer that’s shown up in our logs. We’ve got the range covered: from Seth Godin, to random kids’ Livejournals, to sites about information aesthetics, and modern art with a technological bent.

Even the type of person who would put autoplay Eric Clapton songs on their sites have blogged about us—and perhaps more surprisingly, actually sent us traffic.

Twistori’s appeal is pan-genre.

The takeaway

I don’t want to be lame and do your thinking for you, really I don’t. But I also want to spell it out, because Twistori is as much a proof for myself as it is anything else.

Here’s what I take away from this:

The way you design a thing has a massive impact on how appealing and engaging it is. And by “design” I mean the whole process, from idea conception to shipping. Not so much asking how do I solve this problem effectively?, but rather the 100,000 ft view of what do you mean, “problem”?

Technically speaking, aside from a couple visual effects, Twistori is no different than search results. The things that we chose to search for, the way we present them, and the aesthetic design (what most people refer to as “design”) is what has made it so damn compelling. And they were all very deliberate decisions made based on what we know about human beings. None of it was accidental.

And on this topic I will write some more. Perhaps once we breach 1,000,000 actions. I’d just like to see that happen.

[1] On the day of the number-one-spot-on-digg event, the average visit time was 5.5 minutes. If you’ve ever been on the trammeled end of a digging, you know how insane this is.

PS: This is basically unedited and reads kinda gross. My writing mojo is stopped up. I am too tired/sick to make it as snappy as I’d like. Hopefully the numbers speak for themselves, eh!