Digg Exec: Sorry About the Bugs, But Glad You Care!

Digg’s had a hectic week. The company’s long-awaited relaunch didn’t exactly go as planned; it was fraught with technical problems and poorly received by many users. The Digg team has been working like crazy to fix the site, but Digg VP of Product Management Keval Desai finally got a chance to breathe today and talk to GigaOM. He was able to put a positive spin on recent events, telling us that while the Digg team isn’t proud of the problems with the V4 launch, it is excited that the new platform allows the company to iterate quickly. He said V4 has already been updated with new code and features more than 100 times in the last week, in comparison to the past where Digg would languish for months without change.

Here’s a lightly edited transcript of our conversation:

GigaOM: V4 had been in development since before you joined the company nine months ago and you’d invited tons of beta testers. So why wasn’t it ready for primetime?

Keval Desai: You’re right, it had been around for some time — over a year. A critique was that Digg was slow in launching things, so we wanted to get to a platform that we could launch quickly on. So we went from a SQL database to NoSQL, and did a lot of other things in the stack using open source and other technologies.

This is a massive platform that serves millions of people, where if I’m following you, and you Digg a story, it immediately pops up for me, and if I have 10 followers it goes to them as well, and it’s an instant nuclear reaction. That’s a feature that I don’t think most Internet companies are trying to attempt. I think we’re trying to attempt something very innovative and it carries some technical risk.

As you know, we ran the product in beta for almost a couple of months, and it was running and in parallel to our existing platform. But the beta product was invitation-only, so we didn’t have full load until launch. At launch, we saw higher traffic than on Digg.com on the previous version — from renewed interest or people who had never heard of us who came for the first time. The second thing that happened is that the entire technical stack, some of which was fairly new, finally felt the brunt of the traffic, and saw some stress points that we didn’t see in our load-testing. But every day, the site gets more and more stable. We still have stability issues that I do want to acknowledge, but I think we are reacting. Since we launched, we’ve had more than 100 incremental launches — bug fixes, stability patches — and that speaks to the capability of the new platform.

GigaOM: Do you think Digg users would have revolted against anything you launched?

Desai: I wouldn’t call what happened a revolt. I think the worst thing that could have happened is that we launched a product and nobody noticed. I think what we’re from hearing users and power users is good feedback. On balance, they understand that Digg needs to have a scalable platform, it needs diversity of content, and people want a personalized view of news.

It’s a human tendency to react to something that’s new and that’s fine. What’s happening right now is par for the course except for the stability aspect. A lot of sites launch features every week, and with Digg we hadn’t launched in a while. This will become an incremental revolution.

GigaOM: What’s the distribution on Digg between readers, power users, and new users? How does your product team think about designing for such different experiences?

Desai: With any community site, whether it’s Wikipedia, Digg or Facebook, I think the typical distribution is you have a lot many more readers than writers. On a high level that’s clearly the case with Digg as well. But one interesting thing with V4 is the number of daily registered users has exploded. We’re seeing very high engagement in terms of people logging into the site, so that’s good. We think that bond of knowing who’re you’re following and who’s following you will get people to be more active.

In terms of who we’re designing this for, we want this to be a platform for users with a variety of interests, politics, sports, technology, lifestyle. Readers of all ages, of all nationalities. More than half of our traffic comes from outside the United States. Imagine if we can localize this product; we clearly have latent demand. The vision is to provide a news platform that serves a variety of needs as daily habit. The percentage of repeat users in the last week for V4 is quite high, much higher than we were seeing in V3. I don’t want to take anything for granted, but I think the early findings are quite clear.

GigaOM: Some changes you’ve made aren’t bug fixes or stability issues, but rather, bringing back features from V3 that you’d decided to take out. How do you make those decisions about relenting to user demands for old features (even though there may be more nostalgia around them than actual use — for instance, Kevin Rose said the Upcoming page had only 0.4% of views)?

Desai: In both the cases so far, Upcoming and giving users the choice to make top news default, these were things we wanted to add — but not on Day 1. What we wanted to do was get user reaction and react to it, rather than allowing feature creep to delay the launch of the platform.

GigaOM: Your bio on your Digg profile says “my job is to let the inmates run the asylum here and get out of their way.” What do you mean by that?

Desai: I think the philosophy behind that is when you’re at a job as a manager especially at entrepreneurial companies you need to be a coach rather than a micromanager. I want to hire the best and most brilliant people, then let them do what the came here to do: think of innovative ideas for millions of users, take some risks, launch, get feedback and react.

GigaOM: So the “inmates” are your coworkers, not the users?

Desai: Yes!

See also: New Digg CEO Must Grab the Reins and Go