NewsBlur front page
Google Reader has been in the news of late after its recent changes, which have had a very sharply negative reaction from passionate fans. I had earlier written about how there are no good Google Reader alternatives in the market today and had mentioned NewsBlur then. Since then, I have been using NewsBlur daily, and have been extremely pleased with it. I got in touch with the developer, Samuel Clay, and he gladly made himself available to discuss life as an indie developer, developing NewsBlur, and how he plans on competing with Google.
Samuel Clay is an indie developer of NewsBlur. He just moved from Brooklyn to San Francisco. Previously, he worked at DocumentCloud, where they wrote Backbone.js, VisualSearch.js, Underscore.js, and many other open-source libraries. He is now at Tasty Labs, making a more useful social application on the web. Samuel can be reached on twitter at @samuelclay and NewsBlur is also on twitter at @newsblur.
Techie Buzz (TB): What is NewsBlur?
Samuel Clay (SC): NewsBlur is a feed reader with intelligence. It tries to do two things very well:
- Shows you the original site instead of a context-less feed. Read the original and NewsBlur marks the stories you’ve read as read.
- Filter stories you either like or dislike. A three-stop slider goes between dislike, neutral, and like (red, yellow, and green). Training is super-easy and all click-based (as opposed to you having to writing out what you like in a site, NewsBlur asks you, semi-Hunch-style, your opinions on facets of the site).
I started working on NewsBlur to see if I could do it, put the AI together with the back-end feed processing and fetching, along with the nifty front-end of the original site. This is one of those projects where I just kept pushing in all directions until I felt I had something good, not knowing if I could do it at all, but believing the entire time that I was able to complete the project.
TB: Google Reader is not only the default for RSS reading on the web, but it is also built by the mighty Google with virtually unlimited resources. Why go up against them?
SC: I’m lucky that a competitor like Google Reader exists. It means that new users on NewsBlur already have a place to start and can import over OAuth to have a feed reading experience ready in only a few seconds. Also, Reader has done a great job of educating users about what a feed reader is good for. Without Google, NewsBlur would have a lot more difficulty gaining traction.
The truth is that I wanted something better. I never liked using Google Reader (I was a NetNewsWire fan). It didn’t jive with how I read feeds. I don’t follow hundreds and hundreds of sites. I follow a few dozen writers and want to read everything they write. I also follow a few more multi-author category blogs — New York, Tech, Cooking, Art, Photography, Tumblrs, etc. I knew that I wanted to read some of those articles, ignore others, and make it easy to decide between the two. Google Reader has no easy way of making that happen.
TB: I understand that you have open sourced your code. Why did you do that?
SC: Open-source doesn’t have to mean non-profit. I make money with premium accounts — accounts are free up until 64 feeds. Also, feeds are updated more often for premium users (but only for feeds that they are the sole subscriber, otherwise they get the collective benefit of multiple subscribers).
Sure, folks can host their own copy of NewsBlur and keep it up to date but it’s a pain, and I have 7 servers happily chugging away fetching, parsing, storing, and retrieving feeds for you. Hosting is sure worth a lot.
The main idea behind open-sourcing the code is that a community can develop around the API, new features, and user-contributed code. I’ve had a number of pull requests and issues found in the code thanks to NewsBlur’s many talented developer users.
TB: Is NewsBlur an app or a platform? Why?
SC: It’s both. The app is at www.newsblur.com, and it’s freely hosted for anybody to use. The platform is the NewsBlur API (http://www.newsblur.com/api) and it’s where the future lies. There are already a dozen services that use the API (the Android app and iPhone app being the most notable, but there are also bookmarklets and other one-off tools that folks have written). There’s so much possibility in the API. It’s also part of the NYC BigApps Hackathon, where developers will use NewsBlur’s API as part of a larger effort to make something novel in the service of the city of New York.
TB: Part of Google Reader’s appeal is the multitude of clients available on virtually every OS/device. How are you planning to (or have) combat that?
SC: I’m beginning to cover all my bases. The iPhone app will soon be available (and if you search for it, you can download a beta from the App Store). The Android app is out there, being used by dozens of users. I’ve heard of interested third-party developers on WP7. I’m there to support all of this happening.
Developers can even charge for their app, so they can ride the wave of traffic that comes with NewsBlur’s quickening growth.
TB: Can you talk about the number of users you have?
SC: Sure. I launched exactly a year ago and now have 25,000 users, nearly 500 of which are premium subscribers. Both of those numbers are growing faster each month.
TB: How is the freemium model working out? Would you do it any other way?
SC: Other than fiddle with prices, freemium is clearly the way to go. It’s odd though, since 90% of premium subscribers signed up for premium membership within 24 hours of creating their accounts. The model seems to be working well, but I am very tempted to use the Pinboard approach — one-time entrance fee, but Pinboard also has 18,000/25,000 active users. If I had an entrance fee, I would hope that my actives would closely trail my total count. It’s less likely, since Google Reader is so deeply entrenched, and Delicious is in a world of hurt.
TB: How many developers contribute to NewsBlur?
SC: I’m the only developer, designer, administrator, and marketer. A few folks have kindly contributed code — a Python API library, a few bookmarklets, and the Android app. I fully support anybody who wants to contribute, since they can either charge for their own product, or piggy-back on NewsBlur’s success.
TB: How long was NewsBlur in development before you launched it?
SC: About a year. I am now exactly two years into the five year plan.
TB: As an “indie” developer, what are some tips you can offer others on how to launch (getting the word out, following up, traffic spikes, etc.)?
SC: I never advertised and the only word I spread was at meetups, when explaining what I do with all of my free time. A post on Hacker News a year ago was enough to generate a massive press cycle that brought me into this year and coder to the public’s consciousness.
As for handling traffic spikes, I have moved from hosting provider to hosting provider, starting with a $20/month server at Slicehost, expanding to 3 servers (app, db, and task), moving to Linode and working up to $80/month for 3-4 servers there, and finally ending up with bare metal servers that I rent out for $420/month. This will give me a few months of leeway before I end up having to shard my non-relational data store (in MongoDB) and have to pay upwards of $800/month. User accounts are more than covering that cost.
I would highly recommend staying off Amazon’s servers, since they are older machines, littered with network and I/O contention, unexplained slow-downs, occasional outages, and a huge cost. When I need to provision a new server, it takes a few hours, but I never let my servers run over 10% capacity (save for my task servers, which are workhorses that could use some love right about now).
I also treat customer service as the highest priority ticket. I write down all of monthly priorities and then cross them off the list as I accomplish tasks. However, emails and fixing bugs for users are always top of the list.
TB: Finally, any thoughts on what’s in store for NewsBlur in 2012 and beyond?
SC: Search, launching the iPhone app with training and a River of News, and then building the many social features that I’ve had in mind since the beginning. There will undoubtedly be a major spike in traffic due to the network effect, so I’m battening down the hatches and making sure the foundation is strong before I attempt to build in the most complicated features I’ve ever had the pleasure of designing.