Like many popular apps, Instapaper started its life as a personal solution for Marco Arment. Then it took off as customers realized how great it is. Now, Instapaper faces the challenge of Lion’s Reading List. Mr. Arment sizes up the new competition with our Dave Hamilton during WWDC week.
Dave Hamilton: We’re here at WWDC with Marco Arment. As I recall, you’ve had a very interesting history of developing for Apple. Is Instapaper your first real project?
Marco Arment: Yeah, it really is. After playing with it one night at a time for a few years, it’s my first Apple environment product.
DH: Of course, it’s both a web app as well as an iOS app. Which came first?
MA: The web app came first because I launched it in the fall of 2007 — a few months after the first iPhone came out. When I got one, I quickly realized that I would love to read on the train with this thing — I had a train commute every day — but the screen was really small, it wasn’t that fast at downloading things, and on a train commute you sometimes go under ground and don’t always have good reception. So what I needed was really something I could toss links into during the day, sort of a temporary link bucket, so I’d have stuff in there for the ride home.
So I really wanted something to browse RSS, but back then there were no [iPhone] apps. So one night, I started putting the Web service together and started using it. I didn’t tell anybody about it — I just used it myself for about two months. And then, I told some friends. They loved it. And then I posted it on my blog. And it just exploded. That was the fall of 2007. In July of 2008 is when the App Store launched. I wanted to be in the App Store on day one. I ended making it on day two because they had far more submissions, even by the deadline, than they expected. So even if you submitted by the deadline, you weren’t there on day one necessarily.
I wanted the app to do the same thing as I was doing with the website, but do a few things better. So, from day one it had offline access. That was huge. And the text view, of course. Those were there on day one. Because there were so many times on my commute where I’d go underground and I’d have multiple Safari windows open, but they would be kicked out of memory, and I had to reload — you just can’t do it.
DH: Right. Very frustrating. Yes.
MA: Exactly. Back in the early days, there was no HTML 5 offline support. There was no other good way to get content offline. So the app was just a huge help to the functionality of the service.
DH: Very cool. Since then, how has development progressed? Is there any service that Apple has come up with where you said, I need to bake this is in?
MA: Tons of stuff. Stuff this week that I can’t wait to use. But in the past, there was a lot of UIKit stuff. Instapaper has always been a UIKit app. I don’t do a lot of manual interface creation. Or custom skinning or anything like that. I try to use what’s there in a way that Apple would use it.
I’ve always thought of Instapaper as an attempt to mix Mail and Safari. And that’s kinda the function of it. You make yourself a box full of web pages that you want to read at some point. On the iPad I try to meld it better with iPad design guidelines — where you want to be a little bit more differentiated. But on the iPhone, my app looks a lot like Mail. And then you launch the web page, and it looks a lot like Safari.
DH: Right. That’s interesting. Okay. So not reinventing the wheel, just using it in its own way.
DH: Well, that’s smart. It saves development time for you. And futureproofs you against something you had to do manually.
MA: Sure. And it isn’t the coolest design in the world, by any means. But, everyone knows how to use it as soon as they get it. And so I end up having very few support e-mails related to usability.
DH: Now, are you doing anything on the server side with processing or caching? Or is that all done in the app?
MA: The server does all of the text conversion. The great thing about that is that I can adjust it dynamically, I can adapt, I can do a lot more analysis on what’s generated and where I need to improve it. It works great that way. I love how the server does the heavy lifting. Plus, even when I had the app, the iPhone was slow. The CPUs back then were not that good, and the text output is smaller than the HTML page input, so by doing the conversion on the server, I’m saving tons of bandwidth — over, at that time, the slow EDGE network.
DH: I’m assuming that, at some level, you’re caching the content on your server. You’ve got to go out, re-check, semi-regularly, to make sure a website hans’t be updated. Or something like that.
MA: Sure. In reality, I don’t really have a huge amount of repeat page requests — among all requests. People save a lot of different content. So I actually have relatively few cache hits.
DH: That’s fascinating. I guess you app isn’t about a certain type of reader … it’s about a person with a certain type of lifestyle. And that spans all content interests.
MA: Correct. And you might assume that it’s just the big magazines and newspapers, but in reality it goes way beyond that. There’s a stat I love to quote. I can look at the most frequent hostnames, and usually the top 10 are a few of the big newspapers and magazines and some popular blogs like Lifehacker and TechCrunch. What’s interesting is that out of those top 10 sources, collectively they only represent about 10 or 11 percent of all the saved articles. Which means there’s a huge long tail there. And that’s where most of the activity is happening.
DH: Okay, so obviously this week there was a but of a shakeup during the keynote. Apple announced its “Reading List.” And you’ve actually spoken publicly in your blog about your feelings on this. So. if I may paraphrase, it sounds like you’re actually kind of positive.
MA: I would say that I’m cautiously optimistic. I’m pretty sure it’s gonna be great. But. I can’t say that 100 percent for sure. But the more I think about it, the more positive I feel about it. I really feel that of all possible outcomes here, the two most likely are that either nothing happens or that my sales go up.
DH: I think that’s probably right. It’s akin to the coffee shop owner who has a Starbucks move in across the street. His revenue goes up because coffee drinkers are brought to the area.
MA: Exactly. And certainly, I’m going to lose some customers to Reading List. But the question is whether I’ll gain more than that. I’m fairly confident that I will. I mean, I already don’t have 99 percent of iPhone owners using Instapaper. I already don’t have those customers.
DH: Right. And it’s not like you’re subscription based. It’s just a one-time fee. So you’re not going to lose any customers.
MA: Right. If they bought the app and move over, I’ve still got my $3.50.
DH: So now it’s just the pacing of new sales.
MA: Right. And an engaged customer is more likely to recommend to friends. However, Apple’s Reading List leaves a lot of holes. It’s very, very basic. And I strongly predict to anybody who’s trying to predict my death to try it first. There’s no question, it will be good enough for a lot of people. But I already don’t have a lot of the market share. A lot of people don’t even know that there’s a problem to be solved. And so once Apple shows them that a read-later type of workflow is convenient for them, and might help them, and make them more attentive readers, once Apple shows them that they need this, if they look at all for alternatives, they’re gonna find me. So that only be a positive thing. That’s why I’m cautiously optimistic. I think this will be great.
An equally likely case is that people won’t even use this feature. And therefore it won’t have any effect on me. That’s very possible as well.
DH: We’re finding at TMO, in as much as we can track it, and Apple doesn’t make that necessarily easy, the whole Safari Reader thing — that was released last year here — is sort of inconsequential to us.
MA: Last year, I had the similar feeling. I thought, wow, they did something Instapaper does. But it didn’t matter, it didn’t affect me at all.
DH: Right. Well, we were worried about it, of course, because we wondered if we were going to have to revamp our [business] model, would this change our ad revenue picture. Our view on that is, we’re very much against ad blockers, but if someone wants to consume our content in a different way, because that’s easier on their life, like they do with Instapaper, that’s fine. And if that’s the way the majority of the people go, then we have to adapt. That’s our job. Safari Reader hasn’t given us any indication that we have to make that adaption. And perhaps it’ll be the same thing for Lion’s Reading List.
So. Anything new on the horizon? Anything you’re excited about?
MA: What I’m really excited about is all the new things they’ve done in iOS 5. From two sides. First, it’s going to make the platform more popular. And that’s going to help all developers because our market’s going to get bigger. Typically, when Apple does well, I do well. So that’s great for all of us. Also, all the new APIs in iOS 5 are so compelling. They have done a lot. This is a major release. I’m shocked how much they’ve crammed in here. It’s huge. I just can’t wait until I start requiring iOS 5 and making great use of all this stuff.
Some of it, I can attempt to do with an “if statement” and say, “if you’re on iOS 5, I’ll give you this feature.” And then I have features that aren’t available everywhere. Or I have to write two different ways. And one of the great advantages of [iOS] 5 is that they’ve made a lot of improvements to UIKit to remove the hacks that developers had to do. So once I can require 5, I can delete tons of bad hacks that I was using to get around limitations that 5 has now removed.
DH: Very cool. That also makes it more stable and consistent for the user.
MA: And easier for me to maintain. And leaves room and time for me to add better features to the app.
DH: Is that a wrap?
MA: I think it is.