Y Combinator Startup Podcast - #101 - Brian Donohue

Episode Date: November 14, 2018

Brian Donohue is President of Instapaper and a Product Engineering Manager at Pinterest.You can find him on Twitter @bthdonohue.The YC podcast is hosted by Craig Cannon.***Topics00:19 - The history of Instapaper8:19 - Free competitors enter the market10:19 - How Brian joined Instapaper14:34 - Transitioning from paid to freemium19:19 - Pinterest's acquisition of Instapaper26:34 - Moving to California29:04 - Working on Instapaper within Pinterest32:19 - Spinning Instapaper out of Pinterest42:34 - Jareau Wadé asks - What types of product integrations could Pinterest have done with Instapaper?50:04 - Ryan Hoover asks - I’m curious how he and the team balance simplicity with new feature development/product expansion.54:19 - Raymond Durk asks - I love the rapid reading mode but would also love a voice enabled mode where the Google Assistant or Siri reads it. Speaking of I'd use it on my Google Home to listen to news if that was a skill.57:39 - Brian Kim asks - Any growth hacks that worked well?1:00:04 - Gustaf Alströmer asks - How does it make time for focused time to catch up on everything he saves? What are his best productivity hacks related to this?1:03:44 - Backpacking

Hey, how's it going? This is Craig Cannon and you're listening to Y Combinators podcast. Today's episode is with Brian Donoghue. Brian's president of Instapaper and a product engineering manager at Pinterest. You can find him on Twitter at BTH Donahue. All right, here we go. All right, Brian. So Instabaper has a very curious, like, lineage in history. Could you just explain that? Sure. So Instapaper was started by a single guy named, Marco Arment, and he built the entire product and service and operated it from 2008 to 2013. Marco, beside Instapaper, was previously known for being the CTO of Tumblr. And Instapaper was kind of his side project that he worked on while he was at Tumblr.

So my impression was he would work nights and weekends. He eventually launched this thing. It had quite a bit of fanfare early on, even though it was just a bookmarking service at the time. Yeah. But it was early days App Store. So this is even pre-app Store. So we're talking in January 2008. Markle launches this bookmarking service called Insta Paper.

And essentially just a website and a bookmarklet. So you could save articles using the bookmarklet. And all it was was a website. Coming into March or April 2008, he launched the reading mode for Instapaper using its text parser. So the text parser would take an article from a website. extract just the text and the images and provide just the reading view and experience. So if used InstaPaper, you're pretty familiar with this. It's one of the core value propositions.

And then in August 2008, he launched, with basically the Appster launch, he launched Inspeer for iOS, which leveraged the bookmarking service and the text parser to kind of create the Instapaper experience that kind of were familiar with today. Yeah. And I think it was the first time where you had this combination. of bookmarking experience, this text parser that would extract the important parts of articles, and then also the mobile application where it was kind of all the articles were saved offline and kind of available to you wherever you were. I think in the early days of Instapaper in that 2008 App Store world,

Instapaper solved three key pain points, consumer pain points. The first one was, if you remember early iPhone days, their early mobile networks were really slow. And it was like impossible to load a website. So you'd be waiting forever. Things would come in pretty progressively. And then when it did load, it was impossible to read the content because there's no such thing as the mobile web.

And so you'd pinch the zoom and try to zoom in on this column. And then you'd be like panning around. And it was like this really weird experience. I think the third thing was that the iPhone was kind of primed to be a reading device because it was always with you. But for a lot of reasons, it was not really a good reading device because there were no good reading applications. The screen wasn't really good. And so Instapaper comes along and it lets you save articles.

The articles are perfectly formatted. You can sync the articles while you're on Wi-Fi and then access them from anywhere. And so I think it's solved. And then also back in the App Store days in 2008, there were not a lot of really useful mobile applications, I don't think. There were like a lot of I fart apps and things like this. And so I think at the time it came as pretty revolutionary and solved a lot of pain points. And as such, kind of had a pretty huge addressable market, which is almost anyone with an iPhone, right? And as the iPhone user base grew and so people kind of grew alongside it.

Now, you fast forward to 2018, most of those things are no longer problems. You have LT networks that are super fast. The mobile web has its problems, and there's various efforts to try to solve them, like accelerated mobile pages from Google. But at the end of the day, you can still load a website and read it pretty easily. And then the third one is there's actually tons of reading applications now. And the iPhone is a main reading device for a lot of people, whether you use Kindle or Apple News or New York Times or Wall Street Journal, or you use some aggregator like Reddit or Hacker News or Twitter.

It's fairly easy to read on your phone now. And so I think the market has shifted from something that had kind of like a pretty wide appeal in the early days to something that's far more niche now. And basically the use case for Instapaper these days is you are kind of like a voracious media consumer. You source content from a lot of different places. That might be directly from publishers. It might be from news aggregators like Reddit, Hacker News, tech meme. It may be from social like Twitter or Facebook.

And you essentially need to place to stage all that content temporarily until the time that you have enough time to kind of read it in bulk, and go through it in bulk. And I think the market size for that is fairly modest. And I think that's part of the reason that both pocketed and instant paper wound up landing. at larger companies back in 2016 and readability wound up shutting down because actually I think the market appeal for that product is fairly limited and it's not one that's like massively growing over time. Did you notice when Safari launched the reading list, a decrease in users?

So when that happened, I believe it was 2011 or 12. I wasn't involved with Instaping at the time. and I haven't looked at those historical trends to see if that was the case. But it's not just Safari. It's actually a lot of applications. Save for Later became a feature of a lot of things. So like YouTube, New York Times, Wall Street Journal, Safari. All these things have like this reading list, save for later.

Like I found this thing and I want to come back to it but I don't have time right now. Yeah. And so I think all of those things kind of chip away at kind of like, kind of like the usefulness of this core service. And there is a large user base who's still using it daily, weekly, monthly. And it's been kind of a big privilege for me to have the opportunity to service those types of people. But let's go back to the intermediate time between the creation and now. Because I think that's Instapaper story is particularly interesting.

So talk about the time, how BetoWorks acquired it. Sure. So the story that I've heard, and I joined Baytorch shortly after the acquisition and joined the Instapaper founding team, the Baytorx's Instaper founding team. The story is that Marco wakes up at 2 in the morning. He's like a little restless. He's trying to figure out what to do with Insta paper. I think there were a couple factors there. A couple of reasons why.

And I haven't talked to Marco about this directly. So this is kind of my hypothesis. But I think one was Marco was probably interested in working out some other projects like Overcast, which is his current podcasting service. It's a mobile app and service that he runs. I think also in 2012. And so the time frame for this is late 2012. In April 2012, an important thing happens in this history, which is read it later, Inc.

which was a relator company originally for Firefox, built for Firefox, both mobile apps and so on. They wound up relaunching and rebranding its pocket. And so this is April 2012. And in doing so, they launched, they won, had raised a lot of capital. They had several engineers working on it. And they had launched iOS and Android applications for free in the app store. Now, Instapaper had no Android application, and the iPhone app was $3.99 to download. And in relaunching Pocket in 2012, I think, you know, I've looked at the downloads for Instapaper,

and they just start tanking after April 2012. And so I think, you know, you have this dynamic where you have this service that's really strong, but there's a free competitor that's also really strong, and people will always see. the free competitor. So I think the combination of fatigue for running the service and pocket launching as a free application in 2012 led Marco to rethink what he was doing with Insta Paper. And so the story is he wakes up at 2 in the morning. He creeps downstairs.

He's trying to think about what to do with InstaPaper. He reaches out to John Borthwick, who's a CEO of BetoWorks. Now, John Borthwick and Baderworks had invested in Tumblr in the Vendez. very early days in their seed round. John Borthwick had Marco Arment come to BetoWorks and talk about mobile app development. So they had a good relationship there already. And Marco basically said to John, I've been thinking about what to do with InstaPaper. I've been thinking about a long-term home for it.

And I think BetoWorks could be that home. And so they eventually worked out a deal where BetoWorks would have a majority share of Instapaper, and Marco would retain some owner. partnership in InstaPaper and continue to act as an advisor. And BetoWorks, the understanding was that BetoWorks would continue to invest in Instapaper, staff, a team around it and continue to build it. So that's, that was announced in April 2013. And so you figure that probably took like about four, four months from December 2012,

which is my understanding of the timeline. Uh-huh. And at which point you get hired by BetoWorks to run it. And then you run Insipaper within BetoWorks for a couple of years. And then how did the Pinterest thing happen? Yeah. So I think we should just back up a second. And so I was hired in July 2013. And I was not hired to run Instapaper.

I was hired as an iOS developer for Instapaper. And so I had started a company. The company had never got more than a few thousand people to sign up for it. It was a location-based social networking company. And so I started in 2011. By 2013, it was really not working. I had gone to BetoWorks for an investment and they had said, we're not interested investing in this space, but we really like you.

And we want to work with you. And it doesn't matter if it's next week, a month from now, six months from now, a year from now, just keep us in the loop. Let us know how it goes. And we'd love to have you here. And so when the company folded, I went to Bay, Bay, at works. I was like, look, need a job. And actually, the person I was talking to is Nick Charles, who, you know, as well.

And so I said, hey, I need a job. And the opportunities were done not done, which got shuttered within a few months of my joining. This gift search engine called Giffy, which I was like, gift gifts. This thing from the 90s, like, this is crazy. But I do wonder what that road would have looked like if I had taken that one. And then the third one was Instapaper. And Inspeper was this pretty prestigious application that it was like a long time. App Store. I had used it in the past and I was kind of interested in it. So I interviewed for that

job. I joined as the iOS developer. And I think a lot of the skills I had from starting a company and that perspective had carried over. There were a lot of gaps in terms of who was writing blog posts, who was doing the marketing, who was building product strategy and so on and so forth. And so I just slowly started to pick up some of those things. And there's a whole story about how I became into the leadership position that I don't know if you want to go into. Okay, cool. So essentially what happened is even in my first interviews, I had said to the person running it, who's a BetoWorks partner named Andrew McLaughlin, I said to him that I did a competitive

analysis across Instapaper, pocket readability. And I had said to Andrew that I didn't think Instant Paper could compete as a paid application in a market with strong free competitors. It just didn't make any sense. Now at the time, did they have a subscription model after the paid? Were they doing like Framium? Pocket? So at the time, Pocket was totally free.

Okay, for everything, every feature. Everything was free. Okay. I think readability was the same. And so you had Instapaper, which had a $399 app download in 2013. And going back in time, it had actually been more expensive. And then also $1 per month subscription for search functionality. So I had this full-text search feature where if you had saved an article about Obama, you could search Obama and it would surface the article, right, even if it wasn't in the title.

So even in that early interview, I basically said, like, inspaper should be a free application. And as I was hired, I kind of advocated for this direction. There was a lot of pushback from it from BetaWorks leadership where because the app downloads were 40 to, 50% of the revenue from InstaPaper. There was a lot of hesitation around just eliminating it. And so what I did is I had been in touch with Representative Apple via one of their education programs. And I told her that we were interested in having InstaPaper participate in Free App of the

Week. And so Free App of the Week is a program that the App Store runs where they take a paid application, they make it free, they put it right in the front page of the App Store. or what used to be the front page of the app store and give it like a pretty prominent treatment. A couple weeks later, she got back to me and said there was a slot. It was like sometime in December 2013. And so basically I'd gone to Andrew, who was running into the paper at the time and I said, listen, we have this opportunity to do free app of the week.

I want to test the hypothesis of making instant paper free. And so what I'd like to do is give a two-month free trial to install. to pay for subscription for everyone who signs up during the free app of the week trial period. And I want to measure how many of them we convert after the two months. And what I want the hypothesis to be is to see if we can continue to build a business by transitioning from a paid application with a subscription to a freemium application. And so we had a couple chats. We eventually went through with the free app of the week.

We executed on this plan. In that time, in that one week period, we had more. more app downloads than we had an aggregate over the past two years. And we increased our subscriptions by like 10 to 15%. And so we basically had it so that three weeks before it expired, there was a little notification that came up in the app. And then like a couple days before it expired, there was a notification that went up in the app saying like,

your free trial is going to expire, consider, consider subscribing. So what did that percentage breakdown of new users? How many subscribed? So the way this worked was interesting. I think we had over a million downloads. Of the million downloads, I think only 650,000 people created accounts, meaning that people download it and just... Can you use it without an account?

No. So they never opened it. Or they opened it and were like, forget it. So you do have to create account to get the experience. And then of those people, many churned out. I think we retained maybe 15% of them. And then of the 15% we got some of them became subscribers. Okay.

I want to talk about the numbers because you can work out the percentages. But we had a 10 to 15% increase in our subscription base. Gotcha. And so basically what it proved is that I think it proved a couple of things. One is that the paid app download was a major cap on our growth and that we were not able to compete in the market with strong free competitors as a paid application. And that, two, when people use the application, we could convert them in a pretty predictable, reliable way. And that they would convert and see the value once they started using it.

And so what I had proposed, and this is January 2014, basically after we executed this successfully, John and Andrew were like, you can be a general manager. So I set out the plan for 2014. And the plan is pretty straightforward. it was make the application free and by the first step was creating more features for the subscription package and then increasing the price of subscription from one dollar month to three dollars per month and then making the application free and so over the next eight or nine months we built we built highlighting and then we capped highlighting so you can make i think five or ten highlights

per month but after that you needed a paid subscription we built text to speech which we offered for free, but we had a playlist feature where you could select multiple articles new text of speech and that was paid. And so we tried to strike this balance between the free aspects of the feature and the kind of like the more pro and like what you would have to pay for. And then we did a couple of other small things. And then in September 2014, we made the application free and launched a new subscription pricing.

So we did that across iOS Android and web. and then after that I was promoted to CEO in the leadership position. And that's kind of how I went from like the iOS guy into the leadership position. We took a pretty big hit, I think, which was expected on the revenue side for a couple months. But within six to eight months, we had recouped what we had lost on the monthly revenue. And we were actually back into a growth trajectory. which was awesome to see. And then a lot of my focus, moving from that late 2014 to mid-2016,

when we sold it to Pinterest, was getting kind of the business fundamentals, right, making sure that we had control over our costs. And as we continue to grow our revenue, hit a place where we were break-even and profitable. And that was kind of my main goal moving through 2016. And then so I think that's kind of... Okay.

Well, as much as I know. So, yeah, sounds right. And so do you know what motivated the acquisition interest from Pinterest? Yeah. So we had fielded some interest even going back in 2015. From them? From other companies. Okay.

There were two large media companies in New York and one media company on the West Coast. One of the large New York media companies, sent us a letter of intent for a large sum. Very large sum. It was in the eight-figure range. Yeah, it was big. And so we were like, you know, we were at BetoWorks and we were talking about it and talking about like the long-term prospects of

Instapaper and whether or not this made sense. And I think everyone felt like this was a good outcome, at least for the shareholders. So I think sometimes you think about the people who, use the product, you think about the employees, and you think about the shareholders. I think, unfortunately, in technology, the motivations of each of those groups are usually very different or sometimes very different. And not always clear.

And not always clear. And so I think what's great now with the new incarnation of InstaVor since we've spun it out of Pinterest is those things are very in line. I think we should spend time talking about that. But I think in this case, the most of the most of the case, the most of the most of the thing. motivation was largely around the shareholders. And so we went ahead with this letter of intent. We did all the diligence with these people.

I did like detect diligence with these people. There's like there's a picture on my Instagram from back in 2015 where I'm sleeping at my desk with my feet on the desk. Like someone someone took a picture of me. And I remember it was like I had just launched a big app update and done all this tech diligence. And I was like I was exhausted. I was just wiped. And then I think we had an eight-week quiet period where they were doing diligence and so and so forth. And in week six, they pulled out because of a strategic change in direction.

And it was really rough. I mean, I remember it was actually Halloween 2015. I actually remember it really well because I remember getting pulled into a conference room being told like they backed out. And that was that. I went out and got pretty. pretty drunk that weekend and just like try to take my mind off of it. I felt okay about it until the Monday I got back to work and I was like, oh man, I have to actually like build this company again.

Like I was like, you know, I was like, I was in the mode like, all right, we're like going to do this thing. And it's tough. Like you get all your emotions tied into it. You think it's going to be a big change. It's at least a new chapter. Yeah. And not to mention like the significant amount of money that would have been.

Yeah. That's helpful. Yeah. And so I think. yeah, I think like there were, there were a lot of things where you were on a certain trajectory there. And I was like, I don't know, I was like looking at houses. Oh, wow. You've really got into it. I mean, I was like, it was, it seemed like a done deal. I was like going to put a down payment on the place in like Gowanus or, you know, right,

in South Brooklyn somewhere. And, uh, and then it's just like, boom, done. Now you're back to kind of back to where you were. And I think it was hard to because I like the whole team knew about it. The whole team. There's like three or four of us. But still, I mean, it's kind of rough to go back to everyone and be like, hey, by the way, this isn't happening anymore. And so I was pretty deflated, I think, like, as you might expect. We went back to the other large media company, the West Coast media company, and started those conversations again. And then Pinterest had been talking to BetoWorks separately about things.

Their Corpter have been talking to BetoWorks. They mentioned that they were interested in news. we had a meeting with our corp dev. I had a meeting with Andrew and the guy who was running corp dev at the time. I didn't think the meeting went particularly well. I was like, you know, I was kind of like, this feels like a waste of time. I don't know why we're doing this. It's like spinning up new conversations, whatever.

And then eventually we got connected with the head of product at Pinterest, a senior PM at Pinterest. There was some connection with Andy. Andrew and the senior PM from the Google days. And so they knew each other and it was very friendly. And then we continued talks with them. But these initial talks were in December 2015. And the acquisition happened, like, signed and done in August 2016. So you're talking about a nine-month process.

And it wasn't all on. I mean, it was like we had talks and then it cooled off. And then we came back. And then there was like a negotiating period or whatever. But just to kind of illustrate, I mean, from the time that we fielded initial acquisition interest until a time that we sold the company was over years, 13 months. So these things like take a fairly long time and you may, you know, have a party that's interested and pulls out. And so I learned a lot from the whole process. So like psychologically, for someone in your shoes,

You know, what do you tell them at this point? Like, you know, don't look at houses until you have a check or, yeah, how do you manage the psychology of it? Yeah. So I would say definitely don't count your eggs before they hatch, right? I mean, that's like, that's the expression. I think it's, it's tough when you are so close to something. I mean, like, you know, we were all like when we signed this LOI, right? We were like high-fiving and stuff.

It was like, I was in the impression it was like a done deal. Yeah, yeah. And to back out of an LOI is from like a business perspective, like usually they're non-binding, but it's, I think, pretty frowned upon and it can damage your integrity and your reputation and so on and so forth. And typically it's not done. I learned later on that you can add breakup fees to these. And so, you know, I didn't know a lot about this stuff. And this was my first time going through this process.

And so you can basically say, like, if you do pull. out, you have to pay X. Yeah. And that can kind of disincentivize them from doing all this work. And then I think also like, yeah, in the pullout, there's some other issues. Like it sends a signal to the market. Like maybe something is wrong. Yeah, for sure.

When that definitely wasn't the case. But managing psychology, I mean, definitely, it was definitely a roller coaster. I would say, you know, you have to like be very calm, collected. don't shop for houses until the money is in the bank because there's no point. And yeah, I just like try not to try not to get too down on the downswings. Right. But then ultimately you landed at Pinterest, moved out to California. You got soft.

You got soft. Yeah. How's that going? It's good. So, yeah, BettaWorks is in New York. I'm from New Jersey. I grew up in the New York area. And so it's part of the deal.

it was understood that we would all be moving out to Pinterest. And when I say we all, Instapur was me doing all the engineering and like all the leadership piece. There is someone doing customer support, maintaining the parser, doing kind of a variety of other things and marketing. And then there was a designer. Of the three of us, two moved out to Pinterest. The designer decided to go do something different. And how's it been? I mean, there's a big difference between New York and San Francisco and it's not something I want to get into necessarily on the podcast.

I've gotten really into like hiking and all the outdoor stuff. I think you're kind of into that too. So I think I think that's been awesome. Working at a large Silicon Valley company has been awesome. And I've learned a ton about building products at scale, processes for doing that. even just like some of the methodologies that are used are vastly different than what I've been exposed to at smaller scale startups. I think of kind of my career as like the scale at which I've shipped software. So in the first company, I started out of college, no customers.

We just demoed it a little bit. The company I started after that, we had 6,000 people sign up for it. Insta paper has millions of registered users. and Pinterest has a quarter billion monthly active users. So that's kind of how I think about like the career trajectory of building software. And when I started at Pinterest, a lot of the initial work was around doing instpaper integrations. And so we did a variety, which we can talk about. And then once I completed those integrations, I started doing a lot of Pinterest specific work.

first as an individual contributor, software engineer. And now starting this year in January, I have been managing a team of engineers building a new feature of interest. So I'm managing a team of eight, very talented software engineers, who are great and I care about very much. And, yeah, so I've learned a ton about managing a larger company. I'm curious about how it's gone, managing it within Pinterest.

We had Cortland Allen from Indie Hackers on the podcast a while ago. And now that's within Stripe. And that's been an interesting process for him. And in most part, for the most part, it's been good. Like he's been able to get off the gas in terms of like raising the sponsorship money. And just focusing on the community and the forum and the content. How has it been for you? And how do you manage that thought process while you also have a job working on, you know, product at Pinterest?

Yeah, sure. So in late 2016, we were at Pinterest. We did the acquisition in August. And around November, December, we had a conversation with Pinterest leadership about some options and directions for instance to paper in 2017. And I think I presented like four or five options. One was one was putting it, staffing a whole team around it and treating it as a business unit and trying to grow it. The second one was staffing some. data science around it and seeing what kind of news-based insights we can get on an aggregate

basis. And so I was always pretty adamant about not sharing personal data in any way, shape, or form, but aggregate information about the number of saves on a URL might be interesting, might help to power some features within Pinterest. The third option was to put it in maintenance mode and just kind of have me and Rodian, the other person for the Instapaper, maintain it. And so that would be kind of like more on the side and doing that. And then the fourth option was shutting it down.

And so that was something that we discussed. I felt that it was pretty low investment to keep it running and that there would be a lot of upset people if we shut it down. And that it didn't make a lot of sense when we had people willing to do it. And we could do it in a sustainable way. So it was an option, but a non-option kind of. It was like we tried heavily to steer away from that. And so they asked me for my recommendation. And my recommendation was to put in a maintenance mode at the time and just treat it as a side thing that I maintained.

The reason being is that treating it as a business unit, I just felt that every dollar put into Insta paper could maybe return 70 cents a dollar. It would be pretty challenging to have it kind of be an investment that had a solid return. but I felt pretty strongly that every dollar into Pinterest would yield a $1.10, $1.25, $1.30. And so it didn't seem to make a lot of sense to put tons of resourcing into the paper when there's so much work to do on Pinterest and that there was much more reliability for getting value out of the work done in Pinterest. Right. And then on the data science, we talked about it a little bit and decided I would just do most of that work myself.

And so we got some agreement. that we would kind of keep in the maintenance in the maintenance mode. And then essentially I dedicated three, four hours a week to instant paper and regularly launched new app updates and fix the servers and do things like that. But large parts of my focus shifted to working on Pinterest. Right. And then I forget when you told me you were going to spend it out. But at what point did you guys decide?

you're like, you know what, we're going to take this thing on its own again. Sure. So we had that conversation late 2016 through 2017. We did a variety of updates starting in early 2018, so this year. Yeah. I was thinking a lot about instant paper long-term trajectory. It felt to me that given the current course, it would be on a path for shutdown, just like drawing out things. And I think all things in life have a certain trajectory that could be.

people, businesses, products. Yeah, but dude, that's the sad reality of a lot of these things that get acquired by big companies. They have, I mean, obviously not a tremendous amount of users that love it, but like some amount of users that love it. And then they just like kind of die quietly as people like, you know, get promoted or like buy a house somewhere and leave and go to another company. Yeah.

It's interesting that you're like, no, we want to keep this alive. Yeah. Let's take it. Yeah. I think it is interesting. I think there's a couple of reasons why we were able to make it work. Yeah. Just to touch on what you said, though, a lot of people were angry about the Pinterest acquisition,

and a lot of people on Hacker News, just to say people who were listening to this podcast, were angry about the acquisition. And they had seen, I think, a lot of applications that they loved, gone by the wayside and very similar things. So whether it was mailbox with Dropbox or Sunrise with Microsoft or parse with Facebook, I mean, there were tons of these things. that were these really beloved applications that had a cult following and a lot of users. And they got bought by the big companies.

They're no longer kind of in line with the mission of the original thing. They get integrated or shut down or whatever. Or turned it to open source or something. Exactly. And so I think the reason we were able to make it work for an Instapaper is Instapaper was always run by a very small team. I've always felt that with our existing user base, we could monetize and cover costs. So I think a lot of these things, like some of the ones I just mentioned, they require lots of engineers. They require lots of upfront capital.

And they basically need to continue to have VCs pump more money into them and continue to grow so that they can eventually become self-sustainable. but they're not self-sustainable, usually in the forms that they were acquired in. And I think that in the case of Instapar is very different. It was built by one person, ran by one person for a long time. When I came, there were three or four other engineers as they left. I picked up backend responsibilities, Android responsibilities, website responsibilities. And I basically got to a point, and I think a lot of people think insidpaper is a big thing with a lot of people.

I mean, basically from late 2014 through now, I've been the person doing all the back end engineering, all the iPhone engineering. Android was starting 2015. But it's basically just been me doing all this stuff. And so because of that, I can reduce kind of one of the largest costs in a company, which is the human cost, right? And in doing so, I was of the firm belief that we could cover our server costs with the existing usage. And with like a reasonable conversion rate on the existing usage, I had the benefit of looking back at a lot of the historical data on how many subscribers we had, what the growth rates look like and so on and so forth. And I just felt like this was a type of products that could continue for the foreseeable future. I knew that it wasn't ever going to map to a Pinterest top line metric or KPI or OKR or whatever you want to call them.

And so basically in early 2018, I had emailed the CEO, Pinterest, Ben Silverman, the subject to an Instapaper. And I just said, hey, can we chat? I basically had a conversation very similar to like what we just discussed, which was that it felt. like the current trajectory was for a shutdown. I didn't feel good about that. I didn't feel good about taking away this product from the many people who use it, just because it didn't map to some top line metric or OKR, KPI, that it had been my privilege to serve these people,

and I wanted to continue to serve them for the foreseeable future. And that I thought the two ways we could put instant paper back on a trajectory where it could survive for the foreseeable future is to either treat it like a business unit within Pinterest, which I said in my... recommendation was the same recommendation I made in late 2016, which is that that doesn't make sense. Or we can spin it out into an independent company. And so we talked a lot about it. As part of the discussion, we agreed that it would be like a handshake understanding that this did not mean I was leaving Pinterest.

So the only thing that would change here is the ownership of InstaPaper. And so now that InstaPaper is spun out, I run it in a very similar fashion. I was at Pinterest. I dedicate my nights and one day on a weekend per month. And I fix all the issues, launch new app updates and so on. And I am trying to build into the paper into a thing that can sustain itself for the next 10 years. It's been around since January 2008, celebrated our 10th birthday in January. and I want it to be a thing that can last for another 10 years.

That's awesome, man. And you did raise a little bit of money now that it's spun out. Yeah. Right. So what were the terms on that? What are the expectations? Sure. Yeah.

Sure. So we raised $80,000 from friends and family. Yeah. Very rag tag, like six people. So to kind of give you the average size, I think the largest check was from me, which was 13,000. and then most of the people had 10, some people put in a five. The pitch was interesting because our express purpose is not to try to like grow this into a huge thing and like exit or do anything crazy.

Like our thing is to make this a long term type of business. And again, like trying to align like what our customers want, which is they just basically want the service to last forever, I think, right? What the operators want, which is, again, just to kind of make it so that this is like, the longevity thing and then kind of aligning what the shareholders, the shareholder pitch there. So essentially what we said to shareholders was we structured it as preferred shares. And so we basically said that we're going to sell you preferred shares at this price. And your preferred shares will automatically convert to common shares when we give you back 125% of your investment.

So basically what we're saying is we're going to have a 25% increase on their investment and will automatically convert them from preferred to common shares. And then after that, any dividends to investors would be done across the normal common share equity split. And so the pitch to them was that you have this 25% return on your investment. We said within a year. year. I think based on current trajectory, it looks like maybe more like 16 months. We'll be able to get there, maybe 18 months. And then on a yearly basis, if we can afford to do a small dividend, we'll aim to get you 10% monthly investment return. I mean, sorry, yearly investment return on your investment. And so there was some risk up front here, but I think all the people who were involved, there were like a lot of XBeta works people. My aunt put in money. Um, so it was very friendly, you know, I think.

Right. And, and, and not like life breaking amounts of money. Very small amounts of money. Um, and so there's some risk, but, um, I'm happy to kind of report that things are going very well, that we do have, um, a lot of people who are converting to inspaper subscriptions. We are now, uh, cash flow positive, which is great. Um, I like a couple thousand dollars. Um, a lot of that was around significantly.

reducing our server costs. I got tons of work around that. But also just continued growth in the subscription offering. One thing that I am trying to piece together is that a lot of the subscription offerings, a lot of the new subscribers are yearly subscribers. And so we need to basically account for that revenue as we realize it. So there's an accounting practice called deferred revenue. So while we're cash flow positive, I don't know.

I don't know how we're doing against our deferred revenue. And I think we want to do accounting on a deferred revenue basis, meaning that we would realize the revenue from the yearly subscriptions in each month that we actually offer the service. So if you subscribe for a year in August, then we recognize $3 in August or $30 divided by 12. And then the rest throughout the rest of the year. So we're cash flow positive. I still need to figure out on a deferred revenue basis, how we're doing as a business, because we need to spread that.

money over the whole year and pay for server costs for those for those people over the year. We've been now this is our third month in business. We're starting to set up books and kind of do things like that. So still early scrappy days of the new instant paper. It's cool, man. So you got some questions too. Let's see if there's anything that are specifically Pinterest related. All right.

So yeah, let's do this one that from Jero. Yeah. sounds like you know. What types of product integrations could Pinterest have done with Instapaper? Yeah. So there were two main ones that we talked about. I think I touched on one,

which was taking the aggregate signals from URLs from Instapaper. So how many saves they got, how many reads they got, how many likes they got. We essentially score those for Instapaper, and that powers a feature called Instapaper Daily. Inspaper Daily, each day of the Insipaper Daily gets rolled up into a weekly. email called Inspeaper Weekly. It's kind of like the best article saved every day. Do you name all the stuff?

Crushing it with the names. The products? Yeah. I'm kidding. So the Daily, the Daily was, I think I named the Daily and the Weekly. The Daily was more a group effort. The weekly was just kind of an obvious one. We've talked about doing like a monthly thing too.

What are you going to call it? The monthly. Oh, nice. That was the idea. Yeah, great. So pretty self-describing. So essentially we score those using something called Instoring. which is done by one of the beta works data scientists.

And we essentially took all the URLs and scores from each day for InstaPaper. And we made them available to Pinterest with the idea of having some news-based feature within Pinterest. I think one thing that was interesting is I looked at the intersection between URLs that we saw on InstaPaper or we had signal on an Instapaper, meaning that there was some save read like in a given day. And then compared those to the URLs. URLs that were in Pinterest in each day. And what I found is that on day, on the first day that we saw those links, only 16% of those links were in Pinterest.

So we actually see like a much different set of URLs on a given day. And then over time, I think within seven or eight days, Pinterest gets to like 60, 70, 80% of the links that we're seeing. So there's still like a, you know, different set of overlap there. but given the nature of the types of links that InstaPaper has, which is very news-based, and the types of links and the experience that we were looking to build within Pinterest, which is like some sort of news-based recommendation system, by the time Pinterest gets those links, a lot of those links are out of the news cycle.

Yeah. And so it's kind of the data is difficult to use. And so we built some experiences. around integrating those signals, around having ranking recommendations for news articles using those signals. But there were these big gaps where, like, it took time for these to make their way onto Pinterest, whereas on the paper, we kind of have, like, all these people distributed on the internet who are saving this stuff to their accounts and doing it, like, very fast

as the news breaks. And so that was one reason that product integration, I think, didn't work. I think Pinterest is also the same. algorithmic recommendation engine that's heavily based around things that are kind of more long-term evergreen like recipes fashion fashion has cycles but um home decor whereas there's much different cadence for news and it kind of changes like the way we do recommendations that wait so were a lot of people just saving like short form news articles instant paper is that common i just i would just assume you would read those and then anything that's long you have

actually save. So there's two big use cases for Instababre. There's like the long form one, which I think is what you're describing. But there's also the like, I want to be able to find this stuff again. Right. So it's more of a searchable archive. Yeah. And so people will save things that they have read and I do this too. I save even like pretty small things like I've saved things like campfire recipes. So if I go backpacking or camping somewhere just so I can like kind of get to it really easily. It's offline, which in nature you might be offline. So those are the two key use cases. I've had a lot where I've read this article about startups or entrepreneurship or whatever.

And I wanted to recall it and get the thing back, like pull up the article again. And doing a Google search sometimes, it's like, I can't remember which blog it was on or so on and so forth. But it's much easier to do if you have kind of your own personal search index. and if you subscribe, plug for incentive or premium, you can search through all the articles you save and do a full text search. Okay, cool. So we do get both,

and we do get kind of a different cut versus what Pinterest gets, at least, on day zero, which is when that recommendation is most important. The only other thing that we kind of looked into was kind of doing like an offline reader mode for Pinterest. I think for various reasons, like at the scale of Pinterest and so on and so forth, it would be very difficult to do that. People save a lot of things in Pinterest. And I think the expectations are a little bit different.

Like people expect that all their instant paper stuff will be available. If we built that expectation for Pinterest, we'd have to store like tons and tons and tons of stuff. Yeah, massive videos. And just so many, so much more content because the nature of the services. actually very different. And so it was something we discussed. And actually when we were talking about spitting it out, we discussed like maybe this would be a place of value.

And I talked to the head of product about it. And I said, like, there's so much stuff that is important for Pinterest right now. And this seems like the type of thing that seems like a nice idea, but we'll be infinitely deprioritized behind like the many other more critical things we need to do. And so I don't really see that. as something that's viable, like, are you going to staff people to do this project? And he was like, no. And so I said, yeah, because it doesn't really make sense to do. It's like, and like all those like super technical features. It's funny. Like I think about overcast in this way. Like overcast,

I don't have the numbers, but it seems to be very highly adopted among nerdy people who want like crazy granular control over like notifications per show offline downloads per show. But the average person doesn't care at all. Yeah. And so for Pinterest, like, focus on that, it seems. Yeah. Yeah. And I think Insleeper is actually very similar, right?

I wouldn't describe them as nerdy. I would say they're very, there's continued people who are very particular about the features that are in the product. They want massive control over the features in the product. They're very privacy focused. And I think Marco has built a large audience of these people, right? So I think what's cool about Marco is that he has this audience. And so when he launches products, he launches them directly to this audience.

And they typically go and they download, buy, subscribe, use the products that he builds. And they do it because Marco is the same type of person, right, who wants kind of that control, those privacy settings and so on and so forth. Yeah. So I think it's a great model. It's very cool. A lot of people look up to him. So got some silly questions in here. All right.

Here we go. Ryan Hoover from Product Ton. He asks. I'm curious how he and the team, a very small team, balanced simplicity with new feature development and product expansion. Yeah. This is a really good question. We've built basically our product roadmap,

based, always historically, based off of user requests, so people that write in. And we track all those user requests. And essentially, we build a model based on, like, the most requested features. It was like always a big thing. Even going back to 2014, when I got put into the general manager role for into the paper, I built this like little table, this three by three table.

And the things that I want to evaluate features on were did users request it? Does it give us a competitive advantage? Can we build it into the business model? And then evaluating all the features based on that. We essentially built our roadmap based on what users want, thinking about monetization a little bit. And then also does it give us a competitive advantage? A lot of the features that we've been requested are very power user features. So people want to like do highlights.

They want to create annotations. They want to be able to export all their highlights and annotations to a third party service like Evernote or something like that. The amazing kindle. Amazon. Exactly. I think actually a lot of people want to take all their highlights and put them into one place. There's this notion of like a commonplace book where like all your learnings are kind of in one spot. And so I know Ezra Klein, who's the editor-in-chief of Fox, uses InstaPaper to Evernote,

and then puts his Kindle highlights in Evernote using some service. I think it's called Clippings.io. And together, he kind of has them all one Evernote. So again, it's like this searchable memory bank. Exactly. Of all the things that he reads. And so many of the features are very power user features. Some of the other ones are speed reading.

Text the speech, again, like people use and to paper in their car while they're driving, right? So they'll put the playlist on and just go. Use the iOS speech to text. Yeah. So we use the speech synthesizer. There are a couple new iOS voices that are really great. So there's one called Samantha.

That's really high quality. And the one that most of our users request is called Alex. So Alex is I think it's like a couple hundred megabytes download, but you can kind of go into the voices in iOS settings. And if you have that downloaded, we can detect it and use that by default. So our strategy here for building that is to keep the product really simple for the majority of people. It's already a pretty complicated product where it's like, okay, you're going to use this thing, but you're going to find stuff elsewhere and then put it in here and then come back to it.

It's already like, it's already overly complicated. The browser extension is like, whoa, it's too much. Yeah, yeah. You're like going to install this here and then it's a challenge. And so our strategy for simplicity with new feature development, and it's not a great strategy, is we just hide all the features. We put them kind of in hidden places. And I think our idea and our philosophy behind it is the people who want those features are going to go out of their way to go find them. And we don't care about moving metrics on specific features or whatever.

What we care about is the overall product and keeping it simple. And so most of those features like text to speech and speed reading, if you go into the reader view, you press the share button, they're in the bottom sheet there. And they may be kind of pushed back based on what other things that you have there and what you've configured. And if you scroll on the bottom sheet, you can get to the text of speech or the speed reading features. Many people don't know they exist. I think there's a question. I didn't know it existed. There's a question on here that says, can instant paper be voice enabled?

So we can. Yeah, yeah. We can cross that one out. That's a yes. That's a yes. There was one. Raymond Dirk asked about the speed reading. So I love rapid reading mode, but would also love a voice enabled mode where the Google Assistant or Siri reads it.

Speaking of, I'd use it on my Google Home to listen to news if that were a scale. Yeah. So we do have a voice enabled mode for text of speech. I think the home assistant one is interesting. And so I looked into creating Alexis. skill for Insta Paper in early 2016. Yeah. And I think at the time it was, it was very difficult to do because you can only pass

500 characters in a response and an Alexa skill at a time. Now, Amazon has more recently launched Amazon Poly, which kind of has some of the Alexa functionality for developers on AWS. And more recently, Pocket has done an integration with Poly. So they actually send articles to Poly, and they can. I think what Polly returns is like an MP3. All right. And so you can use Polly with your Alexa scale.

I assume you could do it with Google phone, too. I think the challenge is Polly has some costs associated with it. And I haven't done the math to figure out what percentage of users would use this and what the cost would be. But it might not be feasible for us to do, even with our existing premium model. It's like based on the volume of usage for each individual user. not sure it would make sense. But it's something that I'm still looking into. That's particularly just for the Google Home and Alexa type of environment.

Right. Has that unique consideration, I think. Yeah. But I guess you could always just play it on your home speaker. Yeah, you could always use it on your phone and then put it wherever you want. Right. All right. Cool.

Have you tried any of those like Liarbird type integrations where it does a straight up voice emulation? So I'm not familiar with that, but we did try two different. third-party text-to-speech companies. There was one called Ivona, which I believe is actually an Amazon company. And there was another one we used that's slipping me. When embedding those with the mobile application, it blew up the size of the mobile application like hundreds of megabytes.

And so we looked into downloading the voices asynchronously and like basically you download the app and then you download the voices separately. Kind of felt like it was a lot of work. even just actually linking the text to speech library just to do the functionality blew up the binary, I think, two times the actual download size. Does that materially affect the number of downloads you get? I think only if you exceed the App Store mobile download limit. And so that's changed.

At the time it was 50 megabytes. I believe now it's 100 or perhaps more. But I think at the time, to be it was 20 to 30 megs. Right now it's 50 megs. And linking these Texas speech libraries blew it up to like 70 megs. So at the time, I think it would have dropped app downloads. I think now it's a little bit easier to do on cellular networks and the limits are higher.

So we're basically looking at that and then we have to pay for these things and trading off against this free stuff that comes with iOS devices already. and we can use no problem. And it's like fairly good and probably the quality will improve over time. And so it was a very short-lived type of experiment. Yeah, it would massively change the app for a small percentage. Exactly. Right. Okay.

So Brian Kim asks any growth hacks that worked well. So we didn't focus too much on growth hacks. I think our I think our focus has always been on delivering good products. That's always been our best growth hack. That's always been... That is the best growth hack. Basically, like, when we shipped features that our users wanted that our competitors didn't have, we were able to get more users using that. Now, I think of growth hacks in a more simple way, growth hacks basically is when you take your product, you take a link to it and you posted somewhere else to try to take users from another service.

right and so if your Airbnb your growth hack is posting in a craigslist if you're genius your growth hack is making sure that your lyrics pages are at the top of google searches for those lyrics um there's a variety of other ones but essentially they involve going to this other service and pulling their users um that involves kind of publishing some some content maybe their their apartment listings maybe their lyrics or so on and so forth um the content that we have on the service is saved from our users, and they're from media publishers or bloggers, and we would never republish that content, and that content is specific to that user.

Yeah. And so a lot of those types of growth hacks and tactics don't really work for instant paper. There is a variety of things we could do. You can publish, like, SEO pages of collections, so we could, like, have some collections of articles and try to publish those. It's just never been a big focus area for us. I think probably our biggest growth hack was removing the paid download, right? And demonstrating that that paid app download that we removed back in 2014 was a massive

barrier to our growth. And today, Apple actually goes out and recommends to have developers that they have subscription in their apps and they don't do app downloads. I think in 2013, 2014, that was less obvious. Right. That's the subscriptions were kind of a key piece here. Yeah. Yeah.

Well, it's so much more mature now. Yeah. It was back done. All right. Gustav from IC is curious about your personal reading habits. So how do you make time for focused catching up on everything you save? And what are your best productivity hacks related to this? Sure.

So you need to make dedicated time to do this, I've learned. and I honestly saved more than I read and read less often than I would like. I found that when I was offline, so for instance, when I was in New York City in the subway, back in 2014, 2015, there was no internet service and so there's nothing else to do. So I would read ton on instant paper. That is changing rapidly. I think even in airplanes now, most of the time you'll have Wi-Fi. I've gone back to New York and almost every subway station you have LTE. I live here in the Bay.

I take BART. There's basically service on the entire BART line. And I find that more and more, there's less time I'm offline where I'm going to Instapaper. And so I do need dedicated time on the weekend or nights to actually go through it and read them. There's times, like I'll read a little bit during lunch, especially if it's work-specific articles. I think the best productivity hack I have that's instant paper related is on the mobile application. If you press the ellipsis button in the top right, you can press. There's a sort option, and you can sort by shortest article.

And then I'll go and I'll just blow through a ton of short articles and just do them really, really fast. And I'll leave kind of the bigger, meteor ones for later. I think another one that I do is I'll go through periodically and I'll just delete a bunch of stuff that I know I'll never get to. Especially like there's tons of Donald Trump articles in there that are like just just no like thought pieces about Trump. Yeah. And there's just no longer relevant because so much has happened between now, between today and two or three months when I saved it. Yeah. That it's just no longer interesting or relevant.

And so I'll go through and I'll delete it a bunch of things like that or like this company. raise this money. I'm like, oh, I don't really need this actually. So I think those combination of things are the, the best productivity hacks I have around it, but making time to actually revisit the things that you save is very difficult.

We've thought about building features to help with that, but it is really hard to build that type of feature and not making it massively annoying for the person using it and make them feel bad about not going back to it. It's the same thing as exercise. Exactly. Until you really want to do it. doesn't happen.

Exactly. And it's just guilt. Exactly. Yeah. Do you ever see, I don't want to mess with it? I think his name's Dan Williams in London. He made a Chrome extension that was like send a Kindle, but it didn't do anything. It was like JavaScript void.

And it's like, you could just click it. That's awesome. It would trigger that thing. And I actually did install it for a while because I realized like going back to my Instapaper or pocket or whatever I was using at the time, I could pretty easily ascertain the stories that I'm actually going to read. and having more discipline around that was what allowed me to get. I mean,

I'm never at like inbox zero on any of this stuff, but closer to it. Right. Also, gray scale. Have you experimented with that on your phone? I haven't. I know it's like a big thing people are doing.

Gray scale plus airplane mode, you get shit done. Yeah. Because your phone's just so lame after that. Yeah. You can actually read. I think the last thing I want to talk about is, yeah, all your hiking ambitions.

Sure. What's your, what's your training strategy? lately. Sure. So, yeah, it's interesting. I'm from the East Coast. I was not into nature. I think, you know, if you live in New York, there's, you kind of get sucked into the city. And the vibe in the Bay Area is very different. And I think the thing that I've really loved about living here is the easy access to nature. And so I started just by like exploring and hiking. I did like a lot of my like first hikes here, you know, just like my first like 10 mile hike, things like that.

I think as I started pushing the upper limits on hiking and doing things like going to Yosemite and doing half dome and doing like these 16 mile hikes, I got into more like camping, backpacking, doing much longer hiking. And so I think this year I've done like 10 backpacking trips like all over the place, which has been pretty awesome. Well, you did some big ones. You're not just doing like casual stuff, right? Like you did Whitney, you did Shasta, right?

Have you done other big ones this year? Yeah. So I think those are the two big ones. So from backpacking, it's kind of been like, well, some of these mountains look really awesome, right? And so mountain climbing is a super dangerous thing. And so when I did Shasta, at least where it was like heavily snow and ice, I went with a company that brings people up there. Yeah, I'm not trying to like die on a mountain. Like, you know.

Right. Yeah. It's dangerous if you don't know what you're doing. Exactly. So, but so I did Mount Shasta, which was great. I did Mount Whitney this year, which was extremely hard. Mount Whitney is the highest. peak in the lower 48. It's in California here. It's in the eastern

sierras near the Nevada border. I did that one basically in a day. I went up and came back. It was 13 hours, 8,000 feet of elevation gain. I got a massive headache at the top, which sucked. If I do it again, I'm definitely going to sleep overnight. Because you went from Lone Pine, right? You went from the east side. Yeah. You start, I slept in the parking lot, like right outside the parking lot. And I think that it's like 6,700 feet or 7,000 feet. And then you go up to 14.5. So it's a pretty big lift for one day.

There is a campsite at 10,000 feet. I just didn't get a permit for it. Otherwise, I'm not a masochist. I would have done it. And so I have been doing these pretty intense workout routines for mount climbing specifically. And it started with just the Shasta one. It was like a 12-week mountain climbing routine.

heavily focused on endurance. So a lot of cardio, a lot of like long cardio sessions yesterday. I did an hour on the stairmaster with 20 pounds in my backpack. And just just kind of doing things like that. And I think what I've realized is like you do hit these walls, but the more you can kind of like, you know, you're like where you feel like you want to give up. And then more you can like break through them and get used to like hitting the wall, breaking through it, breaking through it.

and doing that in the gym versus on the mountain is is kind of a good strategy. And so I have I have some more ambitions. Like I don't think I'm ever going to like want to do Everest or anything like this. Like that's a little like real climbing. Like too extreme. Like place climbing. I would be interested in that. I just I want to balance like having a good time outside with being safe.

Oh. Yeah. But there's their whole categories like type two fun. Yes. And yeah. I agree with you. I fall into the camp of enjoying the suffering, not the dangerous part.

Yes, exactly. And I think with Everest, like, if you need to put oxygen on in order to get up, it's like, you know, maybe it's just, you know, I know people do it, but it just seems, I know people do it without oxygen and those people are awesome. Yeah. But it just seems to me like maybe a step beyond what my ambitions are there. But I would like to do like Mount Rainier. There's a couple in South America that I'm kind of interested in doing, at least like O'Neer. And it's kind of been like a whole new thing. thing for me, this whole area. And so it's something I've really enjoyed living out here in

California and exploring. Cool, man. Well, if someone wants to get in touch with you or reach out to you on the internet, what should they do? Sure. Twitter is probably the best place. My Twitter handles B.T.H. Donahue. I usually give out my email too, but it's a lot of people. Don't worry about it. So if you reach out to me on Twitter, I will give you my email address and we can chat. But I think That's probably, we'll start with Twitter. Cool, man. All right. Thanks for making time.

Yep. Thank you. All right. Thanks for listening. So as always, you can find the transcript and the video at blog. Dot Ycombinator. com. And if you have a second, it would be awesome to give us a rating and review wherever you find your podcast.

See you next time.