Startup School Radio, Episode 2: How Became Twitch

In Episode 2 of YC's Startup School Radio, a podcast that features stories and practical advice about starting, funding, and scaling companies, Aaron Harris sat down with entrepreneur, investor, and YC partner Justin Kan, and Mathilde Collin, the founder and CEO of collaborative email and text platform Front.

You can hear the episode in its entirety on SoundCloud here and on iTunes here, and read the full transcript here.

An especially interesting part of Kan's interview details how he and the other founders of decided to focus on two new projects that would become Socialcam, the mobile video platform that was acquired by Autodesk for a reported $60 million, and Twitch, the immensely live game streaming platform that was acquired last year by Amazon for $970 million:

Justin: Then to make a long story short, we were at a point where we were about 25 people. We were thinking, 'What do we do next?' The site had kind of tapped out. It turns out not everything type of content is good live, right? Only certain types of content are really good live, and so we tapped out all of these, all the live content, and we were having trouble continuing to grow. The site was probably about 30 million uniques a month at that point, which is pretty big website, but it wasn't growing. We started working on some new ideas of things that we could potentially be bigger than Justin TV and one of those was one of my co-founders Emmet came and said, 'Hey guys, I think we should work on the gaming section of Justin TV.' The gaming section was people playing video games and other people watching them. At the time, [co-founder] Emmett [Shear] came and said, 'This is the only content that I actually like on our site.'

Aaron: How big was that segment?

Justin: That was 3% of our traffic.

Aaron: Just 3%.

Justin: It was just 3%. It was a couple hundred thousand people a month. And Emmett was like, 'This is the only content that I actually want to watch. Let's focus on this content. Maybe we can be bigger.' The rest of us, so out of the four co-founders, I thought, 'Hey that could be something.' The other two co-founders were very skeptical. At the same time we had this other idea, my other co-founder Michael [Seibel]'s idea, which was, let's work on the mobile part because mobile is growing and there's no good way to get videos off of your phone. He was really selling mobile.

We were at this impasse, right? There were two ideas. We couldn't decide between which one was going to work and so we decided let's do both of these things simultaneously inside our company.

Aaron: That seems like a terrible idea. Like when we talk to startups and they say, 'We want to do two things at once' and they're a small team we say, 'Bad idea.'

Justin: Yes, we definitely tell startups not to do this and I actually think that's right. However, in this case it worked out.

Partnering With Insight Data Science

At YC, each one of our startups brings in valuable data each day that could be used to make their products better. But quickly-growing startups often don't have the time to focus on parsing that data. Early stage founders are fully immersed in building their core products -- and hiring a full-time data science team is not usually a realistic priority.

That's why we're excited to partner with Insight Data Science, the organization that runs a seven-week intensive fellowship for post-doctoral PhDs looking to make the transition from academia to the data science industry. Through this partnership, Y Combinator companies will get $25,000 in free consulting credit with Insight Fellows to work on tackling important data science and data engineering challenges. The partnership applies to active YC companies and alumni companies at any stage.

It's a win/win matchup: YC startups have a wealth of potentially meaningful data, and Insight has an office full of PhD data scientists ready to work on compelling, high-impact data problems full-time. Also, Insight is itself a YC alum from our Winter 2011 batch.

Over the past nine months, 26 YC companies including Pebble, Coinbase, and URX have collaborated with Insight Fellows to great success -- you can read about some of those case studies here. We're looking forward to many more success stories like these in the future.

If you're a YC company or alum interested in working with the Insight Data Science Fellowship, email yc [at] insightdatascience [dot] com.

YC Digest - 7/24-7/30

Top Stories from the YC World - July 24-July 30, 2015

YC News

What Makes Founders Succeed by Jessica Livingston

Private Infrastructure by Aaron Harris

YC S15 Launches

YC Alum

Taplytics (YC W14) adds new ‘smart push’ to get the right message to the right person at the right time

Hickory (YC S15) Helps Employees Retain What They Learn During Training

Launching this week out of our current Summer 2015 class, Hickory is a startup that's created an app to help employees remember what they've learned during their initial training.

According to Hickory, some 70% of employee training is forgotten within five days. Hickory has developed an algorithm that can predict when each employee will forget what they've learned, and sends them targeted quizzes and exercises to ensure that their on-the-job knowledge stays fresh and up to date. At the moment, Hickory is targeted primarily at companies with sales and customer service teams.

Business Insider's Nathan McAlone wrote about Hickory in a story published today:

"Hickory breaks the knowledge you have to retain into 'cards' and continually arranges them in the optimal manner, refreshing your brain with small quizzes every day. These quizzes take about 3 minutes per day, or 15-20 minutes over the course of the week, and have proven effective for remember job training details, according to [Hickory founder Brian] Tobal.

So how does Hickory know when you will forget something? Tobal says the program tracks various data points, like how long you spend reading the information, your track record on answering questions related to it, and what you rate your 'confidence level' at."

Read more about Hickory and how it works in Business Insider here.

ScopeAR (YC S15) Helps Fix Complex Mechanical Problems Remotely Using Augmented Reality

When a piece of complicated mechanical equipment breaks, the expert needed to fix it isn't always located locally.

ScopeAR is a startup in our Summer 2015 batch that uses augmented reality (AR) to dispatch remote experts to consult with field technicians on how best to solve a mechanical issue. Using AR features such as telestration, annotation and 3D models that can overlay and lock directly on to a piece of equipment, RemoteAR provides field technicians with an expert that's essentially looking over their shoulder, giving them guidance each step of the way.

TechCrunch's Greg Kumparak wrote about ScopeAR in an article published today:

"Maybe you’re working on an oil rig, and one of the panels is throwing out errors. 'REPLACE VALVE 6B', reads the screen. You know how to replace a valve! You… just don’t know where said valve is. Your company has experts for this, but they’ve all been called off to other rigs.

...ScopeAR, a company from YC’s Summer 2015 class, wants to help experts be anywhere they need to be via the magic of augmented reality.

The idea behind ScopeAR, over simplified: take a video call, and add the ability to draw on and annotate anything the person on the other end is looking at. As they move their camera, anything you’ve added — arrows, text, custom-made 3D models, etc. — stays locked onto the right place.

Need to know which valve is 'Valve 6B'? Point your camera at that array of valves, and the expert can point right to it."

Read the complete article on TechCrunch here, and participate in the Hacker News discussion here.

Ironclad (YC S15) Is An App That Manages Legal Paperwork For Companies

When companies have to create basic documents such as NDAs and sales contracts, they often task their highly-paid hired lawyers with the job -- even though this is essentially administrative work.

Ironclad is a startup in our current batch that acts as an automated legal assistant, creating standard documents that all companies need. This means that paid lawyers can focus on what they are good at, which is providing legal advice on more complicated matters.

TechCrunch's Fitz Tepper wrote about Ironclad today:

"Jason Boehmig, co-founder of Ironclad, explained that current methods to draft and execute legal documents are extremely fragmented. For example, a startup trying to complete a sales agreement would have to separately move documents between their lawyers, document storage solutions, and e-signature companies.

With Ironclad, this entire process is automated, which lets companies save a bunch of time and money.

Notably, Ironclad isn’t out to replace lawyers. In fact, Boehmig explained that the company’s goal is to become an operational layer for lawyers, essentially consolidating their workflow and giving them superpowers."

Read the full story on TechCrunch here, and participate in the Hacker News discussion here.

Captain401 (YC S15) Makes It Easy For Any Company To Have A 401k Program

The process of setting up a company 401k plan can be daunting for a small business -- so much so that often, it's only larger corporations that offer retirement plans for their workers.

Captain401 is a startup in our current class that makes it easy for a company of any size to set up a 401k program for its employees.

TechCrunch's Matthew Lynley wrote about Captain401 in a story published today:

"The goal of Captain401 is basically to make managing 401ks as simple as other services have done for other functions within companies — and in a more digitally focused manner like Zenefits or Zenpayroll. Captain401 seeks to create a process that avoids funds that have higher fees and fail to beat the market, CEO Roger Lee said. The investing service is automated, helping employees make better decisions about their investments and become more educated about the best choices, he said.

'Incumbents like ADP, John Hanckock, they require tons of paperwork, snail mailing, filling out and signing,' he said. 'Some companies like Fidelity don’t work with small businesses. We think we can do a lot better there, our solution is online and paperless.'"
Read more in-depth about how Captain401 and how it works in TechCrunch here, and see the related Hacker News discussion here.

Chaldal (YC S15) Powers Grocery Delivery For One Of The Densest Countries In The World

Chaldal is a startup launching out of our current class that's built an on-demand grocery delivery platform that serves Bangladesh, one of the densest countries in the world.

Chaldal's service, which is currently active in the capital city of Dhaka and plans to expand to other cities in the future, offers the product variety of a big box store, compared to the limited selection often offered at Bangladeshi markets, which are typically short on space.

TechCrunch's Catherine Shu wrote about Chaldal in a story published today:

"Chaldal was founded by Waseem Alim, Tejas Viswanath, and Zia Ashraf in 2013. After working in product development for Wikinvest and SigFig, Alim began exploring the possibility of launching a startup in his home country. He was galvanized by the collapse of a garment factory near Dhaka that killed more than 1,100 workers.

'That jarred me and made me think, if capital is being invested in a way where people have to work inside a mousetrap, then I want to see what it takes to really make a living in Bangladesh,' he says.

Consumers in Dhaka often purchase small amounts of groceries, regardless of their income level, because shops don’t have a lot of stock. Chaldal’s current business model centers around warehouses that are relatively small (about 5,000 to 7,000 square feet each), but still enable it to carry much more items than brick-and-mortar retailers. It uses a cloud-based inventory system that allows users to see what items are available in real-time."

Read the full story on TechCrunch here.

Foxpass (YC S15) Allows Any Company To Have Advanced Server And Network Security

Large enterprises can often afford to have tightly-managed network and server security across their workforces. But for many small- and medium-sized companies, network and server security practices are either implemented lightly or not at all. 

Foxpass is a startup in our current batch that's created an easy-to-use platform for any company to ensure that each employee has individual, company-managed credentials which must be used for access.

VentureBeat's Ken Yeung wrote about Foxpass and how it works in a story published today:

"So why Foxpass? In some companies, it’s possible that employees share the same login credentials when it comes to accessing internal systems such as servers and Wi-Fi. Obviously this isn’t a secure setup, but Foxpass thinks it can bring things up to best practices without resorting to drastic changes in the layout and design of the tools.

Sandersen told VentureBeat that Foxpass functions as an authentication server integrated with Google Apps that companies can use to grant employees access to specific systems. Because he’s dealing with early-stage companies and smaller entities, syncing with Google Apps makes sense. Using this as the identity layer, someone in the company, whether in human resources or IT, can grant that employee specific access. Should the employee leave for whatever reason, access can quickly be disabled without having to change passwords en masse."

Read more in-depth about Foxpass in VentureBeat here, and participate in the related Hacker News discussion here. You can also check out Foxpass' demo in its first-ever "Show HN" back in February 2015 here.

Announcing The YC Startup School Radio Podcast

We're excited to announce Startup School Radio, a podcast that features stories and practical advice about starting, funding, and scaling companies. In each show, host and Y Combinator partner Aaron Harris talks to two key founders or investors to learn how they got started, what went wrong, what surprised them, and what happened as their companies grew.

In episode 1 of Startup School Radio, Harris sits down with Alexis Ohanian, the co-founder of Reddit, and Kaz Netajian, the co-founder of payments startup Kash. You can listen to the entire hour-long show in the SoundCloud widget embedded above, or find all the episodes on iTunes.

One interesting part of Ohanian's segment was the discussion about how Reddit set out to be the "front page of the internet":

Alexis: The really stand out thing is the commenting system, and the voting system that Steve [Huffman] engineered, which we drew heavily from Slashdot influence on. Just this idea of having community self-regulated voting on content, and comments.

And [Paul Graham] just said, 'All right, well, solve that problem. Find the best way to just inform yourself every morning.' And we're going back and forth with ideas. And then he's just like, 'Listen, just build the front page of the internet.' Steve and I looked at each other and we are like, 'Uh, this guy wants to give us money to build the front page of the internet? All right, sucker.'

Aaron: So, that's actually really interesting, because Paul sold his company at Yahoo. He worked at Yahoo for a while, and a lot of ways, in the late 90's and early 2000's, Yahoo was the front page of the internet. There was no such thing as a search engine, really, when Yahoo started. It was an aggregation of useful links.  

Alexis: Yes.

Aaron: Do you think that's part of where that came from?

Alexis: We should have pitched Reddit as Yahoo 2.0. Probably... I still remember the [meeting], because we kept dilly-dallying around this idea of, what it means to have this front page here, and [Paul Graham] was like, 'Look, the news of record, the most important news of the day for the 20th century was essentially the stuff that made it arguably to the front page of like the New York Times.' Some other newspapers might differ, but for the 21st century, the front page of the New York Times meant this is the news.

Online, you have to be content agnostic. Even in 2005, it was clear that the most important relevant news of the day couldn't possibly come from just one source. The best way to get to that would be having a bunch of random people all over the world submitting suggestions for what that might be.

You can read the full transcript from Startup School Radio Episode 1 on Genius here.