- Cover is a multi-line national property insurance entity that allows customers to insure various items by taking pictures and videos of them.
- The app uses computer vision technology to identify and catalog the property, making the claims process easier.
- The app acts as a front line underwriter and provides proof of the property's existence, reducing pushback during claims.
- The CEO, Karn Saroya, previously worked in management consulting but wanted to transition to building things and taking risks.
- Saroya and his team initially built a body scanner startup but pivoted to creating an app called Stylekick, which allowed users to indicate their preferences for fashion items.
- Stylekick became a marketplace but struggled to generate significant revenue.
- The team eventually joined Shopify and built experimental marketplace apps before deciding to start a new mobile-focused venture.

- The founders of Stylekick tried various industries before settling on insurance
- They realized the potential for profit in the insurance industry based on the high premiums and customer churn rate
- They initially launched a simple version of their app without insurance licenses and quickly gained popularity
- They raised $3.2 million in funding and focused on owning the customer service experience
- They eventually expanded to become a national insurance brokerage in under 12 months
- They prioritize customer satisfaction and have received positive reviews for their services
- They are aware that they are still a small player in the insurance market and are cautious about taking on balance sheet risk
- They acquire customers primarily through their highly rated apps and also through paid advertising and search engine optimization
- They offer additional services and tools to customers, such as price drop alerts and access to defensive driving schools, to enhance the overall customer experience
- They believe in continuously iterating and improving their front-end experience before expanding into other aspects of the insurance industry

- The speaker mentions that they observe what customers are asking for and use that information to build products that support those needs.
- They also read insurance rate filings to gain a better understanding of available discounts for customers and how to make them more accessible.
- The speaker gives an example of a discount offered by Geico for owning a share of Berkshire Class B stock.
- They discuss the decision to buy the domain Cover.com and explain that the ROI on buying an expensive domain can be significant.
- The speaker advises caution when buying a domain and recommends using a broker.
- They mention their engagement to one of their co-founders and discuss the benefits and challenges of working together.
- The speaker explains how they manage their two offices in Toronto and San Francisco and keep everyone aligned on the organization's goals.
- They emphasize the importance of making insurance easier, fairly priced, and providing positive claims experiences for customers.

- The speaker emphasizes the importance of being present when claims are being made in order to provide positive experiences for customers.
- They suggest acting in an advisory capacity and scaling this approach across different lines of business in the insurance industry.
- The speaker, as an international founder, shares advice for other international founders applying to YC or post-YC, recommending having a presence in the Bay Area if possible.
- They acknowledge that visa issues can be a challenge but advise working through them.
- For new YC batch participants, the speaker advises focusing on a key performance indicator (KPI) and ensuring its growth.
- They also suggest avoiding unnecessary hiring, conferences, and parties that are not directly related to the KPI.
- Post-YC, the speaker shares that their experience of going through the program a second time helped them maintain a steady rhythm and achieve their goals.
- They mention raising an A round of funding quickly by effectively communicating their progress to investors.

- Cover is a national property insurance entity that allows customers to insure items by taking pictures and videos of them.
- The app uses computer vision technology to identify and catalog the property, making the claims process easier.
- The CEO and his team initially built a body scanner startup before pivoting to create an app called Stylekick, which struggled to generate revenue.
- They eventually settled on the insurance industry and quickly gained popularity with their app.
- They prioritize customer satisfaction and offer additional services to enhance the customer experience.
- They continuously iterate and improve their app based on customer needs.
- They acquired the domain Cover.com and emphasize the importance of making insurance easier and providing positive claims experiences.
- The speaker shares advice for international founders and advises focusing on key performance indicators and effective communication with investors.

Karn Saroya is the CEO and cofounder of Cover, which was in the YC Winter 2016 batch.

https://twitter.com/karnsaroya/
https://cover.com/

Cover is a nationally licensed insurance brokerage. You can use their app to take a picture of property you want to insure and they’ll connect you with their insurance partners so that you can get the best price and coverage.

The YC podcast is hosted by Craig Cannon.

https://twitter.com/craigcannon

***

Topics

00:08 - Karn’s intro

00:33 - Using computer vision to identify and catalogue property

2:00 - How Karn ended up starting Cover

3:20 - Being a maker vs. an advisor

5:30 - Stylekick, Karn’s previous startup

9:00 - Joining Shopify

9:40 - How the idea for Cover happened

11:30 - The capital-light way to start an insurance business

16:00 - Underwriting

17:30 - Lead generation

19:50 - Product development

21:00 - Buying Cover.com for $750k

24:25 - Being engaged to a cofounder

28:15 - Managing two offices and cultures

30:30 - Being an international founder in YC

31:30 - Advice to people in the current batch of YC

All right, so today we have Karn Saroya, the CEO of Cover, which was in the winter 2016 batch of YC.

Karn Saroya First of all, thanks for hosting me.

So you can think of Cover as a multi-line national property insurance entity.

We ask a couple simple, enduring questions, and they take pictures and videos of things they want to ensure.

So this could be cars, they could walk us around their homes, pets, jewelry, electronics.

We basically make a market for just about anything you can take a picture of.

And it's processed with computer vision, not humans, correct?

Matt F.: On the home side, it's computer.

So we use a TensorFlow-based camera to identify and catalog your property so that when you need to make a claim, there isn't much of a fuss that's put up by an adjuster.

Okay, And now is it assessing more than what the object is?

The value of us being a visual app is twofold.

We're acting as a sophisticated front line underwriter.

We're proving that the property existed in a given time, place, and condition.

And that helps materially improve our loss runs.

And then for the customer, what it means is that there can't be very much pushback in the instance of a claim.

You have proof that your property existed.

An adjuster can't come back and say that, hey, that television that you're trying to claim is actually something that's a inferior model or something.

are your models constantly adjusting per person?

So should I be photographing everything I buy?

It's not necessarily the central tenet of what we do.

A big part of it is to simplify the onboarding of getting insurance, making it a bit more natural on native mobile.

kind of an interesting divergence for you, because before you had a style startup.

So maybe you should explain how you ended up here, because I think it's interesting.

I was a management consultant in a past life.

So I was at Oliver Wyman specifically in their financial services practice.

So I was in their finance and risk practice, a little bit of insurance work.

It helped me pick up a little bit of polish the 2 years that I was there.

Certainly it can model things and build PowerPoint slide decks.

I'm appreciative of the experience that I had there.

But at the end of the day, I kind of wasn't getting what I wanted out of the experience.

The risk-adjusted returns to being in professional services like managing consulting or banking or private equity are great.

But because they're great, I view them as temporary.

And so I'm looking at it and I'm like, in the long run, eventually there's reversion to the mean.

And I kind of want to make the jump to something where I'm building things, I'm taking risks.

And that just gelled better with my personality.

So did you find that you were struggling to care while you were consulting, or you just didn't feel there was sufficient upside?

Well, I was learning to code while I was consulting.

And so I was like, well, I'd rather be a maker of things than an advisor.

1 of the questions that kept percolating in my head was, They're paid millions of dollars to advise banks to do things and to improve operational processes.

And I kind of wondered why the partners that I worked with just didn't build a bank.

And I posed that question to a couple of the partners.

And the answer was, hey, man, we make a lot of money doing this.

We're not going to end up taking these risks.

Quite frankly, the scope of building a bank is way, way larger than the thing that we focus on, whether it's in an insurance context, solvency related, in the banking context, stress testing related, or operational, or something to that effect.

And I was like, well, I'd be the guy that wants to build the bank, right?

And so I made a jump, and it was kind of ad hoc.

The very first thing that I built with Anand, who's our CTO, was actually a body scanner.

And so we used the 3 Kinect sensors, because it's like a depth sensor webcam, 3 points of perspective to get a body form.

And we thought that we could use math to shave away clothes and use that body form to inform sizing decisions on the internet.

And I think if we had made a couple of logical checks before we had ventured off to do this thing.

Like, why aren't there body scanners in every mall in America?

We probably would have realized that if this was something that consumers would adopt, it would be there.

It would be like a war for mall space across America.

And now, was this something that was intended to be a startup, or did you just think it was a cool side project?

We were pitching the Gap and a fair number of large retailers and brands.

And so we stepped back and thought about, hey, what are we actually trying to do here?

We're effectively trying to predict the sizing and stylistic preferences of our customers.

And so over the course of a weekend, we built a super simple app where we scraped a whole bunch of lifestyle blogger and fashion blogger content and threw up outfits and allowed people to scroll and double tap on individual parts of an outfit.

So all we wanted to see was that somebody would indicate a preference for a blouse, or a handbag, or a hat, or a pair of shoes.

So we eventually stepped back, we pulled it back, and then we built out something that would register the unique structural elements of individual items that people would interact with.

If you double tapped on a handbag, we know the price point, the colors, the unique structural elements.

it actually was even more rudimentary than that.

And so that was working, and eventually we started getting to the point where we were predicting what people would want to see next.

And so with Stylekick, actually, we applied to YC 4 or 5 times.

And by the last time that we interviewed, we had 400,000 active users.

And the business was based, was it sort of like supply?

So eventually what we had was designers, influencers uploading looks, And we were earning affiliate income doing that.

Paul At that point, you had 400,000 active users.

How much money were you guys making out of those?

Not nearly enough, just barely to cover our hosting costs.

And the reason for that is a lot of the stuff that was being sold was super high-end.

Back then on a four-inch screen or smaller, selling a pair of $300 jeans wasn't an impulse-purchase.

We ended up acting much more like lead gen to a desktop purchasing experience.

And in retrospect, what we probably would have done and stepped back and looked at was actually working within the marketplace native mobile e-commerce ecosystem.

The things that were working like Wish, so impulse price purchases, 5 bucks, 20 bucks, where you didn't need to think twice about what you were buying.

What was also working, we relied heavily on influencer marketing to grow to eventually a million active.

And so if we step back, we would have recognized that there were a lot of these individual brands that were starting to stand up on top of Instagram.

And we probably could have gone the route of actually building out a fashion brand or selling products of our own accord.

Either you went vertical, or you went impulse price purchase.

Because something in the middle was just never going to happen.

Did you consider a full desktop interface where people could save and then purchase?

I think Net-a-Porter and Farfetch and a bunch of these guys went that route.

In and around the time we were running out of Steam, we got to know the folks at Shopify particularly well.

Craig Miller, who's a current CPO, Satish, who's a VP and a GM there.

We had a longstanding relationship, and ultimately we decided to join Shopify and run a mobile product team.

We were building experimental marketplace apps for them.

We knew that we could build really beautiful mobile products.

Stylekick itself was featured in just about every market on the app store, on the front page repeatedly.

We were driving millions of people through the app, so we knew how to drive distribution via native mobile.

And so we started thinking about, what is the next thing that we want to do?

We knew we wanted to be mobile, because that was 1 of our core strengths.

We wanted to make sure at this point that the basic economic model worked.

We'd drive to northern Ontario and Muskoka, just have a couple of beers and think about what we wanted to build.

We considered entertainment again, because Natalie worked for Russell Simmons and it's kind of the impetus of us moving forward with Stylekick.

And the reason was, we had been working with insurance brokers in the past to help us pick up policies for StyleKick and for ourselves.

But this is not necessarily obvious, right?

Because you were in Toronto and now covers in the US with a more competitive market.

So what convinced you that this could work?

So I think what we wanted to prove was that the pipe was big enough to drive a significantly sized business.

At the end of the day, it was the back of the envelope math that showed us that this could work.

A single auto policy that we sell in California, the average premium is something like $1,600.

Just as a distributor of insurance, never mind underwriting or moving down the margin stack, We earn anywhere between $200 and $300 of that policy in perpetuity.

And so if you think about what churn looks like in insurance, a very, very good broker is churning 10% per year.

So the LTV on a single auto insurance policy is thousands of dollars, which is why you see Geico and State Farm spend billions of dollars a year trying to acquire these customers.

Ryan Yerrick Yeah, because they're just massive.

But you also in the beginning knew that you couldn't afford to underwrite this stuff, correct?

So we were thinking about what's the capital light way of entering this business.

And so the very first version of Cover was super simple.

It was, we didn't have any licenses because we weren't selling any insurance.

We didn't know anything really about insurance distribution.

It was just a handful of views leading to a camera view with a preamble that said, take a picture of something you want to insure.

We launched it and we used the methods that we used to grow Stylekick to grow Cover.

We ended up being the number 1 ranked insurance app in Canada the second day.

And during YCU, we were actually 2 weeks into launch, we were featured under the best new apps in the United States.

Again, we didn't have any insurance licenses.

All we could see was that every 60 seconds we were getting an insurance request.

So how did those conversations with investors go?

Because I think you could, as just a distributor, I think a common pushback would be like, Isn't this just running to 0 margin?

So I think any pure lead gen business is just relying on a marketing arbitrage to continue to exist.

And there are lots of examples of that in insurance.

I don't think that you can build a durable business as just a lead gen service, for a couple of reasons.

1, those arbitrage opportunities eventually dissipate.

The second is you're not building any brand equity or value, and you're certainly not doing a customer's service by vacuuming up their contact information and selling it on 10 times to juice average revenue per user.

Especially as a startup, a venture-back startup.

Yeah, and so we knew exactly what we needed to do.

We were like, okay, well, we need to start owning the actual customer service, the selling experience, and eventually more.

And so during YC, we were like, we're committed to getting license across the country.

We're committed to getting as many insurance markets as we can on the Covered platform, so that when a customer comes in the door, irrespective of what it is they're looking to cover were a competitive place.

And so we stood up a national insurance brokerage in under 12 months.

So we were in 49 states with 30 carriers in under 12 months, which is, I think, enough to show that we could de-risk ourselves from an execution perspective, because that's a non-trivial thing to do.

And so how much did you have to raise to accomplish that?

So if you're in winter 2016, so by winter 2017, you were in that position.

Guy Podjarny We raised 3.2 million bucks.

So Some of it before demo day, a lot of it after demo day.

We actually had most of our seed rounds still when we raised our A.

And a big part of it was we knew that there was going to be a ramp.

So we didn't make our first hire for probably the first year and a half.

We knew that it would take a little bit of time to get everything tooled before we could start to scale.

But yeah, we raised our A as soon as we were in a position to be able to be selling insurance business.

and so when you raised your A, did you feel that you had product market fit?

Ryan Neuhofel I mean, if you take a look at our reviews.

I mean, we do everything we say we're going to do.

A customer comes in the door, we make the underwriting application process super simple.

We do not spam our customers with phone calls or emails.

Every single 1 of our customers gets a dedicated text line where they can text us back and forth to ask questions.

It's a truly, I think, millennial-first experience, which is a cliche thing to say, but it's the thing that we do.

So yeah, we do what we say, and we've been doing well because of it.

And so when it comes to growth, you guys just raised a B in 2018.

There are other insurer techs that have raised hundreds of millions of dollars now, right?

Still a drop in the bucket in the insurance market.

There are $500 billion in insurance premiums that are written just in the property market in the United States every year.

And so, look, that's sort of a concern to us because, 1, we don't take balance sheet risk yet.

Eventually we may be, because we can be pretty opportunistic about the types of requests that are coming in and whether we want to stand up products for those requests.

Paul Matzkoff In practical terms, for people who aren't insurance, what does that mean?

That means that If we see a lot of customers who are coming in for a specific thing, like maybe it's home or maybe it's auto in a given geography and we're writing it profitably, we can actually just step in and underwrite the business ourselves.

Meaning that we take the risk, or we price the risk, take part of the risk, and sell the rest of it on.

And if we take any risk at all, we need to hold the capital in reserve to be able to take that risk.

Right now, we do not, which is why we can be fairly capital light.

And so focus mostly on the distribution aspect of the business and acting as sophisticated frontline underwriters.

We're being actually pretty prudent about the types of risks that we do send to our carrier partners.

You can think of insurance as 3 discrete businesses, like distribution, which is a business in and of itself, underwriting, claims servicing, and regulatory overhead.

Folks who are attacking every aspect of this at once are probably not going to get any of them right and burn a lot of money right out the gate.

And those are the folks that have tended to raise hundreds of millions of dollars to get to that point.

I think our goal right now is to make sure that we're continuously iterating on the best possible front-end experience, making sure that we're proving out that the product itself is a good way, a mechanism to filter out bad risks, and then move into underwriting and maybe the rest of the stack.

We acquire exclusively through our apps, so the App Store, Google Play.

So it's just highly rated, so it's all search?

We were the beneficiaries of pretty decent ASO very early on.

We run a lot of paid, but we take, generally speaking, a portfolio approach to growth.

We would acquire via all the channels you expect, Facebook, Instagram, Reddit, what have you.

That'll probably be a fair material entity for us.

And then we build things that are what we consider to be pre-insurance.

And so these are tools that we think are useful to our customers, irrespective of whether they buy insurance from us right away or not.

A good example of that is in the cover up, price drop alerts.

What we can do is just programmatically re-market you every time we see a violation fall off of your record, or a claim fall off of your record, or you're coming up for renewal.

That's a service you don't need to pay for, but it's pretty valuable because you don't have to think about

Another example is, for certain markets, we're giving away access to a defensive driving school.

So if you go to driving.cover.com, I'll give you a very stylistic example.

In Texas, if you go to a defensive driving school, you can get a violation knocked off of your record, or get access to an insurance discount.

So it can be material, it can be like 10% of the overall premium, so we're talking about hundreds of dollars.

Usually there are entities that charge anywhere from like 25 to 150 bucks to go through that course.

We're a technology company, We just build it and give it away for free.

And that can be funneled directly into our products.

In Texas, there are like 480,000 people over the course of the last 18 months that went through a defensive driving school.

If you think about, hey, where can I be, where are my customers, and how can I build product to facilitate a transition of those customers from those sources to the actual apps, You can build fairly durable distribution?

Paul Woodward And were you thinking about these kinds of programs when you were in YC or is this just by talking to your customers, figuring it out

I think it was observing the types of things that folks were asking, Like, do you have a defensive driving discount?

They will literally ask us, do you have a military discount?

And we're like, well, maybe we haven't considered this and we should go build something to support this.

The other thing I do is I read a lot of insurance rate filings.

But I get a fuller appreciation of the way that 1, the filings are structured, but also discounts that are, generally speaking, available to our customers.

And how we can build product to make it easier to access those discounts.

And so yeah, when someone might be shopping around, they're going to start comparing these things, and this fringe benefit might matter a lot to them.

Guy Co, for example, has this weird affiliate discount where if you own a share of Berkshire Class B stock, you get an 8% discount on Geico.

You mentioned cover.com, and I forgot to put this in my notes, but you guys just bought this domain fairly recently, right?

We've never talked about this on the podcast, but you spent a lot of money

See all Y Combinator transcripts on Youtube

Karn Saroya on the Capital-Light Way to Start an Insurance Business