Glenn Solomon: Max, you’ve talked about the fact that millennials and young people are less enamoured with financial institutions of the prior century, and are open for new solutions within financial services.
Affirm wants to be that next-generation solution for young people, in an efficient and honest way. You’re up against a well-established banking industry. What makes this industry ripe for disruption?
Max: There are lots of things that make it ripe for disruption. Finance is probably one of the largest markets in the world.
Think of it this way. If we can capture even 1% market share of those traditional financial institutions, and convince them to go with a start-up like Affirm, we’re talking about billions of dollars here. The finance industry is a behemoth. But, we want to make a difference even if it’s just for a small percentage of people.
The second thing that makes this industry interesting and ripe for disruption: A vast majority of banking and even non-banking institutions in the US are not software developers. They might have a combination of software packages. Your average bank will have about 100 software systems that were provided by different third parties. They’re sitting on every imaginable mishmash of old technology – Python, JavaScript, fourth D, Lua, Haskell, and Erlang etc.
But what happens next? If the bank wants to introduce a new product, one of the decisions that they have to make is Where would we build that? What will we have to change? Clearly, these traditional, legacy systems would require resources that may no longer be supported. Basically, these crazy combinations of software decisions make these companies incredibly vulnerable to disruptors.
Today’s disruptors enter the industry with a brand new technology stack. They build everything in one system, one language, with one set of frameworks. In Affirm, if anything needs modification, or if we need new features, we can literally call the person who put this stuff together. The ability to execute on the software side very quickly is really powerful.
The third thing that makes the industry ripe for disruption is the general mission around banking. If you’re a top comp science graduate, chances are you’re not exactly looking for a fat paycheck. You’re thinking, what is the coolest thing I can work on? My computer science degree is a key to any door, so flying cars, or AI, or something that inspires me. And the one industry that you normally don’t think is truly inspirational is the banking industry. That’s why fintech is such a disruptor. Fintech companies have an inspiring mission that makes you want to wake up in the morning and change lives.
Glenn Solomon: Fascinating. To be honest, I’ve been wanting to ask you something. Back in 2010 when we invested in Square I asked them how they were able to manage frauds.
His one-word answer was ‘Max’. Can you explain a little bit more?
Max: Keith is too kind. There was an infamous email I wrote to Elon Musk many years ago during our time at PayPal, the subject line was ‘Fraud is Love’. I wrote that email because we were all freaking out about the fraud numbers we were seeing, and all of us were trying to figure out how to beat this.
Personally, I really like tackling fraud. I’m a huge fan of spy novels and doing this is the closest thing you can get to being a spy – if you’re a Fintech/payments nerd like me.
But yes, I helped with fraud management in PayPal and in Square as well.
Glenn Solomon: How are you dealing with fraud in Affirm? Any new developments there?
Max: Yes, definitely. Part of the reason why fraud is so interesting is that it changes all the time. One of the truisms about fraud is that it’s hardest in the early days of starting a company.
Why do I say that? In the early days, the so-called ‘bad guys’ look at you as new targets and attack you. But as you get bigger, the fraud problem stays the same. It’s not growing proportionately to the size of your company.
Regardless, fraudsters are a tricky and creative bunch. They come up with all kinds of wacky new ideas.
During PayPal, the notion of synthetic identity was a theory. The theory was that bad guys will figure out how to get themselves a human-created whole cloth by registering a social security number. But we thought, that’s too much work, probably won’t happen.
I don’t think we ever saw that synthetic identity during the PayPal years – during my time there.
Fast forward to today. Synthetic identity’s a huge issue. Literally billions of dollars lost to humans taking out loans, where the human doesn’t exist. Bad guys have figured out how to create these identities at scale and let me tell you – scalable fraud is the scary kind.
One person figuring out how to steal $100,000 is very unpleasant. But your business can recover from that. If someone figured out how to steal $10 every second, you’re out of business.
There are lots of interesting new stories about fraud. A group of criminals figured out how to temporarily hijack calls and texts from a selected phone number, which is very specific to a carrier. So for example, if I call you, your phone wouldn’t ring – it would ring in someone else’s phone.
Glenn Solomon: Right. So that person can be the second-factor authentication.
Max Levchin: Yes. And that’s a great way of circumventing it. Normally, two-factor auth is pretty bulletproof. Most carriers actually have systems in place to catch that and deal with it very quickly.
But for a time, one carrier had a vulnerability where they’re just in an opening for about 10 minutes. Once the phone was hijacked, 10 minutes can be a long time.
A lot of damage can be done in 10 minutes. You can write a piece of code, you can do a lot of things in 10 minutes. So that was interesting. We lost a few drops of blood on that one. We caught it very quickly and we help the carrier fix their problems. But that was a definitely rude awakening.
Synthetic identities continue to crop up in all kinds of exciting ways and most of them are unfortunately money-losing ways. There’s undoubtedly someone, somewhere, probably in their kitchen cooking up a way to defeat two-factor auth, including, but not limited to things like one-time password generators, which haven’t been defeated – not yet anyway. Someone’s going to figure it out. Wow. It seems like my virtual eyes are lighting up on this podcast. That’s because, despite the fact that it’s so annoying, it’s also really fun to work on.
Glenn Solomon: This is what payment geek fun feels like.
Hans Tung: We have heard from many sources that the checkout processes are resolved by Affirm is truly differentiated.
You also have a lot of competition in this field. How have you been able to continue to create a differentiated product? And what does it take for you to stay ahead of your competitors?
Max: Two different things.
In general, the company is founded by four software engineers, so the DNA of the company is structured so that we should be able to out-product, out-build everyone else out there.
We’ve always had a robust engineering approach. We have always been very good at figuring out new features, new products, new ways of delivering our experience to the customers that are truly different. We’ve been able to measure, test, fine-tune, and iterate at a much higher pace.
One of the things that we did early on was spending the first two years building a full-stack from scratch. Everything from the general ledger all the way to the underwriting systems is built internally. It allowed us to be very nimble with what we want to try, how we want to build it, and how we instrument and optimize it.
The second reason is actually very powerful. We stick to our mission and actually mean what we say. We won’t screw over our customers nor take advantage when they make mistakes.
For example, even though we’ve lent out billions and billions of dollars, we’ve never charged a penny of late fees, which is very controversial in the industry. Ask any veteran in this industry and they’ll say – You guys are leaving money on the table. Late fees are a profit centre, go for it!
Though that may be true, absolutely nothing beats the brand value you get when a customer calls your call centre and says I’m late by a day, I’m so sorry, what’s the late fee. The rep says, Oh, there’s no late fee, don’t worry about it. They’re gonna be like Holy Mother of God, I’m going to tell all my friends about you, right.
So yes, while we are losing out in that sense, we are actively creating something that competitors cannot compete with. Look at it this way: Late fees charged by traditional institutions are optimized into massive profit centres, and most card-issuing banks in the nonprime territory derive about half the profit from late fees. Now, if they wanted to come into our space and say we’re gonna be just like Affirm, we’re gonna cut the late fees out.
Honestly? That can’t happen. They’re going to tank their stock price because they can’t take the P&L plunge. All in all, the mission of not screwing our customers over – that’s been actually an incredibly valuable thing for our ability to compete.
Merchants are quickly wisening up to this new order, I must say. Six months ago, Walmart made a fairly public stand against deferred interest. Deferred interest is a particularly noxious creation of the traditional financial services industry. It’s where you’re promised a zero-percent loan, but in small print, it says that if you’re a day late or short, the interest becomes 29%. However, it compounds from time of purchase which is ridiculous! For example, you could take a 12-year interest-free loan on say, a mattress. If at the end of 12 years, you have not paid up the last dollar, it will have compounded at a 29% rate for 12 years. That’s one easy way of finding yourself near bankruptcy.
I believe this product should be illegal, but that’s my personal opinion. Although I’ll be happy to share this with any politician. Obviously, keeping in line with our mission, Affirm certainly wouldn’t offer anything like that.
Going back to what I was saying: Merchants are starting to realize that these aren’t products where you want your good name and your brand to be associated with. Since most of these traditional financial services products are white labelled, or the merchants’ names are on the loan.
One of the things that we did from the very beginning is made it very clear that we’re always going to have our name fairly prominently in the conversation with the consumer, not because we have to supersede the merchant brand. But it’s because we want the consumer to understand that we’re different. If we treat you badly, your quarrel is with us. It’s not the merchants’ fault. We screwed up and we have to hold ourselves to a higher standard. This has worked pretty well for the Affirm brand.
Hans Tung: Some merchants have told us that sales conversion with Affirm has increased by as much as 30%. How were you able to do that?
Max Levchin: That’s primarily due to iteration. Obviously, there’s also a secular demand for the products. Otherwise you wouldn’t see a 30% increase in volumes, conversion, or ARV.
But people clearly have responded to our product offering. They think it’s good, it’s fair and it’s transparently priced. We’re able to do this because we are first and foremost a software engineering company. We could continually optimize it to higher and higher numbers. It could be even higher than 30%. This is because optimization and iteration is core to our job.
Hans Tung: How fast do you iterate?
Max: We ship code at least once a week, typically a few times a week.
Hans Tung: On the consumer side, you have several million consumers now using Affirm and your NPS scores are considered high – in the 80s. It’s very difficult to achieve for a FinTech company. How have you been able to do that?
Max: Just sticking to our core values. We’ve never actually set a goal of increasing our net promoter score.
I’ve always believed that if we treat the customer right and do the right thing and stick to our core mission, it should speak for itself. By the way, we use a third party to serve our customers. If the score dips, the question we ask is ‘what happened?’ and not ‘what can we do to bring it back up?’. There is a difference. The difference is that we want to locate and fix the problem.
I absolutely believe that a good product should have a very high net promoter score. People using Affirm’s products will hopefully say that was amazing. Because they didn’t feel screwed or that there was a hidden fee. It’s on us to figure out how to keep our core values weaved into everything we do. But we don’t do anything extra to have our net promoter score high.
What we focus on is doing the right thing for the customer.
Glenn Solomon: Max, you’ve also talked about data being a real competitive advantage for Affirm. You mentioned you loan billions of dollars. You must have a growing corpus of data with millions of customers. And I assume that data is growing rapidly now. How do you guys use data as an advantage in your company and when you advise startups, what do you tell them about use of data?
Max: Data is definitely the lifeblood of everything we do here. But we have to be very smart about what we store and how we store it. Obviously, one way of disrupting your high net promoter score is not being careful about data collection and protection.
So far, we’ve been very good about it.
I’m just going to quickly tackle this question since it gets asked a lot: How are we able to underwrite better than the average bank? A typical bank uses a FICO score, but we don’t. So how is it possible that some of these startups are able to underwrite so much better? Could they just be giving loans to anybody who’s asking for it?
The way we’re able to do it is actually through data. One, we look at the standard, typical data the average lender would look at. We tend to buy raw data, and then mine it for features internally and build our models as opposed to relying on external scores. This is how we gain a little bit of a competitive advantage via data.
The most interesting data collection that we’ve gathered over the years actually comes from our own history. The information we gather around our users’ choice of purchase and pricing information. We get skew level data for merchants about what it is they’re actually buying and under what circumstances.
Sometimes we know things like what marketing program led to what kind of a purchase, how the pricing impacted a person’s willingness to convert, how they went about choosing the term.
I’ll give you a scenario. Typically, as an Affirm customer, you get to pick a loan term of 3, 6, 12 or 6,12,18 months. There are many choices on the duration. The duration is based on a principle: That this person is able to safely repay given our ability to estimate their cash flow, which is fundamentally what we’re trying to do every time we lend money. Other answers that we try to glean through data would be: Is this person who they say they are? Are they suspicious, are they trying to rip us off?
We don’t have human intervention. Everything is done through machine learning. And data is what feeds that.
Hans Tung: How have you seen yourself grown as a leader, from PayPal to Slide, and now to Affirm?
Max: I used to freak out about having to speak publicly. I’m a better communicator now.
I’m also more empathetic to my employees. This came about because I had realized I’ve never really worked for anyone, even during my time at Google. I’ve never actually had a relationship where I had concerns about my ability to make my next promotion or worry about my compensation, structure, and fairness. But to be a good manager, a good leader, I needed to empathize about the people on the ‘other side of the table’. Trying to understand their fears and relating to it is something that I’ve learnt at Affirm.
The second thing surrounds core values. At PayPal, we never actually sat down to write our core values. Our core values were fundamentally commercial in nature, things like – We are building the new world currency. In some ways, at that time, it was enough.
But at Affirm, I wanted to do things differently. I wanted to define the company is beyond its commercial aspirations. So we tried doing that. By the time I was done with the first iteration of values, I realized how valuable this was. It’s almost cathartic as you cycle through what matters and what doesn’t. You start to be clearer about how the company is going to be run, what kind of people you want to work with. You start to realize that people who try to bend your core values are probably not going to survive. Core values really matter, and I enjoyed the process of getting there.
Glenn Solomon: Did you come up with the core values before founding the company on your own? Or was it a team effort?
Max Levchin: Well, it first started out with my co-founders. I wrote down the original core values with them. As soon as we started hiring people, I made a point of publicizing our core values saying: Here’s what we stand for, here’s what we will and will not do.
Then came the questions. People asked questions: You have this core value of transparency, What does that really mean? Is salary known to everyone? Can I sit in on a board meeting? We had to quickly differentiate between a core value and an outcome of a core value. Therefore I ended up writing Affirm’s Operating Principles, which were practical ways on interpreting the core values. It was like writing the gospel of Affirm. I must say, it’s a little bit of religious experience.
I’d make people read it before they join Affirm and to say if they were willing to abide by the rules. It’s actually public. If you search for Affirm Operating Principles, you’ll find it. It’s been whittled down since, to be more succinct, but the essence of it still exists.
Glenn Solomon: As an immigrant, we see you’ve prioritized diversity and inclusion as a core preset here at Affirm. It’s definitely a hot topic among startups these days. How have you implemented this and what are the benefits?
Max Levchin: I think wanting to present good DNI numbers is not a priority. It’s valid to want those numbers to be be good. But fundamentally, the business need behind this is attracting the best possible team, and part of that is to get diverse opinions, backgrounds and experience. The diversity is so important, as opposed to hiring everybody who looks like they went to the same school.
Sometimes the best person for the job is not necessarily where you think the best person may be. Go to schools where you don’t normally find a lot of other tech companies looking for talents, because that’s just not the obvious place. I’m not saying that you need to lower your hiring standards and compromise. I’m saying that we need to expand our boundaries of where we’re looking. You need to be able to say, I found you because I was willing to go to some school or some place or some area where people don’t even travel here, because there’s obviously no computer scientists.
Affirm has been very successful recruiting from underrepresented groups, schools and demographics. Each time I conduct campus recruitment, I try to make time to sit down with not just the computer science clubs, but computer science clubs for black students. There’s some amazing jungle out there, and they are so talented. Once you start showing these groups that this is a place where they’re welcome, people will naturally want to join.
We’ve also just announced one really cool female executive and we’re about to announce a few more. I’m doing especially well on my male to female executive ratio right now. I’m quite happy about that.
Hans Tung: Another question on culture that’s been a trending topic in the Valley. Can you work smart or hard? Are they mutually exclusive?
Max Levchin: Founder/CEOs generally work 9-9-9 (9am to 9pm, ‘9’ days a week). That is to say, the amount we work adds up to a couple extra days. I get out of bed at 4 of 5 a.m. and get home by 11 – that’s a comfortable work day to me.
Over the years, I’ve definitely recognized that facetime can be crap. You could be spending hours in front of the computer generating useless code, product specifications or spreadsheets. So yes, I definitely see the value of working smart.
Here’s an example. We were building a product in the middle of the Christmas break with a bunch of engineers. They basically locked themselves in a conference room and worked as hard as they could for seven days. It was necessary, it was hard work, but it was worth it. As their leader, you have to be prepared to lead.
If you’re going to expect that level of hard work, you better be there with them.
But I do think that 996 is not the solution to every problem. It’s not sustainable. There will be times where something needs an all-hands-on deck approach, and the expectation is for everybody to pitch in. Other times, it doesn’t make sense to force people to work through the weekend just to train that sort of ‘self-harm muscle’.
Glenn Solomon: Now we’re gonna ask you a couple of quickfire questions. Who’s an entrepreneur or a company that you admire most?
Max Levchin: Tobias Lutke at Shopify. He is freaking amazing. Yes, I know – the obvious answers are obvious. But I think he gets a lot less fanfare than he deserves. He’s a stunningly successful entrepreneur and very strategic, very smart, super unassuming.
Hans Tung: What’s something you read recently that you would recommend?
Max Levchin: I read a lot of trashy spy novels, and I don’t recommend them. But they’re great to take the mind off the pain in entrepreneurship.
I read a lot of psychology books and just finished a book called Habit. This was a good read, explaining how the brain forms habit patterns and how you can rewire those habits.
I’m also re-reading The Hero With A thousand Faces. If you need a long book to blow your mind and not related to Silicon Valley at all, it’s a great book. That’s what I’m reading right now.
Glenn Solomon: What’s one piece of advice you’d like to give to founders?
Max Levchin: It’s a bad idea to start a company alone. It’s really, really lonely. If you’re a sole founder CEO, you won’t know how lonely it is until it’s too late. And then it can be just a very dark place to crawl out of.
If you’re doing this with someone you don’t know, think of it as choosing a life mate. You don’t know how long you are gonna be together. You don’t want to find out that you really dislike this person. It could be over trivial things like, terrible personal hygiene. But the worst thing you can discover is misaligned ethics.
My advice? Don’t do it alone and get to know the person you’re doing it with.
Glenn Solomon: Great. Thank you so much for joining us. This was fantastic.
