Scott Cook - Founder and Chairman of the Executive Committee, Intuit | Khan Academy
All right, I think we're ready to start. Anyone who wants to—anyone else wants to join us for the talk with Scott Cook, founder of Intuit? So I'll just start. You know, for everyone here at Khan Academy who doesn't know both Scott and Cigna Cook are, you know, some of the earliest kind of believers in Khan Academy and supporters. To a large degree, you know, Khan Academy wouldn't be where it is right now without a lot of their help. And so, you know, I just feel like I owe you guys. But with that out of the way, we're just fans; we cheer from the stands. The players who win the game are in the room.
Very nice. Thank you. The point of having you here—and we know whenever we have people of your stature come to the office, we think it's just a valuable learning experience to have a conversation with you and then to see what's going on in your mind. The place I like to start before we open it up to the rest of the team is, you know, you see a lot of folks who've done big things like start Intuit. When I was a kid, I always wondered how does that happen? How do you go from being, you know, having a normal job and one day you start a company that company becomes a really big company?
So tell us a little bit about that story. How did Intuit get started?
Well, I think for you it was your niece, was it?
My cousin.
Cousin. Yes. For me, it was my wife. Very good. She was struggling with her math problem. No, she had to do—she does our joint checkbook, and she was complaining about doing it. She was very good at it; she's diligent and mathematical. But it was a hassle because she had to pay the bills, and she had to rewrite into the check register, reconcile, and she complained. I thought, "Hah, this was in 1982 when PCs were just starting to explode." I said, "Weymouth, this sounds like the kind of problem everybody might have because everyone's got to have a checkbook; everyone has to pay bills, and PCs are now going into homes. This is very solvable on a computer." This is the kind of task that computers are very good at. So I thought, "Maybe I'll write this software just for fun," but I hadn't programmed a decade. So I figured maybe I should get somebody who actually knew what they were doing. And so I found a student at Stanford, Tom Perrot, and he—I co-founded the company with the goal of building a simple piece of software that allowed people to eliminate the hassles of managing their daily monthly finances.
And how do you go from that? Because what you see a lot of in Silicon Valley and other places is you saw a need, you were able to build a prototype to do it, but then there's a big jump between that and actually getting traction, actually people using it. How did you make that leap?
Yeah, for us, that leap between launching it and actually getting a lot of users was not a leap; it was a chasm. And it was a crevasse. Software at the time was sold in stores in boxes, and you took money. We tried to get VCs to invest two million dollars to give us the marketing money to convince stores to carry it and convince people to buy it, and no VC said yes. They all said no, including VC firms run by classmates. Shows you how unpopular Cydia was, so that was a big problem. We borrowed some money; we got a little bit of seed money from some relatives. We stumbled along; we spent that. We ran out of money; we had to give back rented furniture and rented computers. We stopped paying salaries. It was pretty ugly.
And why didn't you just—I mean, you know, it was stressful. Yeah, our marriages broke up. And I mean, given that, why didn't you just cut, you know, cut your losses?
And I think cutting losses would have been the rational thing. Actually, it did look so hopeless, but I felt I was kind of boxed in because we borrowed this money from some people. I'd taken my father's retirement money, most of it. We had two relatives who invested equity; it was kind of angel money, but I felt I had to pay that back. And I looked at the amount of money I'd have to pay back through honest work if I had quit the business, and so I felt trapped. I can't give up because then I got to pay all this stuff back. So we just kept working; we just kept working at it.
And at some point, how were you living? I mean, how were you paying your bills?
Forcing Cigna had a very good job. In fact, in either this building or in one of the adjacent ones, she worked in a software company, Software Publishing Corp, that had a bunch of these buildings. So she was VP of marketing, so she had a very good job. It was her income that paid the bills.
And then what? So how long did this, I guess, this crevasse—how long was this period, and what was the moment where, actually, obviously, you saw the light at the end of the tunnel?
Well, it's actually even worse than one crevasse; there were two. Something we tried started to work, and then it crapped out. So that's even worse when you get left down. Let's see, we launched the product with great hope in October of '84, and it was the time when I finally felt the thing was doing well enough that I could buy something that wasn't a necessity, something optional that I could splurge on, which turned out to be a CD player. When I finally could afford a CD player, it was in late '87. And if that still was— but that was—and so this was a story; I was still, you know, it wasn't buy—it wasn't like a hockey stick at this point. It was still slow; it was dead.
And so the question you ask—let me answer directly the question. Haskell, so what actually did turn the business around? It turns out it was word of mouth because the way we built the product was different than others had. Nobody up to that point—software is an industrial product sold to corporations. You know, the big software success was Visicalc, which was sold to finance people in corporations, and the UI sucked. You couldn't figure it out without training or reading a book, but that's why software was—we were the first people to do consumer testing, usability tech—we didn't call it back then—but usability testing. So we brought housewives in from the Junior League because, at the time, there were all these women who didn't work during the day, and they didn't use computers. We put them in front of a computer. So I gave them our stuff and said, "Okay, use it to pay a bill." And they'd fumble and goof, and we'd look at it, "Oh [ __ ], oh, oh yeah, now I see." We go back and redesign it, and we try it again and refine it and have people who didn't know the product—no manual—try it by hand. So we had done more cycles of usability refinement than nobody had ever done that before. So our product was actually usable and fast compared to—it. And we're selling to people in homes; no executives telling them to use it. They've got to want it. And so people started using it from our earlier and feeble marketing, and then they'd start telling their friends and start telling their friends, and they started to be demand-pull. And it was the word of mouth that ultimately took off. And then it just is—the rest is history—we started tripling every year. We tripled in size and went into the most kind of obvious adjacent spaces: taxes. And, well, it was—you know, I've learned something about surprises. In fact, the first question we talked about today was what's been surprising. I told him about our burglary situation.
When, yes, I laid out, yeah, there's learning and surprises. So we had—when we ran our first—so we launched Quicken, and even before Turk Hoffman was just, you know, failing and bumbling along, we did a little survey of the few users we had. You ask all these questions, and one question made no sense. It was a—we, you know, at the end of the questions where we were asking about how do they use it, whether they like it or not, etc. At the end, there was a demographic set. So you asked age, location, income. And we asked a question, "Where are you using the product: home, office, or both?" And half the people claimed to be using it in an office. Now, this is a home product. So we thought, "Maybe they don't have a home computer." I mean, this was '84, '85—maybe they're taking their stuff to work and doing it there, so we ignored it. Every subsequent survey had the same answer: half the people were using it in an office.
Now we thought it didn't make any sense. Finally, it started bugging me. Why are people answering this question wrong? So I said, "Well, let's ask those people what they're doing instead of ignoring it." This was years later. So we went in, actually Colson, and asked them, interviewed them, went to visit them, see what they were doing. And by God, a lot of them weren't doing their home stuff. They were running a business on it. Now, we built this as a checkbook manager, not business accounting, and the mindset everyone had is at business accounting. Talk to anyone who's an accountant here. Any accountants? Okay, now it's very good now. And what's the system by which you keep books for a business? In accounting, want this double-entry accounting with debits and credits, journals and ledgers. But any expert will tell you that is the only way. And we assumed that.
So here we found these businesses using a tool that had none of that—no debits, no credits, no double entry of anything. And what? They're using a throw of it; they are running their business on it. It's where all the expenses and the income flow, they're—it's their accounting system. And then we thought, "Why? Why are they doing this unexpected thing?" And it now—it's obvious in hindsight. But of those of you who took accounting in school, raise your hand if you took accounting at school. Okay, how many of you liked accounting in school? Yeah, you and I were both weird, but the rest of you are normal; most people hate accounting. And the people who keep the books in a small business tend to be the owner, the owner's spouse, an unlucky clerk or assistant office manager. Nine times out of ten, they do not know accounting, and they do not want to learn it. To them, the general ledger, which is a feature of all accounting systems, is a world war II hero. So that's why they were using our product, because we were in English and they didn't have to learn anything, whereas all the others had a massive learning curve.
So I finally said, "Well, what would happen if we actually built a product for businesses but made it on the same promise of no debits, no credits, no journals, no—just put it in plain English? You see an invoice; you fill it in. You see a check; you fill it in." So we worked for that on that launch. That all came because of a surprise that we ignored for years, has created a business. And if you add all of our small business stuff—we've added payroll and payments and all that—it's about 15 times the size of the Quicken business. So I didn't recognize it at the time, but there's a pattern about when there are surprises, big upside or downside, savor the surprise. What's really going—what explains that? That's not what we expected. Stop and savor. It's too easy to keep going. All right, well, that's a lesson learned multiple times.
So that's the story on some of our subsequent businesses. So now I know when you ask me questions that there’s like an underpinning philosophy; they're not random; they may seem random.
Very good. And now before I open it up, you know, where do you see Intuit going? I mean, it's a much larger company now. You know, how do you, as the leader or the manager, how do you try to navigate that, and where do you see is the big opportunities and the pitfalls now?
The pitfalls are clear. It is amazing how success and scale make you dumb in an organization. It is just stunning the degree to which the creative, inspiring innovation—invention—suddenly starts disappearing as you grow. You add layers, decision-makers, managers, try to get rigorous and do all the stuff that you have to do to be both efficient and to be reliable. Well, you know, people don't want their—I mean, we did launch version of QuickBooks where your data disappeared. Yeah, randomly, after you—all your data disappeared. Yeah, this is a problem; you don't really want to do this. Don't try this at home, kids!
So you've got to get rigorous, and but something happens to—you know, without continued innovation, where I've continued, how do we make the current thing better? How do we find the next problem to solve? You eventually get stale and die. So how do you be both rigorous and inventive? And most importantly, how do you free your most inventive people, who are often quite new to the company, could be quite young, likely our renegades? How do you free them to invent without the board, the organization, boring them? And so we're redoing how we work along the lines of lean experimentation.
The thing I've learned is if you allow normal hierarchical decisions, you have the problem—all the problems happen—well-meaning middle managers. It just slows down. But what if you tell the middle managers your job isn't to make that decision? Well, let the experience experiments make the decision. Middle—your goal is to put in systems, allow your teams and your brand-new employees to be able to take their best idea to solve your mission and run it as an experiment, fast and cheap. The internet now allows us to run these fast, cheap experiments. So there's this whole wave of change called the Lean Startup that Eric Ries has founded. I've not seen any idea enter the common parlance of business as fast as the Lean Startup. His concept of pivot went from novel to trite in two years. I've never seen it's been absorbed at least as language so fast—but he wrote the book about startups.
I believe there are two groups who need the Lean Startup even more than startups: this concept of lean experimentation, and those are large companies and nonprofits, who will benefit even more than startups. So we're trying to operate as a network of startups inside, where people can try their ideas. And we've got coaches, teams, systems. We put people through a two-day experience where people just sign up, and by the end of the first day, they have their idea up and running in some sort of test to test their leap-of-faith assumption—all to uncork that inventive power that is what has created all great enterprises in South and Ally. So that's kind of—and if we do that right, then I won't have to pick a direction; our people will invent the directions far better than I can.
Yeah, perhaps. Absolutely. I mean on that note, I'd love to take questions. And you won't have questions; I have a bunch more. Anyone— we started the company with a bunch of surveys waiting on. I described one that helped us get the surprise on that led to QuickBooks, but I have become a skeptic on surveys. So I actually don't guide teams to run surveys; I tell them to avoid the survey. We found there's not a high correlation, particularly if you're looking at inventing new ideas and new businesses, new products. There's not much of a correlation between what people say they will do and what they actually do. What they say they will do—they'll do about half the time. Well, so that means it's a coin flip. We've had—we did a product in India, which we have, which we surveyed users, and we're trying to figure out how to monetize it. So we surveyed; we expected 20% would pay for it: we’ve been giving it away for free. We surveyed users, and 40% said they'd pay. And then we actually did the test and took existing users and added a charge, and 2% paid.
Just—there's another example. A team working in the payroll division had a new wave for a business to start up payroll. They mocked it up in screens, showed it to 20 payroll customers—these are employers. Someone here runs a system to write your paychecks, so it would do that person zero of the twenty said they'd use it. Zero! Amazingly, they didn't stop to thank God; they didn't stop there. But because of our system of experimentation, they then got coached. We actually had Eric Ries in coaching that saying, and he figured out a way—you could run an experiment in 24 hours to actually see what people would do, not what they say they would do. They ran the experiment, and 58% of real customers did it. If they listened to the survey, we never would have done this thing because they could take it to a test, a test of just a slice of it. They then learned that customers would—we fleshed out, built the rest of it, and we got the fastest customer growth in ten years in that business.
So what I've learned is—trust behaviors; don't trust surveys. Measure and monitor their behaviors, and then right after behavior, you can ask somebody, "Why did you do that?" or "Why not?" and then you might get something close to truth. So that—and do it more as an in-depth interview of the people who just did something or in a test, just didn't do something, and then you might learn something useful from the interview. But I—we when we go out and talk to customers in an interview and observe, one thing we'll do is go and watch customers. I find it's much more reliable to watch people in what they do. And you've got an opportunity; you've got people at home studying or whatever. We could go watch and see actual behaviors. Don't say anything—just watch. You'll see things they would never comment about.
I'm on the board of the Procter & Gamble Company; they make soap, Tide, and Pampers and other things. And in trying to figure out how to improve Tide, they'd run surveys. And one of the minions—you've got many things you ask about—one of which was the box, the packaging, and they asked, "Is it easy?" and the answers in the survey were, "Yeah, it's easy to open." What? Years later, they did some in-home observation, and they followed some homemakers around, watching what they did. And they then did observe a few people opening a brand new box of Tide, and they thought it was easy if you kept a knife or a screwdriver next year. And because women have nails, and they couldn’t open those things, so they were hacking at it, but he—never come up in a survey, but they could see it.
So the one—the most—the thing I'd recommend is to understand customers deeply. Go ahead and watch them. We call it follow-me-home testing. Leads to follow customers home from the store. Not exactly! We’d ask permission so we could watch them start up. Well, let me answer that in two parts: one, the story of how we got into payments and then the crystal ball gazing. There was actually one guy, an engineer, an African immigrant, who, looking at our QuickBooks users, said, "Ah, they really could use payments." And, in fact, what they need is different from what everyone else is selling. So he got our product manager to work with him, and the two of them cooked up this scheme and inserted it into the product, and by God, it took off because they found a problem for businesses who accept payments when there’s no card present. Because QuickBooks is typically used in the back office around invoices, and some customers wanted to pay with a credit card—like over the phone—but there's no card present, and the whole payments industry had been routed around that. The card being the key, so they invented different economics and different approaches embedded in QuickBooks, and so now we have a payments business that's four hundred million in revenue—sizable—all because they found a problem nobody else had solved and figured out how we could solve it.
Now, on gazing about the future, right now you're seeing a hotbed of innovation around payments, and things are becoming dramatically easier to send money, person to person, to send money for businesses or businesses to set up and accept payments. It's a—it’s long been a kind of monopolized, centralized business with a few major payment utilities, such as Visa and MasterCard, or one central bank-run system called the Automatic Clearing House. And now you're—this rampant innovation's happening around, so it's so much innovation. In most of it, I can't tell you how it's going to turn out, but it's going to get a lot easier and cheaper for everyone.
The place we're focused on is still in the back office because everyone else tends to be focused heavily on the point of sale. But the place that's still being ignored is that business that sends bills or invoices out. And, for example, in rent payments here, how many of you pay rent? Raise your hand. Okay, how many of you pay that electronically? Most of the hands went right there. Especially small landlords are not set up to accept electronic payments, and it's a royal pain for you and for them. They've got to hang around the mailbox waiting for the money to come in, because if it doesn't, they've got to bug you. So a small team—how many people? You mean how many people is that team when they started? Two or three.
Two or three folks. That we can—we can fix that problem! And so they invented a system called Spark Rent, which allows the merchants—the landlords—to sign up, and then the tenants can pay electronically really quickly. The landlord can be anywhere on vacation, and they know everything. It sends reminders out. Cheap; you know, it’s had some hiccups, but it’s now the most popular way that small landlords can get paid electronically, and tenants can pay electronically.
So that's just the beginning of the kind of innovation from that. What we want to do is develop a way so that our system will know that you pay your rent on time, so when you try to get that next apartment in San Francisco and there are five people trying to get that apartment, you've got a track record that you can say, "I'm—you can trust me; here’s my spark rent track record." So we think we can then take the data and make it useful to both sides.
You know, when we're at our best, yes, we can pull that off, but that's a real bit; we're far from perfect on that in our payments group. Again, we've had a group that came at a much lower price point. We've—in our tax group, we’re working on right now—you know, you go to TurboTax, and you got to pay your taxes on a computer. But this group said, "Why don't we use the phone? And why don't we use the phone to snap pictures of those tax documents so you don't have to type it in?" Well, management thought this idea wouldn't work, and outside experts said it wouldn't work, but they kept at it, and they've built something called Snap Tax that works.
It’s an app on your phone, and you snap pictures of your W-2s, and it asks if your taxes are simple—and a lot of you guys are; it answers, asks like nine more questions, and you're done. Nothing we have done in and it's at a lower price point; it's easier. Nothing we've done in our company's history has gotten as higher ratings from customers—it's five stars in the app stores—and we get reviews like you can't believe. One person wrote in that she was in the recovery room after surgery, doing her taxes and so happy about it. Another guy wrote in the review in the app store that it was February in bed with his girlfriend, and she's so stoked she might even wear the gift he just gave it.
So—but again, it was a small team that had the freedom to experiment and invent something that was just—that the mothership wasn't working on. So at our best, we do that. There are other times when I wish we were better. This is the eternal priority-setting question, no matter what size. Herein, before we launched, we had more features we wanted to put in than we could fit in. After we launched—and now, as a four and a half billion dollar company, there are much more stuff we want to get to and that people want to get to than we can.
I have two thoughts on that, and they are kind of opposing thoughts. One is make it fast and cheap to run experiments so you can try more things. If it's big and expensive to try something, then you've got to be choosy. If you can make it fast and cheap to test the leap of faith assumption on which a decision depends, then you can try a lot more things. So this Eric Ries is Lean Startup. Khan, we were already running experimentation, and then I discovered Eric Ries on an online video and said, "My God, that's our idea!" He explains it a lot better, and he took it a lot farther. So I, you know, circulate the Lean Startup and look at some of these videos. It really makes sense.
So that increases your capacity to try new things and to try new radical things, because if it's fast and cheap, it can be radical. The other principle is that the actual important things is a small subset of all the things you're working on. What's truly important? Here's an example. So I learned this technique from another CEO to run a thing called a town hall meeting. So we'll get a group of important customers together—let's say accountants, because we sell a lot of things to accountants—and we'll get a group of twelve of them together in a room. And I'll stand up to the whiteboard, and I'll ask them two main questions: "What do you love about dealing with Intuit?" And each person writes down, private, silent work, writes down a list.
"What do you love about working with Intuit?" Second question: "What do you hate about working with Intuit? What's wrong? What do we—where would you piss you off? What are we just missing the boat?" Okay, silently each person writes their own list. Okay, then once they have those lists, I stand up in front and say, "Okay, let's read off list number one," and everyone read their list—one—and I'm up at the whiteboard, and fill out the whiteboard with all the great stuff. And I said, "Thank you, that’s great, that’s awesome. Now let's go to question two. Everyone tell me the lousy stuff," and I write all those ideas down. And people will have real passion about the stuff that's wrong, that needs to be fixed, needs to be changed, and you'll get 30 items up there, real passion from folks.
And then I'll say, "Okay, let's forget list one, thank you for that. Let's focus on list two. I'm going to give each of you five colored stickers. Go up to this whiteboard and put those colored stickers by the things that you find most important." You can put all five on one. You can put one sticker on each of five; you can split them up. What amazingly happens out of the 30 or 35 items that will be on the board—most get no votes at all, not even from the person who was advocating it! The votes are uniformly incredibly lopsided. Now, they're not collaborating with each other, not comparing notes; they're doing this as individual work.
The votes—there's one always—every time I've done this—one big one. And then there'll be a second that gets half as many votes, and it may be a third that gets half as many, and then it's noise. And then what we do in the meeting is say, "Okay, let's dive into number one. Okay, what's the problem? What's the issue? When does it happen? How does it happen? Tell us more; what's the root cause? We'll go now how do we go about fixing it?" And I've been to number one. So basically, you should work with most of your effort on number one, small effort on number two, and forget the rest. Organizations can't do that. Organizations will work on twelve. "Oh, Suzy wants to work on this—boy, that's a good idea from—" They can't—that customer was so passionate and we got so—you'll get twelve, and you'll underwhelm the important ones.
You'll spend much too much time on the stuff that just isn't so. This is the best example I've ever seen how customers actually are much more focused on what's important than companies are. And it's one big one. You know, I tell folks, "What's the one and a half things you should be working on?" And then run cheap experiments so you keep the idea creation because you'll get surprised out of the cheap experiments; that's your safety valve to get surprised.
But that's—that might be helpful. Let’s do a little show of hands: how many people want to understand all the words in the tax law? Raise your hand! How many people just want to get their taxes filed and get their money? Yeah, we're not in the education business—that's your business. We have found, at least when it comes to finance, most people are exactly like you. They just want to get it over and get the work done. If they occasionally want to learn more, we've got a lot of what we call point-of-need help. We just click. There's a little blue underline thing saying, "What is this?" or "Tell me more," and people can click and learn more. When they do that, it’s typically not because they want to learn; it’s because they want to make the decision right.
So even when they're quote learning more, it's only to make the decision and then move on. So we know what you're doing is noble and really important and crucial to the world, but it's not why people tend to buy or use our stuff. Yeah, so why did we acquire Mint? Given our passion for internal innovation, keep in mind—we acquired Mint before we had ramped up a lot of the experimentation work, okay? So we were—we're not as good as we want now, but we were much less good then. But also, we're very eclectic; we want innovation and innovative teams from both sources. So we both acquire and we teach and coach our internal and enable our internal entrepreneurs. So we do both.
We certainly don't—I think it was Bill Joy who once said almost all the smart people in the world don't work for your company. So we want to be where almost all the smart people are, which is, you know, we're only one slice of the smart people. And in this case with Mint, we had an internal effort to do QuickBooks online, and Mint and some other competitors were pursuing similar things. And Mint just outran the whole field; they produced a better solution that had a different business model and was just better than what we were doing. So we said, you know, why fight a losing battle? Let’s go get the top team and the top approach. Since we acquired them, their business has tripled or quadrupled in size; their ideas have infused much of the rest of Intuit.
That was one thing we were hoping to have—their ideas change the direction of other businesses. We've had them run the QuickBooks business so we could take their learning, and we gave them the QuickBooks business to run. So we really tried to maximize the wisdom transfer, and now we're building a Mint product for small businesses. So again, they have such a good thing for consumers, plus we're adding much more to the Mint data and what Mint can do for you. But we're also saying, "Why can we take that same magic and solve problems for small businesses as well?" So that’s some of the story.
I mean, your premise is exactly right: when you do your finances with any system, it accumulates the data. So we've got a very large repository of data on people's taxes, investments, pending business, at a transactional level. The first thing to recognize is, in my view, that's not our data—it's our customers' data. They allow us to be the stewards of their data only if we earn that. So we can only do things that the customer wants and gives us permission to do, and if we break that trust, people will run from us so—and they deserve to.
The next thing is Mint was another piece of the Mint DNA we wanted: they were making better use of the data than anything we've done. We've largely done nothing with big data. We’ve used the data to solve your immediate problem but we have not—then the way Mint has said, "Oh, based on your data, we can see that you’d be much better off if you did X or Y." So that was another piece that we wanted to do to cross-pollinate from the Mint acquisition.
We now have a data team with a new leader; we took one of our leaders who's very talented and added this challenge to her. So there, the data team was focused on how they can help Intuit with data, and we said that's nice, but what we really want is how do we help customers with data? How can we change their lives? So we reoriented the data team on the mission that you described so that hopefully in a year or two, we’ll have more ways that we can help people.
Here’s just one that’s already in action: it turns out most small businesses apply for a loan at some point—seventy percent of our QuickBooks customers have applied for a business loan, and most small businesses get turned down in a rather slow, long, ugly process where they wind up giving a lot of data, information, and paper to some bank that ultimately turns them down. Plus, there are a lot of lenders who'd like to lend to customers who they don't yet know. We think we can make that happen because of the rich data in QuickBooks. We can help figure out who is a good lender for you and then make it easy for you to apply with a push button instead of filling out all this paper.
And then have a lender be able to rapidly improve and get higher confidence that they're lending to good people because we've got deeper data on your business than anybody. We can only use it with your permission, but this is an area where businesses really want someone to take advantage of their data to give them a better loan at a better rate. So we've got a lending operation; we don't do the lending—we work with lenders—and we're trying to pioneer this lending marketplace using QuickBooks data as the kind of unique asset to make it happen. So that's an example, and we're getting some number of lenders are playing with it.
And the ultimate goal is to build a better credit score based on richer data so that more deserving businesses can get better loans. So I'll just finish up. You know, obviously, you are known as one of the top managers, and not only your industry but in Silicon Valley more generally. And you’ve—you've known Khan Academy since really the beginning. You know, what thoughts do you have about what we're doing? Where do you see the opportunities? Where do you see the pitfalls? And what are your general thoughts and advice for us as a team?
I bet some of you are better aware of the things to be afraid of. I'm not sure I'm close enough to give you a thoughtful assessment of that, but I think one thing that would—I can't say is an issue here but has been an issue in other fast-growing operations. So you grew from about five to six million monthly actives to about ten million over the last year—near doubling. And even, you know, you're expecting a future like that. You're superstars, celebrities, in the press; there is the curse of success where things are on a roll. What everyone wants to talk to you is somehow is a drug that causes people in other organizations to take their eye off the ball, to stop focusing on two things: one, what really solves the customer problem?
Yeah, a customer’s using things are great, we don’t have to worry about that! But companies can persist in doing things that ultimately are second rate just because the second rate stuff was better than what was before, and they don't find out the break. And we think of Overstock.com, who then just got surpassed and crushed by Google robots, by eBay. Or think of Yahoo, which gosh, they were the darlings, huge—the internet company—and then Google just went by to buy them because they didn’t—they stopped focusing. Okay, what's the major problem to solve and how can we innovate and solve it much hugely better?
And Yahoo always was a place people came to to go somewhere else, and their search was not good. So stick to the knitting of what's the funnel problem we're trying to solve, what do people really want us to solve, and how are we the best in the world of that? From what I see you're working on that, so with the dashboard and that.
And the other curse of success can be the curse that Friendster had. So Friendster was the social network but for Facebook. And even before it's the music one out of L.A., MySpace, and they were had taken off as a social network. Kleiner Perkins major VC had invested; they asked me to go down and take a look, talk to the entrepreneur, and that was good, but he talked about the end of the talk—about some technical problems, and dove in to learn more and he wasn’t gonna long with this technical founder or technical guy at all, and the system was crashing.
It just couldn’t keep up with demand; it was collapsing, and ultimately that killed Friendster. One of the things that Mark will say about why Facebook succeeded is he had seen the collapse of Friendster because of technical bumps; they just couldn’t scale with the volume, and he and Dustin Moskowitz would just do—they were determined that wasn’t going to stand in their way. So they limited growth early on, and then they just piled the resources to make sure they were up and reliable.
So those are some of the things I've seen in other cases of companies running afoul. I'm seeing good stuff here; I don't know—don't know—I don't suspect that's happening here, but those would be some watch outs from the history of some companies who, like you, have been shooting upward.
Oh great—no, thank you! Let me just cover one more thing, which is the other half of your questions. Yeah, which is about how do I see the future? Maybe I'm not very good at seeing the future, but I can see today. And there's something—you may have seen this as an article that was in Harvard magazine about a physics teacher there, a guy named Eric Mazur. Anyone? Yep, very familiar. The article was entitled something like The Twilight of the Lecture.
So Eric Mazur was a well-respected teacher on the physics staff there; he'd been teaching physics for a decade. I believe the tests that he gave showed the students learned subject matter—he's all happy. Then he heard about the University of Arizona that had invented kind of a capstone test that tested, kind of like the Common Core, did people really get this stuff? Could they kind of apply it? And the results on the University of Arizona was after taking the semester of physics, that the physics students there just failed this can you actually apply it in real situations? So he looked—and so now it's interesting that those are University of Arizona students, his Harvard students—almost all of his students, virtually all got our number one score, one or two in their AP Physics.
So he just—as a lark, took the test and gave it to his Harvard students after they had finished his semester of physics, and by God, they failed it. He said it was monkey level. If you had guessed red monkeys guessing randomly, you would do as well. As history. Now, Erik is an unusual professor at a research institution because he actually cared. He didn’t say, "Well, screw that and back to the research." He actually said, "What does it say about me? What am I doing wrong as a teacher? I'm failing these students."
Hey, so he dove in and tried to figure out what was happening. He has evolved from that once I—when I guys passed me the article, and then we reached out to him. We actually spent some time with him learning and he basically found out that students don't learn from lectures—even from highly rated professors—they just don't seem to learn from lectures. These are Harvard students. They don’t learn to—I mean, they learn to pass the basic test, but they don’t learn to apply in a deep way.
So he started trying alternative ideas. He tried a very different—he’s basically flipped the classroom, and the students work on problems. And then he can figure out through a system which students got it and which didn’t, and then he'd have the students who got that section explain it to the students who didn’t. And he said they’re much better at it because the student has just tried it, didn't do it, and now they’re hearing from someone who just learned it, who just is walking in their shoes.
And he says he's now—now his students ace this test. I think there’s a revolution in learning from something that was done to you. Education was done to people; lectures are done to people, too. Turning education on its head—learning on its head—so people now it’s up to me. It's now become something you do, and you learn at your speed, and you learn by doing yourself, not by being reliant on someone else to pour learning into your head.
You learn by using; you learn by doing. That’s the way real humans really learn fast. You guys are in the vanguard of leading that globally with the internet, with the raw inventiveness of the dashboard, the on-demand, the lectures to understand how the students progress, creating a platform so that math, physics, and other subjects can be taught by learning by doing. This will change learning and lifelong learning. This is the invention of printing by Gutenberg applied to a field that desperately needs it. And more than that, this is—the sum total of IQ points in the world is fixed; it's what genes have delivered to us.
But how you turn those IQ points into productive human potential, into skills to be able to solve problems, to be able to see problems, to be—that is entirely a function of people's learning, and can be dialed up or down dramatically by the learning method. Compare when you’ve been in some class which was boring, badly taught, or the great research professor didn’t really know English and your learning rate is so slow. Your love of learning just goes into the toilet. And then compare that to a class where the teacher was fascinating where you could move and learn and try things and they worked and you got this feeling of accomplishment. It’s night and day!
Igniting the fire of learning in billions of people is a force of power far greater than the printing press, and I believe you guys are the world leader in doing that.
Well, that gives us a little bit of pressure, but it's exciting at the same time. And so I'll just, on behalf of the whole team, I just really thank you. That was really a special conversation. Thank you.