Alistair is an analyst, writer, and startup accelerant. He likes Big Data, clouds, and the web. He is currently involved in Strata, Cloudconnect, Startupfest, YearOneLabs, Bitnorth, and Decibel.
→ His background as an analyst. writer, and start up accelerant
→ Finding the most valuable and interesting solution to a problem
→ Creating something that truly addresses a real issue in a significant way
→ His background in Strata
→ Slides created for the book “Lean Analytics”
→ Used in workshops and conferences
→ Used to share insights and best practices for using data to drive business growth and success
→ How does that affect lean analytics?
→ How he defines growth hacking with the Venn diagram
→ And a whole lot more
Bronson: Welcome to another episode of Growth Hacker TV. I brought in Taylor and today I have Alastair Croll with us. Alastair, thanks for coming on the program.
Alistair: Thanks for having.
Bronson: Us. Absolutely. So, Alastair, your Twitter bio says that you’re an analyst, a writer and a start up accelerant. Unpack that for us a little bit. What exactly do you do?
Alistair: I somehow turn it into a career. So I’ve started companies in the past. The one that was the longest and the biggest out was a company called Gradient, which was a web performance management company acquired by BMC in 2011. But that was a ten year story and we learned a lot. It was a tumultuous but exciting ride. I help run conferences. I’m not really sure how I got into that, but events like Cloud Connect O’Reilly Strata Conference, which is the big data conference on the content chair for that, and one in Montreal called Startup Fest, which is now in its third year or will be this year in its third year, which is a week long festival in Montreal around startups and building new companies with people that come from all over the world. And then writing, I have written four books. This Lean Analytics is my fourth written or coauthored, and I tend to write about web performance technology and really the intersection of technology and society and how the two affect each other. Mm hmm.
Bronson: Oh, it’s great. You really. You really have turned ADHD into a career. You got your hands in a little bit of everything. I love the title of your blog. You write Add solve for interest income. What does it mean to solve for interesting because I read the about page on I was like, that is great.
Alistair: So it’s a funny joke. When I when I called the blog software interesting, people thought that was a crazy name and like, you know, go to the bank and setting it up and they’re like, that does that’s not good English. So it’s actually something that’s relevant to what we’re going to talk about today, our online analytics. If I were to tell you that I’m building the next Facebook. Mm hmm. Let’s say legitimately. Seriously. I say right now, Brandon, I’m going to build the next Facebook. You’re probably a little skeptical.
Alistair: Your skepticism doesn’t come from unlike I’m guessing you look at my bio, I built companies before. You think you can probably get some technologists to build the technology, but no one will care. Mm hmm. Right. I mean, yeah, the interesting thing. Who cares? Right. And so the most risky thing for almost every startup these days is attention. Mm hmm. So there’s actually a guy in the 1970s who was an economist, Herbert Simon, who observed that we don’t live in an information economy, he said. We live in an information society. Mm hmm. But what we economy is there to find out what’s scarce. So in, you know, fuel economy, you know, oil is scarce in an energy economy, in an agrarian economy, it’s access to land fuel economy, it’s access to castles, that kind of stuff. And so he said, what information consumes is attention. We literally pay attention. And so we live in an attention economy. And he said this in the seventies, which was pretty good foreshadowing. Mm hmm. Really? We’re in an attention economy. Google has your attention because they can direct you to where you going to go for search and they’ll charge money to do so. Right. We literally live in this attention economy. We joked when we did our last big presentation for analytics that Kevin Costner was a crappy entrepreneur, because if you remember Field of Dreams, his line was, if you build it, they will come. And the lean startup model is if they come, you will build it. Right. So attention is the scarce thing. And so the reason I call it self interesting is that when you have too many things to pay attention to, literally paying attention. Mm hmm. You focus on what’s interesting. So all the good stuff in my life has come from trying to solve for what’s interesting rather than, you know, solving for risk or solving for profit or whatever I try to solve for. Interesting.
Bronson: No, I love it. Like I said, I read it and I was like instantly, okay, I got to put that, you know, it’s a new mantra of some sort for me. I love it. You sent over some slides that we’re going to talk to in just a moment. But I have to admit, when you sent this over last week, I open above and I just instantly thought, I’m looking at something really important here. I mean, some of these slides break down what’s going on in a way that we haven’t seen before and how clear it breaks it down. So let me ask you this. These slides are going to go through in a minute. What do you use these for? Typically, are these things you share with start ups or mentoring? Are they things you talk about at conferences? Where do these slides normally make an appearance for you?
Alistair: Well, when Ben and I finished Lean Analytics and the blog for the website is for the book is Lean Analytics Books. When we finished the book, in the last throes of editing, Ben went off and poured through the whole book and found all the errors and so on. And I made a 420 slide, slide deck and we didn’t use all of that anything. We ever used all the slides, but we literally turned it into a series of slides. Mm hmm. In many ways, the book came from our experience with Year One Labs, which was the startup accelerator. We ran in Montreal, and we learned a lot of stuff and codified some of it so that we sort of built the deck and the book at the same time. So we’ve done a lot of workshops. We did one last year in December at Eric Reis Lean Startup Conference. We actually just announced one that we’re going to do a startup fest the day before with Patrick and Brant, the authors of. An entrepreneur. So we have a lot of content for this. But in many ways, when you’re writing a book these days, especially a book that’s so built by others, like, you know, we have 150 coauthors, really, because we talk to dozens of founders and entrepreneurs who give us great feedback. It’s an iterative process, just like Lean, where you’re putting slides out there, you’re in a room, someone shoots them down, that affects what’s going in the book and so on, right? So a lot of this was built in that in that sort of mindset.
Bronson: Now that’s great. Well, let’s jump into the slides a little bit and I’ll be showing these on the screen to the people watching as we talk through it. But the first one is you talk about how startups don’t really know what they’re going to be when they grow up first. What do you mean by that? And then second, why does it matter? How does that affect, you know, lean analytics?
Alistair: Sure. If you look at most of the great companies that are out there today, especially recent startups, but this is true of many others, they don’t really know what they’re going to be. Right. I mean, IBM is a global consulting company, but it started out making tabulators for the census. So there’s there’s always an element of not just knowing how you’re going to solve a particular problem, but not even knowing what problem you’re going to solve. And this, I think, was the great realization of lean. So if you’re building a nuclear reactor, you know, you need to, you know, the design specs of the reactor like the control panel and you know how to build it. And that’s a very specific thing. And that and that’s where we traditionally had like waterfall engineering for where you have a very well known goal of what you trying to build. Mm hmm. And they’re very known. Very well known way of building it. And what happened in the from the seventies to the to the mid 2000s is that we learned that we often didn’t know what how to solve the problem. Right. So we knew we were let’s see your Amazon. You know, your goal is to sell books online, but you don’t necessarily know the best user interface for selling books. So we would use Agile to kind of build something and then iterate a few times and figure out what the best interface was. Because by the time we built something, the market may have changed a bit. And so we had to be able to tweak it until we we satisfied that. No need. Yeah. The real lesson of lean is we don’t even know what the know need is. Okay, so it’s not just we know how to solve the problem. We don’t know a problem we’re solving. Mm hmm. That’s a very hard thing to understand, right? It’s sort of this Zen, like, kind of. No first that you don’t know. Mm hmm. So if you look at these companies, right, PayPal was going to be a payment system for Palm Pilots. FreshBooks, a very successful billing and sort of mini CRM for small businesses. Is an invoicing. Was the invoicing half of a Web design company that just got fed up sending invoices to people? Wikipedia was going to be written by experts. Mm hmm. All these companies write Hotmail. The founders wanted to build a database company, and this is a fairly well known story. They want it to work from their day jobs on this database. Companies, they built themselves a web based email tool, and the VCs went, Well, your database sucks, but we’d like to see your web based email. Right. Flickr was an MMO and they were smart enough to look at players who are uploading pictures and they were in the picture business. One of my favorite ones of all time, Mitel, is a Canadian telecommunications company founded by Terry MATTHEWS and Michael Copeland. Mm hmm. And they were originally going to make robotic lawnmowers, so it was Mike and Terry’s lawn mower, Mitel, and eventually realized that the idea of robot powered drones driving around the suburbs of spinning blades was a bad idea. And they kind of said, All right, maybe we should get into telecommunications. So all of these companies didn’t just know. They didn’t just not know how to solve the problem they were they thought they were trying to solve they didn’t know a problem they were solving. Yeah. And so we joke in the book, but I think this is a very true statement. You’re not building a product. You’re building a product to figure out what product to build. Mm hmm. And once you understand that, then you don’t fall in love with your product when you’re your baby dies a horrible death, but it’s taught you something good. You celebrate the learning and move on to making another baby. Whereas as founders, we love to build things right. So we love to create stuff. We hate another building. And this is a good way to sort of overcome that objection.
Bronson: Yeah, the lean analytics, does it play into it because you’re really relying on the analytics to tell you what’s working and what not, what’s not working, as opposed to just going with the product because you love it. Is that how the analytics kind of play into this?
Alistair: Yeah. Mark Andreasen said markets that don’t exist don’t care how smart you are, which I thought was I love that guy. He’s so brutal. So the build measure learning cycle of Lean Analytics obviously is a big part of the measuring and learning. And what Ben and I tried to do was get more specific about what should you measure based on what stage of your company you’re at, based on what kind of business you’re in?
Bronson: Yeah, it seems like what you just kind of outlined for us in terms of, you know, not even knowing what problem you’re trying to solve. It seems like a very Internet specific kind of new thing that companies have to deal with, because I can’t even imagine this conversation making sense 50 years ago, and maybe I’m just not being creative enough, but it’s hard for me to imagine a group of people coming together saying, we’re going to start a small business, go get out a small business loan with the bank. And we don’t know what we’re doing yet. We’ll figure it out because we’ll build, measure and learn. Is this just is it that new? I mean, is there any correlation to this in history that you can think of where you just don’t even know what you’re going to solve? But you’re an entrepreneur, so you know you’re going. So something.
Alistair: Right. So I mean, you can take examples. Mystery like the skunkworks, right? Which was a division of Lockheed Martin. As World War Two kind of rolled across Europe, the Americans realized they needed a jet and they put a bunch of smart guys in a tent next to a circus tent next to a paint factory, which is why it was called the Skunkworks. And it smelled really bad and it was kind of prototype. And within ten months, sorry, within a month, had a plan within ten months on a prototype. So in times of extreme pressure or urgency, you can relax a lot of the normal constraints because business wants predictability. Right? But you can look at organizations like Procter and Gamble with Swiffer. So Procter and Gamble, another big organization, I mean, 100 years ago. But but, you know, it’s been around for a while. They’re in the business of selling soap. And it turns out that the market needed was not more soap. It was a better mop. And it took like an outside design agency watching how people picked up coffee. And these guys were really mean. They went up to Grandmother’s and spilled coffee on the floor and watched how they cleaned it up. But, you know, that’s pretty good getting out of the office. And it turns out that there’s a there’s a an important lesson to be learned there.
Bronson: Now, that’s great. I’m glad you brought up those examples. Now you kind of see growth hacking as the intersection of sort of three different disciplines, three related but different disciplines. Talk us through kind of how you define growth hacking with your Venn diagram here.
Alistair: Sure. So first of all, you know, the word cloud computing is probably the most nebulous the i.t organization has come up with has come up with except maybe big data. And I think marketers are now trying to get themselves a pay raise by calling themselves grow growth hackers. So nine times out of ten, a growth hacker is not a growth hacker, they’re a marketer. Growth hacking specifically has a technical aspect to it. That’s where the hacking comes from. And I think that there are some interesting parallels between growth hacking and literal hacking. I joked at an interview event last week that marketers need to stop writing press releases and start hacking their markets. What does it mean to hack a market? What it means is to find a vulnerability in the existing market and exploit it. Just like a hacker finds a vulnerability in a in a server and exploits it a growth market or needs to find a vulnerability in an existing market and exploit excuse me. And that’s that’s difficult. It’s one of the reasons people don’t talk about it because good growth hacks like zero day exploits, you want to save them for something special, right? So it’s really is intersection of guerrilla marketing, which is sort of unconventional marketing approaches, data driven learning and a certain amount of subversiveness. And I really do think that this is an important part of this. I mean, when Farmville grew, it’s because they figured out how to exploit the vulnerability in Facebook, where you could where an app could post things to your friends feeds that was an export. And then Facebook patched the exploit like it literally is hacking like that. You know, for example, sorry, Andrew Chen talks a lot about how Airbnb makes it easy for someone to post the listing on Craigslist. Mm hmm. Well, it’s less well known. Is that years before Airbnb went and crawled Craigslist and sent all of the listing owners a message saying You should just let Airbnb be, that’s a hack. And in fact, in violation of the terms of service.
Bronson: Yeah. And you know, Craigslist has no API. I mean, it’s not like it’s easy to plug into these things. They really are hacks here. You’re screen scraping, you’re finding, you know, loopholes, you’re doing all those things. So I like the way you kind of see those three coming together there. Now, you mentioned that the five things that people need to know and I’ll list these off here and then I’ll let you kind of talk through them a little bit. As we go on here, you talk about the need to understand what makes a good metric, and I love your slides on that understanding cohorts and segments and the difference between them. I’m also understanding AB versus multivariate, those kind of things, the business model flipbook, the lean analytics framework and picking the one metric that matters. And honestly, if no one remembers anything from this interview, if they can remember how to pick the one metric that matters, they’re going to walk away with with with a lot. So let’s jump in here, talk about metrics for a minute and maybe the evolution of metrics. You know, it began with hits, but but why is that broken and how is it evolved and where are we at now?
Alistair: Well, I think for a lot of people excuse me.
Bronson: Now, we should tell our audience, you’re recovering from a cold here, so I’m not doing a troop or helping us out here.
Alistair: That’s right. It’s a good thing I’m not at home with you. You have the plague, too. So the nineties call and they kind of want their metrics back, right? I mean, anybody who put a hit counter on the website, that was a pure vanity thing. Right. And so what’s happened is we’ve started to realize that, like, I don’t want followers, I want minions, I want people who do my bidding. And so if I have a mailing list, that’s nice. If I have a mailing list and I know with certainty that when I mail them, 20% of people act on what I asked them to do. That’s a superpower. So really when you look at metrics, you have to make sure that your metric is going to be something that will change how you behave. That’s one of the most basic rules of metric. If you look at a metric, someone asks you for a metric your receipt and says, Hey, you know, could you measure X? The best thing to say is, well, what are we going to do if exchange of. Their baby had a test they did a while ago where they took photographs of professional photographs of some of their houses. And it turns out that if you have professional photographs of your house, your two teeth, you get 2 to 3 times more and a lot more rental. Mm hmm. That was a great experiment to run. And if you look at where they actually hit their hockey stick, it’s right around that time. So that was a metric what percentage of houses we rented with professional photography versus owner or renter photography. And it turns out it was huge. And they knew if this number is high, we will make this a standard feature. So, you know, when you’re getting a metric, how is it going to change my business? And so vanity metrics tend to be ones that make you feel good but don’t actually change what you’re going to do. Yeah. Mean VCs are often guilty of this and I was at South by a couple of weeks ago and a PR agent came up to me and said, Look, for me, the number of followers is a is a real metric, not a vanity metric. I said, You’re wrong. She said, No, my client. Gets comped by his boss based on the number of followers they have. For me, getting him followers is a real metric because it drives my revenues. I’m like, Yeah, you’re right. And that’s why so many investors have a hard time today, because we really like simple rules, right? And so, I mean, I can tell you how to get 10,000 followers right now. You know how to get to those voters right now. Anybody? All you do is you go on Twitter. You look at accounts that have the same number of followers and followings. And those people are extremely likely to follow back. And you follow them all. Yeah. I mean, it’s why did written out Mitt Romney by followers? He didn’t need to just follow people who follow back.
Bronson: Yeah, absolutely. And that’s why when I see 23,000 followers, 23,000 that I’m following, I know they’re not thought leader even though they have a massive audience because the audience is fake. It’s like, I don’t care to interact with them. I don’t care what they think about anything. I don’t want to get them on this show because the ratio is off. I know it’s broken, but other people don’t see that yet. Talk us through why some of these metrics are broken. You talked about followers and that’s a good one because we don’t we still think that matters when it doesn’t. But other things like unique visitors, I mean, it still has a lot of kind of people feel like it’s important. I think we’ve moved beyond hits and page views and maybe even visits, but unique visitors, I mean, we’re still in that world, I think. Why is that?
Alistair: The metric so unique visitors is actually okay and in fact, Pages is okay in one case. And if you’re displaying ad inventory and we say on the slide because then it’s inventory which is part of your business model. Right. But you still care about are those visitors likely to do the thing I want? So if they’re drive by users, let’s say you’re you’ve got an app that you’re launching and you have like mailbox. Did you get hundreds of thousands of people? Did Dropbox buy mailbox because they had hundreds of thousands of people in the queue. I’m guessing the Dropbox guys are smart in that they probably would. They had hundreds of thousands of people in the queue and a significant percentage of people who get the app use it and replace the regular mail clients. I’m willing to bet if you asked, that’s what matter. So unique visitors is nice, but unique visitors that stick around and are engaged is very useful. So what you actually care about is unique engaged visitors and you’ve got a measurement for engagement like they do the thing you want. They return every day. You know, on Reddit, 91% of people that read it come back the next day. That’s incredible.
Bronson: And huge. Now, that’s great. What about emails collected? Because when I when I saw your slide here, that was the one where I was like, huh, maybe that’s why I need to change my thinking, because in my mind, that’s still a good metric. It doesn’t feel like vanity. In my case, that’s my Oh, but Twitter followers still matter. This is my example of that.
Alistair: So, you know, this is wonderful, but you’ve got to know what your open rate is, right? Remember that the bottom line is you’re trying to take the spreadsheet like you should be able to define your business model on a spreadsheet. It shouldn’t look a lot more complicated than a lemonade stand. And, you know, so I buy lemonade and I sell it for more than it cost me. And I have enough street traffic that people will buy sufficient number of lemonade to pay my rent for my little sister, whatever that business model is. Once you’ve defined your business model simply like that, everyone these metrics probably fills in a cell. Mm hmm. And if it doesn’t, it’s not that useful. But it does. Now you know what to optimize, right? So emails collected is very important. If you let’s say in your case, if you send it out and 80% of your listeners click on the thing where they get an email and you’re listening to podcast. That’s awesome. Mm hmm. Right. But even then, I would say, okay, does the listening to the podcast generate add clicks or does it generate purchase of videos or whatever? And that’s really the challenge here is unless it ties back. So so the reason we care about followers is because we think in the absence of knowing anything else about the person. Mm hmm. We think, well, maybe 10% of our followers will do it. Or maybe 1%. So if I tell you a million followers, that’s still 10,000 people. That would do something I want. Mm hmm. Which is a lot of people. So that’s why, in the absence of anything else, we still look at followers. We got to remember that. That’s nice to know. But it’s only part of the answer, which is, you know what percentage? It’s like if I do how many things that I sell, I also care about what was the value of the shopping cart? Mm hmm.
Bronson: No, absolutely. And that kind of leads into the next discussion, which is, if we’re not looking at vanity metrics, then it seems like what we’re really looking for is correlation and causation. Talk to us about those two things a little bit. What is correlation? What is causality? And really, how do you move from one to the other?
Alistair: Sure. So first of all, things that are causal or correlated are really good. Mm hmm. If you have a metric that’s causal or quality, that’s good. I’ll give you an example of two things that are correlated. Drowning and ice cream are very strongly correlated. Also, this is a true fact.
Alistair: Right. So does that mean that we should pair ice cream so we reduce drowning deaths? Probably not. All right. Does that mean that when there’s an increase in drowning, we should call the ice cream shop and tell it to stock up on Rocky Road? Probably not. Both ice cream and drowning are caused by summertime. Gotcha. So in summer, it’s warmer. You go swimming and eat ice cream. Mm hmm. There’s a strong correlation. You can predict if you see increases in ice cream sales, that you’ll have more drowning deaths. Yeah. So I can predict the future with correlation. And that’s good, because, like a friend of mine in San Diego owns a restaurant called Solari, and he’s a successful tech entrepreneur. He’s had, like, five exits. Used to be general manager of Teradata. So you can imagine he runs his restaurant with a little bit of data. Mm hmm. We actually open the book with this case study because it’s a great example of a non-tech company. Mm hmm. But in his case, at 5 p.m.. His son says, Hey, Dad, we got 50 reservations. Randy knows that for his business, there’s about five times the number of covers that night, a year of reservations at 5 p.m.. So if he’s got 50 people with reservations at 5 p.m., he’s going to have about 250 people that night. Mm hmm. That’s a good correlation, right? Because it allows him to go, oh, it’s 5:00. I can hire an extra an extra waiter for the evening, or I can go out and get some more food. That’s a good thing to do. Mm hmm. But if Randy somehow doubled the reservations to 100 people, does that mean that he’d get 500 people in restaurant that night? We don’t know. Mm hmm. So be very cool for him to run an experiment to try that. Right. Because if it does, then it’s causal. And that’s a superpower. He can fill the restaurant just by getting reservations up.
Bronson: Yeah. Is that how you move from correlation to causation? Is through testing, individual kind of things and seeing what’s the causal relationship, if there is one.
Alistair: And this is the this is one of the things that like when people talk about big data and you don’t need correlate, you don’t need causality with big data. You do. Correlation shows you what could be related. And then you run experiments if you’re doing DNA and you find that a particular gene predicts short sightedness. Mm hmm. Now you’ve got to go do tests to see if changing that gene reduces shortsightedness. Right. And so, in the case of causality, I’ll give you another silly example. Houston Airport. Houston Airport had constant complaints about how long it took the baggage to come out. Mm hmm. Thinking like 8 minutes. And when they surveyed people, the number one source of complaints was it’s taking a long time for the bags to come out. Mm hmm. So they did everything they could to improve the baggage handling system. They got it down to 6 minutes, which is a 25% improvement from the time it gets off the plane until it’s on the carousel. Sounds pretty good, right? It’s a lot of hard work and a lot of money. Everyone still complains. Bags are taking too long, and then they decided to park the planes further from the baggage carousel. Complaints went away. So two things to learn if you’re an entrepreneur. And number one, what was causing the delay was time spent waiting, not duration of bags to the carousel. Yes, right. Number two, if you have two possible experiments and one is re revamp the entire baggage handling system and the other is ask a few planes to park further away and run an experiment to see if that makes a difference. Yeah. Yeah. Look at the two experiments. Which one can you do more quickly and cheaply? Try that. At first.
Bronson: Yes. They assume the causality, like any of us, would probably make it seem too obvious that making that system more efficient was the causal relationship when really it wasn’t. That wasn’t the thing. And if they had tested earlier on, they could have saved themselves the time of, you know, making it 25% more efficient, which was irrelevant.
Alistair: Right. And in growth hacking, I mean, one of the things that came out of the growth hacking conference last year was that all these big tech companies have some kind of indicator, like someone on Twitter who within their first visit follows a certain number of people who within a certain time, follow them back. Mm hmm. Now, that predicts whether they’re going to stick around or on Zynga with somebody plays the game and then comes back the next day to play to. You predict whether they’re likely to become paying customers. If it’s causal, then Zynga should do everything it can to send you emails telling you come back and play the next day or Twitter should do everything it can to suggest people who are likely to follow you back and you can run these experiments. And that’s why I say that growth hacking has the subversiveness and also the data driven angle is you’re trying to find a thing today that correlates strongly with an outcome you want down the road, a leading indicator that you can use to change the future.
Bronson: Yeah. Yeah. But so correlation is not a bad thing. It’s it’s getting us on the right track possibly, but it could also be completely irrelevant at the same time. So correlations a good starting point, but it’s not the endgame.
Alistair: Well, correlation is a very useful thing. Right. You can have a business is working great people like your stuff, but you don’t have hard causality and it’s very, very seldom you find causality. Yeah. For example, if you think about leading a lagging indicators, churn is a lagging indicator. It tells you people left after you close the door, the horse is already gone, right? Whereas let’s say it’s users, they go to the support page three times in their first week or users that don’t use the reporting feature of your app, or users that only visit for 5 minutes instead of 20 minutes. Mm hmm. You can predict who’s going to churn. That’s a leading indicator, right? Because I can predict how time. I can use that to try and help those users, give them a better experience, reach out to them, offer them a discount, whatever, and keep them around. Yeah. So. So correlation is very useful, even if you don’t know the cause. Yeah.
Bronson: And something you just said kind of made me think of something interesting. You know, we’re getting more used to event based analytics with things like Mixpanel and things like that. But it’s interesting, we always kind of look for the positive correlations. So when they do this and they do this, then this happens. But there’s also when they don’t do X, Y and Z, this happens. So maybe even not just looking at the events that happen, but look at the events that don’t happen like they don’t contact customer support or they don’t, you know, do this kind of thing. So it’s really looking at the positive and the negative event correlations and eventually causality.
Alistair: Absolutely. You know, those LinkedIn those little LinkedIn things that say, you know, what, do you endorse this person for these skills? Mm hmm. Do you ever have a friend that that isn’t really that good at them. And you just politely call click the X and cancel it. You think LinkedIn isn’t counting those?
Bronson: Yeah, that’s true.
Alistair: That’s good. But I guarantee you, they’re counting every one of those. Right. And they’re doing it because this is a great way for them to find out. Oh, you don’t think that highly of that person? That’s actually a negative vote. I mean, it’s not a down vote, but it’s a down vote for them, right? Yeah.
Bronson: And you know, it’s funny, like, I.
Alistair: Want to see a list of people who don’t think I’m good at what I said I was, because that’s the people I’m not going to vote the.
Bronson: Party. Exactly. You know, it’s funny you say that like I use Pandora, right? And I keep wanting them to do better. You know, I keep wanting it to work like I needed to, but I want them to know more about what my clicks mean. I get a thumbs up or thumbs down, which means they play it more or they stop playing it altogether. But I want them to understand that when I skip a track, it means I’m losing interest slowly in it, even though I’m not thumbs down yet. But I feel like there’s not that, you know, it’s not nuanced enough. They’re not taking into consideration all my clicks, just the really explicit ones. And I think that’s maybe a trend where I want to see more companies take into account all the, you know, all the the small nuance things to give me a great experience.
Alistair: Well, there’s actually a really interesting rathole. We can go down here, so I’ll just I’ll just.
Bronson: Yeah, we like rattles.
Alistair: Go forest dance on the edge. Yeah. Often the app designers design and their business model or add ons in Pandora’s case, and I’m not saying this is how Pandora does it, but hypothetically Pandora wants to stream new music to streaming music. They have to send you something and you buffer it when you hit forward. They have to refill your buffer from scratch. Their bandwidth goes up. Mm hmm. So people who forward a lot leads to a bad behavior if they encourage that as a way of sending a signal. And you knew that maybe you’d forward more concrete. Example, I use Expensify for my expense receipts. It’s really good. But their tool to scan a thing and crop it and rotate it and and read it is absolute crap. It’s a very like if you look at this picture of the of your pic of your receipt on their web interface. Mm hmm. You try and adjust the crop. You hit crop. You’ve got to wait, like 3 seconds. Then you got to rotating or another 3 seconds. If you try and change any of the fields while it’s rotating, you lose the data. It’s really bad. So they make money. By selling you a service that will scanning sherry receipts for $0.10.
Bronson: Makes sense.
Alistair: I would. So. So I’m not saying they do this, but when you they could very easily have their mobile device allow me to do crop and rotate to the picture once I’ve taken it. But that actually doesn’t help their business model. And so in many cases, especially with freemium or free products, you got to realize that a lot of the behaviors are there’s a decision going on, not just about usability, but about how it affects their business model. And I think people aren’t aware enough of that stuff. They just think, why isn’t this well-designed? Rather than what is the company trying? What behavior is the company trying to produce on my behalf?
Bronson: Yeah. Yeah. It’s not just pure altruism. Why they released this thing into the. Into the world. Absolutely. There’s there’s some commerce going on here. Speaking of commerce, you talk about understanding what mode of commerce that somebody is in. Talk us through this a little bit. You talk about the different modes, acquisition, hybrid loyalty. What are the differences those? And is it possible to move from one to the other or just kind of stuck in the one you’re in?
Alistair: Sure. Good segway, by the way. Very smooth.
Bronson: Thank you.
Alistair: So so this is an example of a useful leading indicator. It is. There’s nothing bad about being an acquisition focused e-commerce company. A wedding ring vendor. Is acquisition focused unless you really, really want like wedding ring every month for most of the world, at least at the time. You want only one running. Right. So a wedding ring cost us wedding ring companies unlikely to a second sale in a year. Mm hmm. On the other hand, a grocery store is in the loyalty mode. They really want people to come back. Okay, so you’re in loyalty. If you’re a grocery store and someone says, my fruit was bad, you probably ought to refund the purchase and send them a basket. Mm hmm. If you’re in the wedding ring business and they say, I don’t want your back, you say, I’m sorry. All purchases are final. Right. By contrast, when that person buys the wedding ring, you want to sell them insurance. You want to make them tell all their friends. You want to gouge them for as much as you possibly can. Mm hmm. In the loyalty case, what you want is to give the person extreme satisfaction, build features like favorite produce lists, like pay with a single click features, because you’re really trying for loyalty. Okay. There’s nothing wrong with either of those models. Mm hmm. Looking dirty about being in the acquisition business all as it is, there’s everything wrong with not knowing which one you’re in. Yeah. Because if you’re a wedding ring company and you start to put together, like, which of these many wedding rings is my collection of wedding? These I like to get, that probably sends a really bad message to your spouse. Right? If you’re in the, you know, the buy with one click feature, you know, you don’t want that. You want to offer them lots of other stuff, maybe upsell them to wedding planners and so on. So Kevin Hellström, one of the many people who helped us to this book, talks about measuring a simple metric. How many of your customers buy a second time in 90 days? Mm hmm. Pretty simple, right? Most people don’t look at this. If less than 15% of your customers buy a second time in 90 days, you’re in acquisition mode. Okay. And you should focus everything on low customer acquisition cost, very high checkout rate, because you will lose that purchase. High margins on the sale and so on. Mm hmm. You mean once the purchase has happened, maybe it’s okay for you to take longer, you know, procuring the product. Maybe it’s being shipped from China. Once you get the order, maybe that’s okay. You don’t carry a lot of inventory, right? That’s okay. You have the purchase and so on. You’re like most retailers in this case, if you buy, if you go play paintball and you think it’s the greatest thing in the world, the odds are you will not continue to play people, you’ll buy your equipment. It will sit in your garage until you sell it. Yeah.
Bronson: Guilty as charged. I have a couple of people that don’t get. Yes.
Alistair: It’s the classic example for people that have been married for a long time. So if you’re looking at loyalty, for example, and mattresses is another one for acquisition, right. Loyalty is like Domino’s Pizza. What’s the conversion rate on Domino’s Pizza is incredibly high because that’s where I go to buy pizza. It’s not that their site is better. I decided I wanted a pizza. I know what I’m getting. I’ve had it before, but I have it again. It’ll be exactly the same. Maybe with a weird feeling this time, but pretty much the same. Yeah. So if you’re over 30%, that means that most of your customers by the year, I mean, you’re just extrapolating out from 90 days. You don’t wait a year to get this number right. Yeah. And so if your customers are buying from you several times a year, you’re in the loyalty business. This is what Amazon is it? And you need to do things like expanding your inventory so you can sell them more things, frequent wishlist stuff, loyalty programs. And so a hybrid is like a Zappos where you’re buying shoes, you’re not doing it that much, but you need to increase the number of returns from a user over time and try to move yourself toward loyalty while still doing acquisition. So Zappos is kind of in the middle. It’s amazing to me this is a really simple metric. It’s dumb and easy to calculate and very few people do the math and it changes your entire marketing strategy for e-commerce. So this is a great example of a leading indicator for analytics that shapes your entire business strategy.
Bronson: Yeah. No, it’s so clear when you when you communicate like you just did. So it’s not cool. At first glance to this slide, I was thinking the goal that you’re going to talk about is moving us from acquisition to loyalty, when in fact it’s not about moving us. It’s about understanding where you’re always going to be because of the kind of business you are and who your customers are, and really embracing that and now doing things that help you instead of hinder you.
Alistair: Right. And you can see how like the wedding Ring of the Month club would. Fail. Right. So there are some examples where you can’t move it. There could be others. Like wine. Mm hmm. Maybe you’re an auction house that sells very expensive bottles of wine. People don’t buy that often. And you want to move towards people who are enjoying more wine in the month club. What Kevin told us is that in his experience and he’s done a lot of work with ecommerce providers, it’s very hard to move the needle more than 10% as far as how many customers buy. But so so if you’re in the hybrid, you wind up having to spread yourself a little because you’re you’re trying to be very good on support, but you’re still trying to be clever about customer acquisition, which is where someone like Zappos did a great job. They built loyalty. That is some clever marketing, and then they grow. And that’s why Amazon acquired them, is because they saw that engagement happening.
Bronson: Yeah, that’s great. Now, your next slide here. This is the one filled with all the buzzwords, right? Right. We’re talking about segmenting cohorts. And these are things I think a lot of people throw these things around because they feel like they’re supposed to but really grasp and understand them as something different. So talk to us first about cohorts and segments and what the difference in those two are, because I think it’s an important difference.
Alistair: Yeah, absolutely. So in this diagram, you see these five funny bands and the one on the bottom, imagine the one on the what it was all your users from January and they’re kind of dropping off as you go left to right. And then the next blue one is all of the February and then March, April and May. So all the users that joined you in March, in April, the teal colored thing that says cohort that represents those users and their lifecycle through your company. And what the reason the cohorts are so important for a startup is that the company you are in April may be entirely different from the company you are in January you’re releasing something new. So the experience of a user in April is very different if you’re buying dish soap. Your experience is probably similar from the experience of someone who bought Dish. So three months ago. If you’re using Flickr, let’s say your experience with Flickr as an MMO may be very, very different from the thing they released next time on a picture sharing site. Right. And so it’s unfair to analyze the users who had the experience with your first company in January as those of your first company, i.e. April.
Bronson: Yeah. And iteration almost makes it where you have to do cohort because you’re changing so much.
Alistair: Yeah. And that’s the whole point is you can’t really tell what happened. In fact, I didn’t include it here, but one of the examples we provide is we show two companies, one company, where the average revenue per user is like $5 the first month, then 450, then 433, then for 50, then for 50. And another company where in the first month users spend five bucks, the first cohort spends five bucks in the first month, and the next cohort spends like seven bucks and then eight bucks and nine bucks. So clearly one company is floundering. No one is doing really well. And then we make the point that they’re actually the same company. It’s just that all of those users who came in January and had a crappy experience are diluting the revenue. If you take the average revenue, I gotcha. Of the users that have joined you and me and are having a great time. Yeah. So I’ve seen people who go pitch by season. They show them average revenue in the VC throws amount, but we’re doing great. You’re looking at two different companies. You’re just analyzing them differently now.
Bronson: That’s great. Well, how else can you use cohorts besides like months? Because the way you just described it seems like, okay, that’s a great way to use it. The obvious way to use it. Are there other ways? Is there any other similarities among the groups of people that make a useful cohort analysis?
Alistair: Sure. So, I mean, anything these are called longitudinal studies because they follow the life of the users across along their lifecycle. And so a longitudinal study might be everybody that came from Twitter versus Flickr as part of campaigns. Right. Or everybody who is shown the first versus the second version of the product. But a segment of your users is when we say take all of the users and slice across them. By some attribute, they came on a rainy day versus a sunny day. Men versus women could be a demographic thing, whatever. But you’re slicing across those users, right? If you intentionally create two segments to test out. So let’s say some people are shown a red button and some people are showing a blue button that’s called a Navy test. And people like Google can do that because they have a lot of traffic. It’s more common for an early stage startup with less traffic to do what’s called a multivariate test, where I’m showing several things. I’m saying you’re going to be sunny and red, you’re going to be cloudy in green. You mean sunny in green and so on. And then you use regression analysis to see which combination of factors is most correlated with the behavior you want. Yeah. So is it Sunny, people who saw the red button that are most likely to be engaged? Great. That tells me I should do my marketing on Sunny Days with Red Buttons, for example. Yeah, and you’re right. These four words, like, people throw them around. They don’t really understand. A longitudinal study, which is a cohort study, follows along the lifecycle of users and a latitudinal study which is segmenting is basically taking all users and dividing by some metric.
Bronson: Yeah. Now, let me ask you a follow up question that kind of goes along with all these words. Do you ever find that the story that a cohort analysis might tell is different than the story that the segment analysis tells? Do you ever find that the data is at odds, or does it always kind of seem to make a cohesive, you know, story when you look at it? What’s been your experience with actually crunching the data and dissecting it?
Alistair: Well, it’s often at odds. And so I’ll give you an example from the book about a. A website called Circle of Friends. So Circle of Friends was a kind of did Google circles on Facebook before there was such a thing. And the founders had about had millions of users, about 10 million users, but very little engagement. People would come in, they’d set up a circle, and they just wouldn’t return. Mm hmm. And in sort of frustration, the founders went and mined through the data to look, to find out whether there was a a cluster of of a pattern of attributes that made the engaged users different from the disengaged users. Mm hmm. Turns out moms was the difference that moms were way more likely to invite. Other is way more likely to click on stuff, way more likely to write long posts, way more likely to start posts to form groups and so on. Like, like 100, 200% more. Incredible amount more. They actually changed the name of the company from Circle of Moms to circle of friends. Sorry, circle of friends. The circle of moms after this. And they shrink way down in userbase, right? Because they were alienating most of their users and then grew back up to 4 million heavily engaged users and then eventually had an exit to sugar. Mm hmm. This is a good example where if you’d analyzed all users along some segment, like, you know, people who came through Facebook, you’d have had a very different outcome from that analysis versus moms analyzed on Facebook. And so you can actually do some kind of a segmentation, especially if you’re looking for behavior you like like engagement along a particular segment like moms or star moms. Once they made the change and now it was circle of moms, you probably see very different results. So it’s usually switching between the find a segment and then see if it’s true for all cohorts or just the most recent cohort. Mm hmm. It’s usually the intersection of those two. It’s really interesting. You may find that with the new application that the feature you’ve added makes a huge difference. But for older users, it doesn’t make any difference because they’re already disengaged or they don’t use that feature. And so it’s it’s a lot of juggling between those two. Yeah.
Bronson: And it seems like there’s so much data that you’re wading through that you could tell a false story if you wanted to, to yourself, to VCs. I mean, you can come in with a cohort this segment, there’s a B that and there’s enough data that if you just don’t understand correlation causality back to our original discussion you really could tell a lie with the data. Is do you feel like that’s accurate or does it is it hard to make the numbers say something they’re not saying?
Alistair: It’s hard to lie to a smart person. Hey, I tell you, I have the fastest gray Volkswagen in Verdun, Quebec, parked on the left side of the kid seat in the back. Mm hmm. Maybe true. Right. Yeah. Not that useful. So if you go back and say we are number one with all single moms under a 19. Mm hmm. Who are worth more than $2 million on paper. Mm hmm.
Bronson: Yeah, that’s a small group.
Alistair: Right. And so. So. So the core doesn’t lie if you tie it to addressable market and the current market and how much that. So if you if I tell you that this cohort and I can reach a million of them and half of them will do what I tell them to do. Now you’re interested, right? Yeah. And it’s really identifying caught in this case. Moms. Can I make money from marketing to them? Absolutely. Can I reach them? Yes, I can. They’re identified very easily on Facebook. And so if you know you have a cohort, you don’t have a segment, you only have a thing you can do to produce a business outcome you want. Mm hmm. And that group is significant enough that it works.
Bronson: Yeah. You just mentioned producing a business outcome that you want, and that kind of comes back to the one metric that matters. Right. That thing we talked about before, you have this really great chart of moving through the the five stages of lean analytics and you really define the one metric that matters. Talk about this for us. Talk about the X and the Y axis here, the stage you’re at and the business you’re in and how this all works together here.
Alistair: Sure. So this is the kind of core of the book, and this is also why the book is so damn big. Because what the book should be is an app that says to you what business you in, what stage you at and then gives you 58 pages. We may do that. We’re not sure yet, but Erik Reis has these three engines of growth, stickiness, virality and price. We believe you need to go through all of those. We believe that every business first needs to get inside the head of its users to understand their needs and then to understand whether you’re the solution you’re proposing works. And we call this the empathy stage is very qualitative. Then you need to make sure people will use your product, you know, to the dogs, eat the dog food. And so if I’m doing that, we’re talking about stickiness. When they arrive, do they keep using it? If I can’t get a hundred people to like my product, I probably can’t get a million people like it. Mm hmm. Once you have those people and you’ve got a relatively low churn rate, less than 5%, definitely less than 5%, possibly less than 2% a month, which is the number, the magic number for churn. And this is a lot of what we did in the book is find out what’s the right number before you can move on. Yeah. Yeah. Then you focus on virality, getting people to tell others or designing features in the app that when you invite someone to use the app, they become a user. Mm hmm. It’s like a project management tool, like a sauna where you invite someone to a task. They become a user once you’ve got virality. Which means that for every user you acquire, you actually acquire, like, 1.6 users. Then you go to revenue, which is I make money and I pour some of it back into acquisition. And the reason you do revenue afterwards is because then you’ve got virality as a force multiplier. So your acquisition cost goes down because you’re bringing out more than one user at a time. Only once you see that working. Do you have a sustainable business model? Then it’s time to scale. Now, if you’re a restaurant owner, this may be time to franchise. If you’re an enterprise software company, it may be time to acquire a direct sales force. If you’re B2B started B2C company, maybe it’s time to build APIs. Scaling means different things. So those are the stages of lean analytics and they change depending on the business you’re in. So if you’re trying to make money from transactions like an E commerce or a two sided marketplace, you make money differently. You make it from someone. Buy something in a software. As a service or a mobile app business, you make money from subscriptions and in-app purchases like a cloud computing provider, for example. And then in user generated content or media, you make money from click through is ad revenue, banner advertising. So you care about the intersection of these two things, which is actually what the next slide talks about defines what your one metric that matters should be. Yeah.
Bronson: Now talk about kind of what the one metric is for these different companies. I’m sure you don’t want to talk us through every single possibility here because I take forever.
Alistair: I want you to go by the book.
Bronson: I mean. Exactly. That’s what it’s for. Talk to us about a couple of them. What are some of the surprising ones are the ones where companies don’t get that for whatever reason. With that on the X and then on the Y axis, they missed that. That is what they’re supposed to be doing. Give us just a couple examples of that.\
Alistair: Sure. So in the case of software as a service, what you really care about on the revenue stage is that you can acquire a customer for roughly a third or less of their lifetime value. So if a customer is going to bring in 500 bucks in their life, you want to spend a third of $500 acquiring them. At most, a good metric from that is the time it takes to pay off customer acquisition. So one of the companies we talked to said how many months until the customer has paid me back? Mm hmm. Right. Let’s say it’s eight months until they’ve paid me back for acquiring them. Mm hmm. Using that eight months. And then the projections for growth, you can tell how much money you need and you can go to a VC and say, Here’s how much money I need. Mm hmm. And there’s no more supposition or debating. It’s fact. And if you do that to a VC, they will go, Thank you so much. You understand your business. Like, that’s so refreshing for them, right? Here’s why I need the money. Because I want to get this many users, and it takes me this many months to pay back their cost. This is what they bring in. You have everything that we see needs to go. I know what your worth and I know how much money. Yeah. Nobody said.
Bronson: Yeah. Yeah. Why did they not do that? It seems like it’s. It seems obvious in a sense, but why do they not do that? Why is it going to the V.C. when they’re still there? Well, it’s that.
Alistair: It’s because they’re hard work. It’s actually I’ll tell you the reason why. The reason is, once you’ve done that, the VC will try to force a multiple discussion looking at comparables in the market. Mm hmm. And oftentimes, you know, people say hope and hope is not a strategy, but hope is a good valuation. So a lot of times people want to avoid stating what revenue is, right? Yeah. But even Twitter, when when Twitter was considering advertising, which was going to be its model, they would measure things like the number of times someone looked at their feed. Mm hmm. Now, that’s a vanity metric, really? Except they knew they were going to insert ads there. So all of a sudden, it became a useful metric. So one more from this graph. From this chart, I mean, spam. So what we found in talking to read it and we actually have a two part case study in the book on Reddit is that early on flagging was enough to defeat spam. But later on, Reddit spends 50% of its engineering efforts combating spam and, you know, robot upvotes and so on. So if you’re in the user generated content business, one of the metrics you care about early on is the ratio of good content to spam, because that predicts whether or not people will find the content on your site useful. And it’s that’s one of the places where people don’t necessarily necessarily spend a lot of effort until all of a sudden they have like a spam outbreak and they have to go fix it. So to these things, like in the stickiness stage, when you’re using user content, you care about the engagement being good, the community being good. Whereas for software as a service or for for a two sided marketplace you care about, are there enough listings in, let’s say, your real estate site to have enough property listings so that when someone searches, they get a result back? Yeah. So each of these things is different.
Bronson: And I think this is so important because, you know, entrepreneurs, we go online and we read the newest blog post by whoever and we just want to apply it to our business when they’re in a totally different stage and a totally different business with totally different things that are four causality correlates like it’s just so different. And yet we think it’s a best practice. So we try to adopt it, use it, and we can’t figure out why we can’t build something because they don’t understand our business. Right. I see you smile. Right.
Alistair: And the thing for each of these is there’s a line in the sand, right? Like I mentioned, 2% churn. Mm hmm. When we talked to one company that was killing themselves because they were losing a quarter of their customers a year, it turns out that’s really good. I mean, that’s 24%. That’s great. Go worry about something else. It’s time to move on. Right. But if you don’t know that, you’re just getting diminishing returns on something when you should be off fixing acquisition cost.
Bronson: Absolutely. That’s why I like that chart a lot. Now, you talk also about how it’s there are some different things to think about when it’s enterprise focused and you’re looking at the lean analytics. The stages are the same. You go from empathy all the way through scale, but you talk about it in terms of do this and kind of fear this or don’t do this. Walk us through the enterprise view on this real quick.
Alistair: Sure. So, I mean, most startups we see think of themselves as B to C, and there’s reasons for that. Right. It’s hard to get virality from an enterprise self. I just sold your big ERP system. You’re not going to go tweet out, Hey, I really love my Europe, right? You’re going to maybe do a case study for me or you’re going to use word of mouth and refer me to someone that you trust so I can sell to them. So the virality stage for enterprise focused sales is very different, right? Usually, or in the early stages to learn what people have. If you go and look at most of the B2B startups, it’s someone who worked in an industry and understood there was a place to take advantage of something or an opportunity to do better. And so early on you started doing consulting. Then you have this awkward moment where you’ve got to tell your consulting customers that they can’t have what they want. They can have what you’re selling them. Stickiness is less about will they use it? I mean, for example, let’s say you worked for a company and you had to use the software that you were told to use. You may hate it, but you’re still using it. So I can’t measure your use of the software as a form of engagement because you have to use it. And if I pop up a survey saying, Do you hate your job, then the guy that bought the thing is going to be like, Dude, stop serving my employees, so you can’t do that. But what you can look at is, how easily would I sell you the thing? Does it integrate into what you’re doing? How much training do people need on the product? Because that’s a measure of Can I integrate it easily? So do I get a lot of support calls to spend a month getting it to work with your existing systems or does it just go in? Yeah. And then later on, revenue is things like pipeline revenue recognition, your comp plan, those kinds of things that you got to worry about. Right. So there’s the same stages. They just they just manifest in very different ways. Yeah. Yeah.
Bronson: And then you also talk about how you view it for an entrepreneur as well and you add a stage for insurance or why do you add a stage? Well, what’s that all about?
Alistair: So you have to remember that an entrepreneur. Someone who’s working within an organization to effect change. Mm hmm. The difference between being a rogue agent and a special operative is if someone wants you to do it. Right? That’s correct. I mean, you look at any spy movie, right? That guy’s gone rogue. Means he doesn’t have executive buy in. He’s our special ops guy on the floor. That guy’s got buy in.
Bronson: So it’s the same.
Alistair: It looks the same same behavior. It’s just one of them sanctioned. Right. And so what you need to do beforehand is get buy in for what you’re doing from a high level executive who’s on record as doing it. Otherwise you got to do a political fight. The thing that people do understand of entrepreneurialism is that, first of all, it can be a lot harder, right? You may, when you launch your minimum viable product, have to comply with certain regulations that a startup wouldn’t because you have existing customers. You may want to make the products more viral and involve sharing, but be subject to privacy regulations that prevent you from connecting your users to one another. You may. Let’s take Microsoft, for example, trying to build a software as a service version of Office. Mm hmm. Well, that’s nice that they’re trying to do that. But how do you think people feel about how do you think the Microsoft direct sales force that sells software licenses feels about that? That’s a pretty awkward thing, right? Yeah. And so you’ve got to deal with channel conflict, sales resistance, existing established agreements. And the worst part of all this is that at the end, you don’t have an exit, you take your product and you give it to the rest of the company, which is the whole reason you were created as the rest of the company can’t do something right. So you’re going to hate what happens to your baby and you have to be okay with that. And if you don’t like this list, don’t work for a big company.
Bronson: Don’t be an entrepreneur. Right, exactly. Be an entrepreneur. Yeah. That’s great advice. Alistair, a credible interview. I have one last question for you. I’m kind of a high level question. What’s the best piece of advice that you can give to someone watching this right now who’s in a startup? They’re an entrepreneur and they’re trying to they’re trying to grow their startup. What’s the best piece of advice you have for them?
Alistair: So I’ll say what we say at the end of the book. You know, the story of Archimedes.
Bronson: Refresh my memory.
Alistair: So Archimedes had a king who said, Look, I think I got ripped off. I don’t think this crown is real gold. Go measure it for me. And Archimedes went, Oh, you want to measure the shape of volume of an irregularly shaped solid? That’s hard. And then he sat in the bath, and the bath water went up and he went, Eureka! I figured it out. I can measure the shape of a solid or the volume of a solid by putting it in water and watching displacement. Archimedes had taken baths before. He wasn’t some smelly, gross guy who thought of a bath. Right. But the fact that the king had asked him a good question. Mm hmm. Meant that as soon as he got in the bath, he noticed the water rise and found the answer. Mm hmm. In the old days, business leaders were the people who could convince you to act in the absence of information. They were the most convincing person in the room. Mm hmm. Today, the business leader is the person who knows what questions to ask. And so the hardest thing for any entrepreneur is to ask really good questions, because the answers are all around you. Analytics is easy once you know the question to ask and the hypothesis, which is the human thing is inspiration. And then the machines are just optimization. So the heart of the single lesson an entrepreneur is to just sit down for a while, talk about your business, find out what the riskiest thing is, and then ask the right question to identify the riskiest part and mitigate it or quantify it before you can move on. And that’s a skill that is, I think, in great demand and will be in more so as we move into a sort of data driven society.
Bronson: Yeah, what incredible advice. Alastair, thank you so much for coming on the program. It’s been a wealth of information and I got a feeling this interview going to be watched quite a few times. Thank you. Great.
Alistair: Glad to be here. Thanks again.
Get the strategies, motivation, and in-depth interview with all the details every week!