Aug. 9, 2021

Shark Fins

This week we're going to be jumping the shark. That is we're going to be looking at why tech companies and their success, their growth, their user numbers often resemble a shark fin: how they scale up really fast, then taper off then fall really,...

Apple Podcasts podcast player badge
Spotify podcast player badge
Google Podcasts podcast player badge
Overcast podcast player badge
Castro podcast player badge
PocketCasts podcast player badge
RSS Feed podcast player badge

This week we're going to be jumping the shark. That is we're going to be looking at why tech companies and their success, their growth, their user numbers often resemble a shark fin: how they scale up really fast, then taper off then fall really, really fast in a straight line all the way down.


Will Page: Welcome back to Bubble Trouble, where we explore the inconvenient truths of how financial markets really work. And in this week's episode, we're gonna be jumping the shark, that is we're gonna be looking at why tech companies and their success, their growth, their user numbers often resemble a shark fin, how they scale up really fast, then taper off, then fall really, really fast in a straight line all the way down. That shark fin is the basis for this week's Bubble Trouble, back in a moment. Welcome back, Richard.

Richard Kramer: Hey, Will.

Will Page: So this week we're gonna go to the tech startup world, an area that you know a lot about, an area that you've expressed a lot of skepticism and cynicism about. And we're gonna go to visualizing a shape of a shark fin to capture where bubbles appear and trouble reappears. And this shark fin idea, we all know what a shark fin looks like. I wanna cue the theme music for yours here, but we have this shark fin, which goes up in a sort of convex growth function up and then straight down.

Richard Kramer: It's also called an S curve, Will. That's an easier.

Will Page: We have tech companies that go up, create a lot of hype, then seem to taper off at the peak, find their peak and then fall straight down and crash and burn. So, Richard, what I wanna get to here is we're now basically, we're a decade anniversary of this Andrew Chan article, Facebook viral marketing: When and why do apps "jump the shark?" and it's 10 years. So what have you learned in those 10 years?

Richard Kramer: First of all, Will you realize that all companies start out at the bottom end of that S curve and the back of the shark before they go up the fin? So everybody's gotta start somewhere and it always comes with an idea, with some proposition for the consumer that's looking to scale quickly. Now how it scales has been turned into both a fine art, and that's the art of marketing and product proposition, if you will and a science. And a science comes in to the algorithms that are used to find people, to download an app, to engage them with notifications or other growth hacks to try to rope them into using it, and I think one of the key things we've learned is just how much of a heavy fashion element there are to these apps.

They always wanna find the next cool app and whether that's in the e-commerce space where maybe Amazon is well established and eBay was established even before that, now you've got dozens of other players. And indeed we've just seen one of the tremendous UK company, Depop get bought by Etsy, something that's really captured fancy of a whole generation of teenagers and, and millennials that really wasn't on the radar five or 10 years ago.

Will Page: So the reason why a lot of tech companies shark for these, not necessarily because they failed, is because another companies succeeded and fans are fickle. Is that a fair comment?

Richard Kramer: Absolutely. And again, e- when we have established behavior, this sort of muscle memory of digital engagement with a particular app, we can also get bored of it. And I think you could probably find plenty of examples in your own behavior of stuff that hey, and that was really compelling for a period of time, but you know what? I'm looking for something new.

Will Page: So let's go back to this 10 year old anniversary of the Andrew Chan essay, Facebook viral marketing, interesting thing about Facebook stock price 10 years ago, where that's got to today. That's not-

Richard Kramer: Hmm.

Will Page: ... shark fin by any means, it's an imagination, but what made that essay so interesting [laughs]. And again, there is a plethora of essays about tech back there, and there's an even bigger plethora of essays about tech today. But this essay in particular included some maths, in fact, Andrew Chan just includes some mathematics. He allowed us to work with a spreadsheet to understand what viral meant, that viral is a topic I cover in my bootcamp, draw lessons from Tupperware parties in the late 1940s, how Tupperware parties had a viral coefficient to spread across America, 15 people in a room sharing the use of Tupperware, sharing those use cases with their other 15 friends. We've had this term viral for decades literally, what do you make about the term viral? 'Cause I love how you look at these terms and you don't necessarily drink the Kool-Aid and think that's something revolutionary or that's something new. So when somebody says my startup's viral, what's your gut reaction?

Richard Kramer: Well, look, everybody wants to have a viral success because it implies that your success is being taken care of for you, that someone is spreading the word for you without resorting to expensive app install ads on a whole bunch of platforms to try to find users. When you get that word of mouth it's by far the most powerful marketing. And if you go back, for example, when I early in my career, as an analyst was watching the growth of mobile subscriptions, it was following a phenomenal Fibonacci sequence where-

Will Page: Wow.

Richard Kramer: ... one plus one equals two and one plus two equals three and two plus three equals five and so on and so forth. And if you added zeros and millions onto those numbers, that was exactly the spread of mobile technology in the first 10 years that followed that Fibonacci sequence. But then naturally you reach a saturation point where you can't go past the four or five billion people who can afford some form of mobile access subscription on the planet. And there are limits to that virality, but the whole attraction, fatal attraction in some cases of that viral equation is that instead of having to spend a bucket load on sales and marketing, word of mouth does that hard work for you.

Will Page: I'm hearing that. I wanna dig a little deeper right there because with Spotify, I remember we have a free tier and a paid tier. And back in those early days, there wasn't much marketing budget. And everyone said you lose money on the free tier. You make money in a paid tier one. One of the responses I had in my head was, well, we don't spend a lot of money on marketing, but we do lose a lot of money on the free tier. There's a counterfactual there, but can you extend that logic of viral replaces need for marketing? Do you think it's situations where people want to spend money marketing their viral message, does it feel full circle?

Richard Kramer: Well, certainly the marketing will be spent to craft the message that you hope and pray will go viral. And I'll give you a great example from another industry that we all know really well. And that's the food industry, the restaurant industry. Now, when a local restaurant started up 20 years ago, they used local word of mouth in the neighborhood and hope that people would pass by, and actually there's not a lot of repeat business that restaurants get. Now when restaurants launch in a big city, they have a PR firm and they try to get the word out to food blogs and, and invite influencers. And it's not just the half dozen restaurant reviewers in the newspaper that they need to influence, they have a lot more channels to spread their message.

And there's kind of a paradox here because on the one hand, the restaurant is probably definitely afraid of being completely swamped with interest people to come because they'll have to turn half of them away and they'll never come back, but they also desperately need to be swamped because otherwise it's an extremely tough business to make money. And if you don't find a lot of people clamoring to get to your restaurant, you'll never get the additional sittings and be able to charge premium prices that you need to make your rent and, and that's a very tough business model to make work.

Will Page: Let's get back to this essay, the, the Facebook viral marketing essay, 10 years on, since Andrew Chan introduced this viral coefficient to us. The way I wanna illustrate it for our listeners using non-technical language, breaking it down is to think of, I gotta start up. I need to throw something at a wall for it to grow. So what I throw is it marketing budget? Is it a free tier? Is it an invite code? What am I throwing to get traction with my startup, to start it to scale, the shape of that shark fin again, how often do I need to throw it and how much it needs to stick? Because what Andrew Chan's core premises is when you're looking at startups scaling, it comes down to retention and a retention coefficient drops below 50%, that's when you shark fin. That's when the only way is down. You're on the slope of hope and cashing your coupons, come in number 13, your time is up. Talk to me about that whole retention element, do you see that as the key ingredient of this or is there other stuff we need to consider too?

Richard Kramer: Well, obviously there's at the first level a quality filter. There are 2 million apps on the Play Store, an Apple's app store. They are clamoring for attention and something that we can go back to and review as a recurring theme in Bubble Trouble, which is this hyper-competition, this overproduction-

Will Page: Interesting.

Richard Kramer: ... in every area. So the key, first and foremost is to have something that feels or looks unique to the consumer. Now, maybe that's in a category that's very familiar like mobile casual games of which there are literally millions on these app stores. And many of them are very similar in the way they work. But, you know, as long as you find something that's new and you can entice people to come in to, to, to give it a go, then that can be very successful. Now, the next thing is obviously some level of engagement and for a gaming company, that engagement is getting you to pass by, playing all the levels and to keep progressing through the game. When it's a social networking service, it's giving you some sort of feedback and this is where the term growth hacking comes in, where you're gonna send your users notifications. So I'll give you an example of an app I used recently to sell some used books. It, it was a way to clear out books that I just couldn't take the time to box up and take down to my local Oxfam.

Will Page: Did you include mine in that box?

Richard Kramer: No, I didn't. I, I kept yours on-

Will Page: [crosstalk 00:10:45].

Richard Kramer: ... the shelf Will. And the reality is they give you kind of a nice little bit of money, a book you might've bought for eight or 10 pounds, they'll give you a pound or two and you feel kinda good that you're getting 25 pounds back from a box of books that really you just wanted to get out of your house. And the second time that you use it, you'll realize that, hey, hang on a second. Instead of giving me one pound 50 or two pounds for those books, they're offering me 80P or 40P. And then all of a sudden it doesn't feel so attractive anymore, and then the engagement is like, could it tail off?

Will Page: Sliding scales.

Richard Kramer: As long as you're hitting that sort of dopamine reward system in the brain saying, you've got something from opening up this piece of digital technology and engaging with it, then you'll keep going. And that's why likes and reactions on social media are just so important for cementing that usage and, and maintaining the virality. Because once you get those likes, you wanna get them from as many people as possible.

Will Page: Now I have to admit I don't actually use Facebook. So when you're talking about likes, you lose me a little. Can you bring it back to restaurants for me 'cause I'm really at home with that analogy?

Richard Kramer: Sure. And let's go back to that food market. It's extremely hard for restaurant to get repeat business. Let's say a couple eats out once a week, 50 visits a year. They want variety. They want experience outside of their neighborhood restaurants, what's the chance they go back to a place more than once a year? And when you think about London, just around where you live in Kentish town, not to be too precise, there are hundreds of restaurant options that you have to go to and you're not far off from getting into town where there are thousands more.

So you have that same sort of cultural overproduction in terms of restaurant choices. How does a restaurant get you to come back? Well, how many of them actually communicate with you? How many of them have an app? How many of them have a way to notify you, hey, you know what? We've got a special on haggis today and that'll get you running.

Will Page: Running away, not running from [crosstalk 00:12:45].

Richard Kramer: Running away. Well, you know, there's a reason, I don't know any Scottish restaurants in London, isn't there Will?

Will Page: Do you fry food, doesn't go with running skills. So I guess-

Richard Kramer: Exactly. Fried Mars bar is for dessert and haggis for the mains. How does a restaurant really engage with its customers when they know they're only likely to see them once on average. They're relatively few, they're gonna come back. Well, if you give someone exceptional service, they might be more likely to come back. If you identify them as a local, someone who's walked over to the restaurant, as opposed to arrived some other way, you might be more likely to get them back. So it's that sorting algorithm of figuring out who of your users are likely to be repeat users that will engage with you more frequently and focusing your time and attention on them. Now the app world is obviously much broader in terms of the netted casts, but with a platform like Spotify, what I've noticed is they, they keep trying to do things that get me to engage with the platform even though they don't really appeal to me.

But I can see that kind of growth hacking going on saying, hey, we haven't seen you in a while. Maybe you've been listening to the BBC or Mixcloud or SoundCloud or some other apps, come on Mark and spend some time with us or let us figure out a way to do viral marketing [laughs], because we wanna see what playlist you'll share with your friends or that's shared listening function that they just launched. We wanna be able to see who you're sharing with and see if we can market to someone that, in their friend circle.

Will Page: Yeah. And Amazon Prime are doing shared viewing, you know, as well, a function of lockdown where you can watch with your friends content and presumably social features or bull arm as well. So this is kind of clever because you're coming at the word viral and you're coming out with the word frequency in it. It's not just restaurants, is it? I mean, if you think about travel apps like Skyscanner and Kayak, the one thing they always had to solve for was frequency. Skyscanner is a great app, allow you to scan all of the skies, including [inaudible 00:14:47]. How often do you need to travel is a similar question to how often do you need to dine out? And when you do wanna dine out, how often do you wanna go back to a place you've already dined out in? You could have the perfect product, but rather than think about retention, think about frequency of the customer.

And one other point I wanna unpack there, which you mentioned is you raised hyper-competition as a recurring theme. How do you explain hyper-competition in a matter of five or six words? And he said it beautifully, he said when quantity goes up, quality goes down. I think that's something we can come back to in part two. So far, we've discussed sharks and theory of your like. Part two, we'll go into sharks notion, back in a moment. Welcome back to Bubble Trouble where this week we're jumping the shark. We desperately need that jazz music in the background just to make sense this one, because we're discussing tech companies, which scale up then falls straight back down again, resembling that famous shark film from the movie. Part one, we looked at the theory, you know why the original essay from Andrew Chan, which introduced us to the concept of the viral coefficient talked about retention and Richard very cleverly made us pivot from retention to frequency.

Is this the problem you're trying to solve when you're trying to avoid your tech company resembling a shark fin, but we are still seeing a lot of companies fail to jump the shark. We're still seeing a lot of companies go up and straight back there. And in this part two of this week's episode, I wanna explore the reasons why, why is this still happening? So let's go from success to failure. Richard, if I look at viral coefficients or frequencies, I can learn a lot from the success stories. Here's to Facebook and Dropbox. If you think about how the viral coefficient works, so the frequency function as you were discussing works for Facebook. I joined Facebook, I've been, I spend the first two to three weeks inviting all of my friends and then it begins to taper off. So now I've got all my friends in from inviting 100 people a week to one or two a week.

Whereas with Dropbox, it's the backend at [inaudible 00:17:11]. So I joined Dropbox. I give them roughly $100 a year. I start using it and the more ingrained they become in Dropbox, the more friends I invite. The more collaborative projects I build in Dropbox, the more I need other people to use it as well. So the coefficients can work in different ways, but Facebook and Dropbox have both been successful companies. What lessons can we learn from the success stories we see? What companies have avoided the shark fin?

Richard Kramer: One of the things you mentioned that was a common theme between those two companies and many more. And I think LinkedIn is another great example on the professional side-

Will Page: Mm-hmm [affirmative].

Richard Kramer: ... is that you-

Will Page: 100%.

Richard Kramer: ... invest time building a network, and then it becomes a sunk cost. And then you say, g- unless there is an API or some other lift and shift algorithm that takes all of the investment I've made into building this network and dumps it somewhere else that I can make use of it, where I would see a better value proposition, then I'm stuck. I'm gonna stay there and I'm not moving because frankly, the other thing that we haven't really talked about in terms of the major limiting factor for all of these apps and the single currency we have to measure from an investment perspective is time spent. And the limiting factor is we've only got 24 hours a day minus the eight hours we're supposed to be sleeping and minus a few other hours of stuff to do.

And you really don't have that much time to spend on those apps co- in a collected sense, all manner of content offerings compete for that time spent. You've got the purportedly viral networks that you're engaged in, and they're always pinging you to basically say, hey, I'm here. Spend some time with me because I need to be the one that you're devoting your attention to in that attention economy.

Will Page: Interesting. So time spent on LinkedIn is rewarding time spent or getting something from it. Time spent on those apps, which crash and burn clearly isn't producing the productivity that you want and the value-

Richard Kramer: Oh, but...

Will Page: ... of that time.

Richard Kramer: Let's not privilege productivity over everything else. I mean, we all want entertainment and productivity. The time spent on LinkedIn, the time spent on casual mobile games, no one is suggesting that that's for productivity, but those are the most sold category of apps on the app stores.

Will Page: Interesting. And this is something I discussed at a conference, which took place yesterday with 4,000 people in an open marquee. And I used the, the off con framework, which I highlight in my big tars and economics for thinking about attention, which is not time spent, but relative attention paid to activities and relative importance of activity. So I just wanna take your example and kind of unpack it a little bit. So during lockdown last year, a colleague at Netflix showed me that time spent watching dramas or documentary was going up, but the relevance of that time was going down. It's just an interesting way of self-breaking down time spent into two functions, how much time and how important was that time? Yes, I've given six hours to watching tigers in captivity on Netflix, but no, it wasn't that important the use of six hours.

Does that give an opportunity for another app to come in and compete for a better time spent as well. Let's now go to what we can learn from the failures. So there's still countless others who have crashed and burned. I mentioned Dropbox and Facebook at the start. You added LinkedIn, we've got our success stories. What themes are emerging from those which jumped the shark and fall straight back down like a shark film of telos.

Richard Kramer: There is no doubt we hear endless iterations of the success stories and what incredible visionary founders have been able to accomplish. And one of the things that's usually left out of that story is just the sheer amount of continuous innovation in the platforms that's required. Now, if you recall when Facebook went through its early days and growing pains, and it's early on auspicious debut as a public company, but there was the concern that Facebook needed to shift from desktop to mobile. And famously the company came up with a list of 600 potential mobile applications they could launch and had to relentlessly winnow those down until they came up with a few that were as compelling on a smartphone screen as they were on the desktop.

And indeed they rearchitected what they did on the desktop to remove that right-hand rail and, and reorient the user experience such that it looked fairly similar on a desktop and on mobile. Then you had very interestingly them not only buying Instagram, but very quickly, they made Instagram, the setfacl visual user experience standard for the web. So if you go onto any of the major sports websites, sports league websites in the, in the US, you'll see they look a heck of a lot like Instagram when you look at them on a mobile device. And indeed when you look at ESPN or some of the other sports websites, they've all conformed to the visual identity of Instagram.

So part of the effort, once that initial network is set up and the sunk cost is made by the users is to give them a, a familiar terrain that they can ingest other incremental content of interest and where they can go, where they, if they go elsewhere they feel that warm and fuzzy glow that draws them back to the mothership of the key app. And I think if you look at many of the failures, whether it's the Myspace and Bebo's, if you wanna go back that far, or many of the other apps that-

Will Page: My empty space is what I like to call it.

Richard Kramer: ... it's empty space now, but it wasn't a pre-Facebook. It was the place to go and they just didn't satisfy that continuous innovation. And so some other shark was able to swim into the curve and start eating the tastiest fish.

Will Page: The fish analogy Will, right there. Coming to the end of this week's podcast, we always touch on smoke signals. I mean, the whole purpose of Bubble Trouble is to give the audience some smoke signals they can hold onto and think when they're next placing a bet on the stock market or looking to invest in a startup or looking to work for a startup, what are those signals they can look out for to think something's fishy here, and that was the shark analogy. Something there's some red meat in the ocean, and that shark is gonna swim up for it. So smoke sequences system, I wanted to toss out one, I'm gonna ask you to toss out another. For me, the big one here is manipulation of the active user base. You rightly impart one point to frequency as issue, how many people are using this app?

How often are those people using this app? 'Cause anyone wanna stop the clock and say how many of those people are unique? So for example, if I'm measuring my audience by IP address, which should, well, I think I've used five IP addresses this morning. I'm not five people. And for me, that's a classic smoke signal of, we have X million users in the first month of launch. Okay, what is a million users here? We'll draw it back to unique. How many people of pulsates are actually using this thing and how often are they using it? And I often think are really important smokes signals that people boast about their numbers and say unique. Give me people with pulses-

Richard Kramer: Hmm.

Will Page: ... not IP addresses, not MBUS, not apps whilst they were open, but not being used as a dormant app.

Richard Kramer: Indeed. One of the things we look at closely is the ratio of daily active users to monthly active users, because that shows some measure of engagement. However, the other key point is the measurement of those active users, whether daily, monthly, weekly, what have you is very willy. And like you said, it could be apps that are open. It can be banks or phones emulating the usage of the app or bot traffic on sites.

Will Page: Yep.

Richard Kramer: And this is all either growth hacking or just fraud. And there's lots of it. My analogy here is this sort of sausage factory one. You never asked what goes into that sausage. And indeed, if you're an ad buyer and you're buying 10 or 100 million impressions across a huge mobile audience, you don't check that all of those users were real users. You just hope that a sufficient portion of them actually are. And it's clear that with companies that are eager to show the market that they have a hockey stick curve of monthly active users, there are lots of ways to get people to, to show up as an MMU, which doesn't really mean they're devoting their time spent to you.

Will Page: I think we're really well aligned to what this smoke signals are. They love that word impressions. What on earth does an impression mean when you're deciding whether to invest in a startup or not, that needs to be thrown to the sharks as well.

Richard Kramer: Well, if I had to have a smoke signal to throw out here, I would say, remember virality. And the concept of virus is something that we build immunity towards.

Will Page: Mm-hmm [affirmative].

Richard Kramer: So virality is something that can be manufactured by clever PR, and every social media post that gets put up there by an influencer is desperate to figure out a way to make themselves go viral. And that doesn't mean that people can't build immunities against it and overcome that vitality. So the collective influence of influencers would diminish with each end plus one influencer that's competing for time. And the collective aim of every new app to go viral is gonna be diminished each time a new app comes out because you can't keep restocking your phone and filling it up with more and more and more apps.

Will Page: So are you suggesting when we bring the shelter close, we discuss smoke signals. We should swap words around and call it the vaccination program instead?

Richard Kramer: Well, we're trying to vaccinate listeners against a lot of-

Will Page: [Laughs].

Richard Kramer: ... the, uh, a lot of the noise and relentless marketing of, of these concepts when we know from hard evidence over decades that the vast majority of these companies are likely to fail. And what we've seen unusually in the market in the past, let's just say a decade or more certainly since the global financial crisis and the slightest tech boom is that many companies simply launch as a concept and then hope to do the hard work of relentless innovation that we've seen demonstrated by the companies that we know so well, having started out as viral apps, but having established themselves as hundreds of billions or trillion dollar companies.

Will Page: Fantastic. Thank you so much, Richard Kramer. Bringing us to a close, I think in the first part we discussed sharks and theory and in the second part, we worked through some sharks and action. And it's making me think about that old analogy and it's a chicken and egg analogy here of, do you come for network and stay for the tools, or do you come from the tools and state the network? Definitely a theme I wanna explore in future Bubble Trouble episodes as well. Also, Richard brought up the term, hyper-competition something we've stumbled on whilst making this podcast, something which has got real traction, potential bigness topic, but yeah, when the quantity goes up, quality seems to go down and think about why companies go up suddenly and down rapidly could be a result of hyper-competition too.

But lots of food for thought, lots of food for the sharks to think about too, and you've been with myself, Will page and Richard Kramer, and this has been Bubble Trouble. If you're new to Bubble Trouble, we'd encourage you to follow the podcast wherever you listen. Bubble Trouble is produced by Eric Nuzum and Jesse Baker and Magnificent Noise. You can learn more at See you next time.