Network Effects in Financial Services (#8)
This podcast was recorded at FinTECHTalents’19 Festival and in it we’re exploring the potential of unleashing network effects in financial services. Ben Robinson is joined into the conversation by: Evgenia Plotnikova (Principal VC @ Dawn Capital); Martin McCann (CEO at Trade Ledger) and Oliver Prill (CEO at Tide Business Banking).
Full podcast transcript:
Ben Robinson (BR): Welcome to the Aperture Podcast. So for this one we’re at FinTECHTalents and we’re talking network effects. And so we just did a panel on network effects, which we’re going to basically continue, but have the chance to go a bit deeper on some of the themes. So I’m here with Dr. Oliver Prill, CEO of Tide which is UK challenger bank focused on the SME space, which I thought had got to 100,000 customers, but I’m told that it is already at 120,000 customers, so exponential. We’ll come back to how it’s growing exponentially in a second. I’m here with Martin McCann who is CEO of Trade Ledger, a system intelligence for lending. You comfortable with that?
Martin McCann (MM): Yes.
BR: Yes. System of intelligence for lending, which boasts customers in the UK, Hong Kong, Australia, including several of the world’s largest trade banks, and also Evgenia Plotnikova, partner at Dawn Capital, a London-based VC firm with numerous investments in FinTech including iZettle, Sonovate and Soldo, a business pay and spend management platform.
So the premise of this conversation is that sustainable competitive advantage in any connected industry has moved from supply side economies of scale to demand side economies of scale as consumers have become connected, and as computing has become ubiquitous, there’s this ability now to use that information flow to create network effects. This won’t help the listeners, but we’re going to go left to right here. So each of your businesses, could you talk about the kinds of network effects that you’re seeking and Evgenia in your case, the network effects you’re seeking as a VC, plus the network effects that you look for in the companies you invest in. So Martin, we’ll start with you please.
What we’re essentially trying to do is create a network of quality data that provides the potential for innovation in terms of services which can be built on top of that data. But the problem we’re solving initially is that convenience and trust problem of both sides of the network. — MM
MM: So Trade Ledger is a LendTech business. We provide a white labeled lending platform to banks and to non-bank lenders as well, anybody who provides a credit service of any type to a business. And where the market is today is basically we’re providing lending automation capabilities, both on the origination and on the loan processing side of the equation to the suppliers of credit to the marketplace, trying to bridge the 1.2 trillion pound global gap in working and growth capital for businesses around the world by working with banks who can scale as opposed to trying to be another alternative lender ourselves. And one of the reasons is to do with scalability and I guess the inflection point in the industry, which leads to potential for network effects moving forward. The core problem that we have is data. So there’s not very good data on particularly smaller businesses, and the problem that causes for, I guess, the two sides of the network — the supplier of credit and the consumer of credit, the businesses — is there’s a lack of trust and convenience.
So the convenience side, it takes on average 90 days from first application for credit to first drawdown, 30 hours of work. That’s not very convenient in the modern digital age. On the lending side, the banks actually can’t get hold of enough data in order to correctly assess the credit worthiness easily of the SME. So if you can solve those two problems using the same data, then there’s lots of ways you can add value outside of just that sort of I guess inline value creation and delivery providing.
BR: So just to clarify, so what you’re saying is that the banks already have SME customers, so you’re not linking the banks with the SMEs, what you’re doing is using data or creating data network effects to enable them to service those SMEs better.
MM: So the problem is the quality of the data that actually, there is no network of quality data between financial service providers generally and SMEs in particular. Even with larger companies, it tends to be proprietary networks. So what we’re essentially trying to do is create a network of quality data that provides the potential for innovation or potentially innovation in ways that we don’t know in terms of services which can be built on top of that data. But the problem we’re solving initially is that convenience and trust problem of both sides of the network.
What we’d like to distinguish within our business is three things — network effects themselves, then the virality factor and economies of scale. […] in my view, you can actually create significant businesses without having network effects -EP
BR: Okay, Evgenia, same question to you. So are you seeking network effects within the VC business, and are they even possible? And then secondly, I think you are principally a B2B investor, right? So what kind of network effects do you look for in the businesses in which you invest?
Evgenia Plotnikova (EP): So let me start maybe with the second question first. So as we kind of discussed a little bit during the panel, what we’d like to distinguish within our business is three things — network effects themselves, then the virality factor and economies of scale. So with network effects, the positive ones. It’s about adding incremental value to the product every time the usage increases by adding an additional user. With virality, it’s more around the shared distribution and every incremental user increases that adoption of the product. And then finally with economies of scale, it’s more around the reduction in the marginal cost. So as you become of critical mass and grow into a bigger behemoth, you can drive drive the economics and decrease your cost basis effectively.
So I’d argue that within our business, all three can be extraordinarily valuable. With one of our largest exits, which is iZettle which was sold to Pay Pal for over $2 billion, I don’t think there were significant network effects. This is a business where if you are a merchant who has a point of sale dongle, it doesn’t necessarily mean that the next merchant will get a bigger benefit out of it. That might be distribution challenge or distribution acceleration through that virality factor or economies of scale scale, but I don’t quite think the network effects was there. So in my view, you can actually create significant businesses without having network effects.
BR: iZettle? Was that a payment scheme or was that card acquirer?
EP: Point of sale payment solution. So mobile sale, mobile payments. So as a representative, if you were kind of using a MasterCard product or Visa product, then yes, absolutely. As a user of that, the more it gets accepted everywhere, the better it is for you. But from a merchant having the dongle, it doesn’t necessarily matter whether the other hairdresser has it or not. And it became obviously a very, very big business. I think what we see in terms of specifically network effects is perhaps more was Martin was talking about. In a lot of our B2B businesses, it tends to be by data. So within software, a bit on the credit scoring side as an example, or collecting user data to make a more valuable product ultimately.
And to your first question, I do think we actually have a little bit of network effects within our own fund, so we’re focused on … we do B2B only, B2B software and B2B FinTech. So as we continue to scale and every portfolio company joins Dawn family and we partner with them, they can benefit from the collective knowledge that has been assembled through over half a century of board seats that we have accumulated between all of us. There’s certain commonality in customers that tend to multiply, certain commonality in channel partners and distributors, and so arguably, I think by being a focus, we seem to experience a little bit of those network effects.
BR: And do you think that’s the single biggest thing that differentiates one VC from another, which is how strong those network effects are that they can plug other portfolio companies in to?
EP: Yes, it can be. A lot of my esteemed colleagues and frenemis take a different approach being generalists and a lot of them have done fantastically well in the market. It’s just the route that we have chosen. I think that being focused and being specialists allows us to offer a better, more tailored help to the founders that we’re privileged to partner with.
BR: Oliver, what kind of network effects is Tide seeking, and the fact that you’ve gone from 100,000 customers to 120,000 customers in the space of weeks, which suggests that they’re kicking in. So what are those network events that are kicking in?
The platform generates a lot of data […] so we invest a lot in data science and machine learning to basically do all non-simplistic decision making. […] Every time you can’t make a simplistic decision, we deploy machine learning and machine learning gets better the more data points you get from different users because the models become whack. — OP
Oliver Prill (OP): Yeah, so maybe first of all, just to finish off on the VC, I would actually buy into the notion that with VCs focus increase their network effects. The three VCs that I’ve had since I’ve been CEO really pushed — Anthemis, Augmentum, SBI are all FinTech specific, and this may be different in other interests, but given that a FinTech has a lot of specialties around it, it is actually a lot easier if you able to leverage the network of it, and indeed you and I met at one of those Anthemis events, at a great offsite where we talked about network effects in FinTech. Anyway so just to say, I absolutely endorsed that. Now Tide, what Tide is: Tide is effectively a platform. We say we are a business bank but really, we are a horizontal layer that spans anything that a business would refer to as admin and finance. So it’s anything from payments to credit to a deposit, but also payroll, expense management, invoicing.
Basically what we consider the operating layer of an SME, not the application layer. So we would never go in and try to tell the restaurant how to run a restaurant or whatever. We want to be the go-to place for anything of their finance and admin needs. And fundamentally what we are trying to do with this, the problem we’re trying to solve, that a small business, 5.7 million out of the 5.9 don’t have a formalized finance function, which basically means they spend 48 days a year basically doing finance and admin, which is either the owner, even if they’ve got more staff, there’s usually one trusted lieutenant that basically does it and it presents a huge opportunity cost. So really when we started out, our mission and it remains, is to essentially save entrepreneurs, businesses, small businesses time so they can get on with what they love, which is really running their business and having a bit of spare time.
And what we do is we orchestrate across in the platform, we orchestrate across all these different solutions. So in general, we do not manufacture the core product. We work with lending providers and we only basically self-manufacture if it’s trivial like invoicing. If there is no provider, you can’t do it optimally. Right? For the rest of it we are there. Now network effects really we ventured into because we saw them naturally occurring and we see two sorts of network effects. One is I think technically called two-sided platform network effect. So the more members we’ve got, every incremental member effectively provides a further incentive for product partners to come on because we’ve got more product richness, more providers and conversely they’re more product partners we have on attracts more and more members. So that’s it, although I would fundamentally agree. Much of what we have seen so far is is virality and scale economies rather than networks. But we see that effect coming.
The second one is essentially data network effects. So let me just explain how we look at it. Having lots of data, how we started off the platform generates a lot of data and with that data we pass it on. That itself does not generate network effects. We believe it generates network effects if exactly the incremental user adds to that, and that’s really primarily in our work done through machine learning. So we invest a lot in data science and machine learning to basically do all non-simplistic decision making. So these are credit decisions where we do eligibility scoring. This is about what we call our internal app network where we basically expose the right products to the right customers. Every time you can’t make a simplistic decision, we deploy machine learning and machine learning gets better the more data points you get from different users because the models become whack.
BR: And do you use those data network effects to underpin the two-sided network so that you can better match customers with services?
OP: Our rule actually is a lot of what we do actually comes … we’re learning about the network effect as we go. We didn’t design the business around it. So our notion was we believe simplistic decision making is one of the reasons why businesses are under-served. We said everything that needs non-simple heuristic decision making, whether it’s on the supply or the demand side, doesn’t matter. We deploy machine learning and so we actually have, our machine learning is a utility that gets applied to different parts of the business. So it’s anything from better targeting of marketing, better eligibility scoring, better internal marketing. It’s all of these points and it doesn’t matter and it just happens to get better the more data points we’ve got, the more richness around from different users of data we’ve got rather than us having designed the business around.
BR: So one of the questions that I asked the panel that I want to go into a bit more detail now is this idea of internalizing/externalizing network effects. So are you familiar with that Bill Gates line where somebody said to him, what’s the difference between an aggregator and a platform? And he said a platform delivers more value to the network than it takes for itself. And I think that could be another way of framing that question I asked earlier, which is when you externalize the network effects, you’re sharing the network effects with the network in a way that enables them to have some control over pricing. For example, if it’s a marketplace concept, and contrast that with somebody like Uber, which very much sort of commoditizes the supply side of the marketplace by not allowing the supply to differentiate itself, charge different prices because you know when you call a driver, you just call a driver. You don’t have any choice whether it’s your favorite driver, a driver you’ve had before, somebody who has different characteristics, like might have really good chats or whatever. And the literature sort of says that the platform business model or those networks that externalize network effects tend to be much more durable, much more longstanding, sustainable. So in your businesses, do you think about that contrast between externalizing or internalizing network effects? Where do you stand on it? So, maybe Martin again.
We’re internalizing, but I think other people who sit on our platform can externalize. So it’s too big a problem for anybody to solve, so we’re trying to do the bit that we know how to do really well, which is to organize the data services and provide them so that other people can create other business models on top of that. — MM
MM: Yeah. So I think our market is so nascent that it’s hard to know what the steady state model is going to look like. And I think we’re probably five years away from anybody having a successful steady state model, which could be called a platform model in what we do. So we’re trying to start with the basics of how do we actually create sclae, how do we create standardization of data between what will become a two sided network? Because it’s basically the consumerization of business banking and trade. That’s kind of the way we see it, and right now there’s nothing there and I don’t think that we’ve even started to imagine what the types of value add services could look like if the data was there because nobody’s had the data before. So you can’t be too prescriptive on what that looks like.
So I kinda think what we were doing is a little bit like what Google has done with searching the internet. At that point in time when Google was adding value, who knew that it was going to decimate the entire advertising industry in sort of the late 90s, early 2000s, but that’s what’s ended up happening but that’s come because they’ve managed to get skilled. They effectively created a monopoly on internet search and organizing the Internet’s information. So if you take that concept and look at trade data, we’re in a similar, I guess, point in time where the combinations and permutations of trade processes and trade data is so large that it just can’t be standardized, but the technology is there today to think of interoperability, a normalization of data that runs across those different connection points and the ability to create connections much more simply than you did previously. That’s a platform that’s getting created today. And there’s lots of different approaches to that happening. So if you look at all of the banks, all of the trade banks have invested in at least half a dozen consortia. A lot of them are DLT and blockchain based. A lot of them are not, and they believe that that’s going to be a platform enough to standardize trade data. They’re not. There’s no reason for corporates to join those services. It doesn’t solve the convenience and trust problem that I mentioned earlier, so there’s going to be a different solution which is more organic and less prescriptive in our view.
We look at it as like we’re trying to create the platform. We’ve got a view as to how we create short term value, which is basically we charge the banks for the efficiencies created by having digital data and automation of lending, back office, and the efficiencies in terms of the cost of origination. There’s a very clear business model to do that. What we don’t know yet is what is the opportunity for creating value for the corporates on the other side of that model and how we can charge for and what types of services then can be created by the participants who sit on top of that network and how do we deliver the new data services, which actually create that value and how we charge for it. My view is we’re going to be looking more at internal versus external, but other players will have the opportunity to create that external business model based on the plumbing of the platform that we’re providing.
BR: So you’d argue you’re internalizing to externalize, if that’s not a contradiction in terms?
MM: Yeah, so we’re internalizing, but I think other people who sit on our platform can externalize. So it’s too big a problem for anybody to solve, so we’re trying to do the bit that we know how to do really well, which is to organize the data services and provide them so that other people can create other business models on top of that.
OP: Maybe just to build on it, let me give you a good perspective. We come at the thing a little bit, not only from network effects or whether we internalize, but also diversity. So SMEs are very diverse. So let me give you an example of accounting with just this one product vertical on it. We’ve taken a deliberate decision not to limit the number of accounting systems that we connect to the entire platform. So we’ve got Xero. We just launched QuickBooks. We have five or six that you can do on it and we call these Tide Connect. So in a way we externalize. At the same level, the problem with too much externalization we realized is 50% of all SMEs don’t have any accounting. They don’t want to even engage in this thing. So they actually have like, at best, what you could call a cashbook, like a scrap thing.
And so in addition to the Tide Connect, which is with all these, we actually internalize it as something called a Tide Tool. So we say you can actually own the platform, do simplistic accounting. Now the reason why we do this is they will just not go out and even want to engage with Xero. They’re just scared of running down. They want like a cashbook equivalent on the platform, highly integrated with all their other services. So we actually give members choice, and but in the end we would probably tend towards the externalization, not because so much the network effect is what we’re trying to do, but we actually believe for SMEs to actually maximize lifetime value is we need to give them choice, and an SME that today may come to the Tide platform and do effectively the Tide tools or simplistic way of accounting may over time want to go to Xero to do their self-accounting, or even to an accountant and we wanted to make sure we can get these stages of development of an SME covered.
BR: Yes. I think the definition in a two-sided marketplace is about whether those — what do you call them, Tide Connect — whether they’re branded or whether they’re white labelled and whether you give the SMEs choice and it would seem like on that definition then, you are very much externalizing them.
OP: As a matter of fact, we are handing them off. So the biggest challenge we’ve got in there is actually not so much our strategy and this is actually where the real network effect comes in. So when we started off, and I don’t want to mention any names here, let’s say some of these accounting players had issues, because they saw competitively we want to own the space and so on. Now they’re saying, well the 120,000 we can’t really get past them. We want to be part of that network. And that’s why I was saying there’s a network effect and people that a year ago sort of were saying, Oh well we will build all of these items, found themselves not being able to generate that volume and are now coming on. That’s what I’m saying. There is a network effect arising over and above just pure economies of scale.
Secondly, we are very open to it because we didn’t think about that it would have positive network effects. Glad to hear that. But we actually think if someone wants Xero for their accounting, it is in the member’s best long-term interest that you could have connectivity to that and in that case, clearly they’re choosing Tide. We would do the same with the … I don’t want to mention too many players that do expense management. We have our own simplistic tools, but we’re quite open and we are in discussions with other Tide Connects where we say, look, if you want that one that is probably more a mid-sized solution, you can connect it to Tide. We actually think providing this richness on the platform is a positive thing rather than very narrow thinking where you say, Oh well by having Xero connected where I don’t get a monthly transaction fee versus our tool where I charge them a fee, I minimize customer life. A matter of fact, we say the opposite. We actually argue if they stay with us through their entire journey, we actually maximize customer life.
But the key for us at this stage, we don’t think too much about, I guess the detail of what the network effect will be. Because if we don’t get the momentum and the traction and the scale, we don’t get the opportunity to create the network effect. — MM
BR: Just changing tack slightly, I want to talk about the chicken and egg problem. So again, we briefly discussed this on the panel, so the definition of a network effect is where the network gets more valuable with every additional user, but the network has no value if it has no users. So what are the proven techniques for overcoming that chicken and egg problem that you’ve seen that work and which ones are more difficult to execute in banking? So Evgenia, starting with you.
EP: Sure. So the two ones that we’ve talked about for me is one is either going small first or starting your company or your business within a defined geography or a predefined small sector. I guess the most famous example will be Facebook. So you’ll start at Harvard and then gradually you exponentially grow from there. The other way that we have seen that has been successful is creating a fantastic product. So a tool that users want to use, then they will then come back to you and kind of multiply themselves and then with every user, you will start having network effects. And to come back to the social networks example, I think Instagram has proven itself extremely valuable where when they first launched they were the first, compared to other consumer products to offer the filters for free that people could play with the product, fall in love with the product and then grow the user base virally to then start introducing network effects off posting on the same network over and over again. Where we’ve seen those things around kind of virality and liquidity and creation of values has been a lot in peer to peer, particularly with products like TransferWise or Venmo, where I think those approaches have worked quite quite well.
BR: Martin, what’s your strategy for solving the chicken and egg problem? Is it this kind of creating great service? So follow the rabbit strategy where you deliver I guess in your case, you deliver real cost benefits to banks. And then once you’ve done that, then you’ve got a critical mass of data to then kick-start the network effects. Or is it more around using the banks to, I suppose, kind of the same thing, but piggybacking on banks networks to get there fast?
MM: So I think in our market, there is no opportunity for network effects today because there is no network and there is no homogeneous data at scale. So the first thing is how do you create momentum and scale? It’s only when you’ve created a momentum and scale that you can actually start to look at what the opportunities are for network effects. So for us, we’re taking a very traditional model to try and create scale in the model, which is where a very traditional sort of SaaS enterprise software provider to the banks, but the way we do it is different, which gives us the opportunity to add additional capability and additional value through the banks to the corporates, or bypassing the banks to the corporates. Either a model is possible with the way we’ve delivered the platform.
But the key for us at this stage, we don’t think too much about, I guess the detail of what the network effect will be. Because if we don’t get the momentum and the traction and the scale, we don’t get the opportunity to create the network effect. We’ve kind of mapped out what that product strategy looks like and what we think the options are, but we think we are probably two years away from anybody having enough sort of momentum and scale to properly deploy any significant network effects in our particular market.
I mean we currently open more business current accounts than Lloyds or RBS group per month on a flow measure, and the reason for that is very simple that the network effects just don’t exist. There used to be a degree of virality, which really was just brand awareness. You would default to the Big 5, because you just didn’t know that was another. — OP
BR: So your cases are almost like — so I’m pointing against Trade Ledger and Tide. So your case is almost like polar opposites in the sense that you’re building for network effects that you think might be two years or more out. And in your case, you build a business and once you’ve reached critical mass, it just so happened that network effects kicked in. So would you argue that you weren’t seeking to solve the chicken and egg problem at the beginning? Or would you argue that you were, because in your case, it seems like a lot of the network effects coming through what people call consumer evangelism, which is the consumers on-boarding other consumers either, I don’t know if you offer referrals or whether it’s just genuine.
OP: I think I agree with you. I would call that virality. So I think that is definitely the way we’ve seen this is in the beginning really we were just trying to build a decent product. I mean partly to prove product market fit to attract customers. Honestly the time it would track problem we’re trying to solve at that stages show that we can prove, you can save time. And given that we worked in a market where banks hadn’t really understood network effects, we’re not competing with any network. No one had network effects and it sort of worked. The next stage of growth was effectively around virality, which really for us is a referral, and what we’ve actually learned is there’s active referrals so someone actually going out, but actually in small business that doesn’t happen that often, but it’s the passive referral.
So I’m looking for a bank — I heard of this Tide. Is Tide actually a good one? — So that’s what we call a passive referral, versus you going out actively saying to all your friends, go to Tide. So we see 50% are active or passive, passive being the bigger one and that is in a way a sort of virality. And then what we’ve seen is really for the network effects. It’s really not so much bringing people onto the platform, but once they’re on the platform they get more and more value from being on it. And we’re trying to nurture this and that’s probably how we would, if we had to do it again, we would do it. I think the strategy probably has to be different if you have a business that effectively is a network, like a social network, or if in the core part of it is who you actually try to compete in a business, in a marketplace where there has already strong network effects, but we were sort of facing and probably still today in business banking, this is not really as established as it might be, which is one of the reasons why we see at least on the flow measure, the big banks very, very rapidly losing. I mean we currently open more business current accounts than Lloyds or RBS group per month on a flow measure, and the reason for that is very simple that the network effects just don’t exist. There used to be a degree of virality, which really was just brand awareness. You would default to the Big 5, because you just didn’t know that was another.
BR: But it seems to me you could design these kind of same-sided network effects of social network effects I think are more difficult to achieve in banking than in other industries. I think you still could design them in. So I don’t know what they might be in your business, but this idea that there’s more value for another SME to join your platform if there’s some mechanisms which would make it easier for SMEs to deal with other, trade with each other, if you could net off accounts like working capital between two SMEs.
OP: The thing is, here you have to represent how strong is the network effect once you’re on the platform versus getting them onto the platform. So give you an example. We have various functionality of member to member payments.
BR: The same sided network effect, yeah.
OP: Exactly. That’s a good one. But that’s on the platform. Now would you join Tide? We live in a world where there’s a Faster Payment system. So basically what a member to member thing gives you is richer information. You can actually pull this sort of direct debit and it’s a great thing once you’re in the network. Getting someone to join, you have to switch their bank account. It’s harder. Would you do it just for that? No, it’s because you would basically do choose it because of various fundamental attractors, which at this stage I am not convinced whether the attractors are there. There maybe a little bit on the richness of the product offering, but it’s really once people are on the platform and I think every time you’re on a platform, often I see this with investors where we have this debate, we fundamentally differentiate between the propositions we need that are strong enough to pull people onto the platform. And then the propositions that generate the revenue model and the unit economics once you onto the platform and it, at least in the SME space, those are two very, very different challenges.
It’s the regulator. It could probably be more market share with network effects — because the winner takes most of the market — which is actually one of the debates we’re having with all the social networks, should the regulator not start to intervene because they have all of us, right? But we fundamentally believe in financial services. There may be niches where this doesn’t apply, but in general, 15 to 20% in any country exposes you to regulatory intervention risk. — OP
BR: Let’s talk about exporting network effects because one of the differences if we compare Amazon or Facebook or other platforms is that these are platforms that can operate across national boundaries very easily, whereas banking, you still have a lot of regulations that make banking quite domestic market specific, and so if we think about two-sided network effects, for example. They’re not so easy to move across national boundaries, because you’ve got a network that’s of, in your case, SMEs and partners that’s quite specific to the UK, which would be, not so easy to, if your next market was Germany, to just simply to recreate or to alter or to transfer across national boundaries. But whereas when I think about Trade Ledger for example, I could see that it’s probably an easier business to explore across natural boundaries because you’re providing a data layer that sits across many different types of banks. It doesn’t matter if your customers sit in Australia or the UK because you can still provide a data assistant intelligence that underpins businesses in different national or different jurisdictions. So, having talked your ears off about services, I would ask Evgenia to answer that first. When you think about portfolio companies and companies you invest in, and you think about as a business that’s where network effects are really important, do you think ahead to how do we make that business then scale across now national boundaries?
EP: The good thing about FinTech is that sometimes you don’t have to, in the sense that with a lot of financial services businesses, you could well be a national champion and still become a multi-billion dollar business. So for a lot of the businesses that are kind of very specific market driven, that can actually be okay. So if you think of certain banking businesses, some sort of insurance businesses, it could be a health insurance business, which arguably is a very national affair and it can become a very, very big business by being a French champion or an Italian champion as an example. With others, particularly with software, I’d say that those are a lot more exportable. When you start having data network effects, a natural transition for a lot of those businesses for our case will be to make them, or to encourage them to go to the US which is a very, very large market,and there it’s just sort of nailing your go-to market really and that can accelerate exponentially very, very quickly. So to divide businesses into different camps. The ones there that are more kind of software driven, which can be global and universal, and the one that I believe can be a national champion or regional.
BR: And so when you think about national, regional champions, does that bias you towards investing in companies that are focused on large domestic markets?
EP: Absolutely. So if you think of again coming back to the example of iZettle. iZettle never went to the US and you know Square has a fantastic presence there. Both are large businesses. iZettle is the Pan-European champion with Latin American presence and that’s enough. If you think about Soldo, one of the businesses I’ve backed most recently which is in spend management, they are targeting 20 million SMEs across Europe. That’s more than enough to become a multi-billion dollar mama. So will you ask will it be eventually on their roadmap? Perhaps. Who knows? Arguably it doesn’t have to be.
Platforms take a while to build. This is very different to a product vertical that you were talking, in Europe with 20+ countries or, you know Revolut that goes into 30–45 markets. Effectively, they are single product or similar product-centric propositions that very rapidly can go across borders because all they do is marginal change. — OP
BR: So, Oliver, how do you think about exporting your network effects? Because you know you’ve got international ambitions, but you’ve been very sort of reticent to tell us what they are.
OP: First, I do agree. Well, we haven’t decided what countries but one of the things I do agree with absolutely is, I think for many FinTechs when you are in a reasonably large market, there is not the immediate urgency. And so what worries us is the speed at which we’re growing, especially the flow metric is we believe from a regulatory perspective, there’s a natural, maximum market share, which probably in the UK, given that spin regulatory intervention in the small business it’s probably somewhere around 15 to 20%. So there’s no hard thing. But if I was an investor, and clearly we are all driven with that lens at a stage, if you look at our flow market share, you know you can do the maths and you can say that will probably translate into a stock market share at a level to …
BR: May I just interject at this point because that idea of you peak at 15, 20 percent, isn’t that an industrial age stuff?
There is nothing to stop us from creating a monopoly. So we’re not like a financial service provider. Whether we get to do that depends on the decisions we make about building up the business. A lot of that has to go back to what I said earlier about momentum and scale and the ability to keep creating additional value add and value adding services for the participants of the platform, in different regions. — MM
OP: No, no, no. It’s the regulator. So the problem is, no, I absolutely agree. They probably could be and maybe even with network effects — because the winner takes most of the market — which is actually one of the debates we’re having with all the social networks, should the regulator not start to intervene because you have all of us, right? But we fundamentally believe in financial services. There may be niches where this doesn’t apply, but in general, 15 to 20% in any country exposes you to regulatory intervention risk. So that’s number one.
Number two is, and that’s the second thing is because platforms take a while to build. This is very different to a product vertical that you were talking about Europe with 20 you know or Revolut that goes into you know 30 markets, 45. These are very effectively they are single product or similar product centric propositions that very rapidly can go across because all they do is marginal change. If you’re a platform like ourselves, it needs to be rich and deep and the markets we look at are really mega-markets that could actually each be quite significant and we will take our time, and the interesting thing coming back to the network effects of we learning cause I’m out there speaking to quite a few, the interesting thing is clearly on the members side, so there’s no network effect, right? However, interestingly on the product providers side there’s massive, because they look cross-national and where you know we would have seen in the UK even 12 months ago, certain people saying we’re not looking to speak to you, in many of these markets, people are moving very, very quickly to work with us because there is a degree of halo effect from the UK or wow, I want to be there and I want to be part of the early one.
So there is a degree of transferability across that. There’s clearly the whole thing about the technology and the learnings you can transfer, but I’m with you. It’s not strong network effects. On the other hand, I would argue that maybe a social network also, I mean I’m not sure, I’ve never seen the statistics, but when Facebook came to Europe, how many people were really interacting with the US? It must’ve been in the small percentages that have connections to the US of the whole European population. So network effects can be built if you’re there early. You start in the same way, bringing great product, bringing your learnings, and by just being very good at, articulate, going into that and being smart, which I absolutely… with going to deep accessible markets rather than very small ones. I think that good strategies can be built around them.
I love the fact that we’re a technology company. I mean, I would never want to be a bank because I think that model has so many constraints with it that you have to work within the system, whereas being a technology company, you’re completely unbounded and you can re-imagine completely what the value proposition in the marketplace looks like. — MM
MM: Yeah. Look, I think we’ve been very explicit about the fact that we’re trying to build an exportable, scalable business model with the capability for network effects going forward. So when you look at our business versus like a regulated sort of financial service provider nationally, there’s some very, very different facts and boundaries around our business. The biggest one being there is nothing to stop us from creating a monopoly. So it’s not like a financial service provider. Whether we get to do that depends on the decisions we make about building up the business. A lot of that has to go back to what I said earlier about momentum and scale and the ability to keep creating additional value add and value adding services for the participants of the platform, in different regions. Now in all probability, it’s quite difficult to create a monopoly, but there’s certainly economies of scale which will give rise to certain types of network effects, so what will probably happen is that you’ll fill certain niches and have much more penetration in those so it’ll be almost an effective monopoly if we execute correctly.
So I love the fact that we’re a technology company. I mean, I would never want to be a bank because I think that model has so many constraints with it that you have to work within the system, whereas being a technology company, you’re completely unbounded and you can re-imagine completely what the value proposition in the marketplace looks like. So we’re dealing with, I guess sticking to our knitting, which is we understand data, we understand supply chains, we understand enterprise software. So we’re trying to see how we actually mash up all of those topics that we have knowledge on to come up with a completely different way of earning value, which is based on digital data at the core of how supply chains work and how financial services get originated and distributed into those supply chains. I think the nature of that sort of, we call it open innovation means that we don’t actually know necessarily what the end destination is. We just know how we need to make money and what the next milestones are and we need to get through in terms of traction, scalability, etcetera in order to have the right to continue to explore that model.
BR: I want to revisit something you said, which is should the regulators intervene and break up Facebook? And I want to ask that question from the point of view of do we actually think that’s necessary or do we think that … So it’s a different sort of curve than supply side economies of scale, but do we believe at some point that network effects peak and they go into reverse? Because you could argue that we might be getting to that point a bit with Amazon for example. Because if you go onto Amazon, you want to buy like, I don’t know a colander now, you know that 20,000 colanders that you could choose from and so in the end it becomes too crowded. And if you think about something like Facebook, you can lose the customer trust if you don’t always act in their interest. And so the question is, do we think that in banking, it will be difficult to sustain network effects unless we are very careful about ideas of trust, for example. And how do you think about that in the context of your investments or your own businesses? So this idea of sustaining network effects and making sure that they don’t go in reverse.
OP: So from my perspective, monopolies are bad for the economy and we can go here or theoretically why that is. It just creates different value creation models internally because a monopolist is a monopolist and they can act. And as a matter of fact, I always say when we have certain competitors who I shall not name here, who are also trying to attempt Aurizon(?), it’s actually a greatest gift that we’ve got because it keeps us on our toes. Right? And I think that’s how very much when you look at the breaking open of the business banking market, what a regulator did or the government did by intervening with the PCR. They actually said this does not work for the economy as a whole. And as I think when you look at it at from that point of view, I think there’s a market reality around that that is proven that probably is more sensitive in financial services than in other places. But I just would be very, very wary. So I would not bank on it that the regulator and regulator, I don’t meet the financial regulator. I mean just the political class that deals with ensuring that there is sufficient competition.
On the network effects, I think that can be a thing where they can go into reverse. I’m not sure whether fundamentally I have seen it. I mean you could potentially argue and you’re probably better expert on it as I`m going a bit out of my territory, specialist networks coming in, so by Facebook being so generalist, maybe a LinkedIn was able to emerge because just like I’m for example, not personally affected. We decided our children should not be sort of exposed. We don’t want to post our family pictures or not. However I need to be and I want to be on LinkedIn because it’s a specialist network and is very much suited to my thing rather than a generalist network like Facebook that started off and is now trying to branch off into all these specialist arenas. So there may be a sort of converse effect from that, but I’m not sure.
I’d say in the beginning, it can be poor product that can accelerate network effects. You mentioned LinkedIn, you might remember monster.com. So that kind of became a large business driven by network effects before LinkedIn came. And so I think that’s actually a case in point of eventually poor product falling on its own sword despite the network effects that existed there. And so for me, ultimately, it’s not about sort of internalizing and externalizing. Ultimately it’s about category dominance. And so the way that we think about opportunities to invest is we don’t just look at a market that’s large. We’ll look at a specific value chain and where the business positions itself within that value chain. — EP
BR: So I agree with you. I think there’s a tendency to sort of unbundle big things, which I think we might be going through now. But for example, how would you feel about advertising on Tide? You’ve got 120,000 SMEs on there. That’s a valuable market for advertisers. Would you allow advertisers? Do you think that starts to introduce this conflict of interest between you and your end customer that might ultimately weaken some of the…?
OP: So for us, we are at that level. We wouldn’t even think of network effects. We would think of, is it in the interest of Tide? That’s our interest of our members and in the interest of Tide? In the interest of our members, we always think that’s the operating system which we want to have and we want to rethink the values there to weave everything together. And then there’s the application layer and we do two things on the application layer. So these are things like advertising for third party products. One is we have a community where we actually for free at the moment allow third parties, like some marketing agencies to promote their services and so on. And we also believe that over time we will allow a developer API that basically allows people like inventory management and things that are not really relevant for the platform itself — we call this the ecosystem around the platform — to interact with it. However, and maybe that’s a scale and prioritization story. For us at this stage, it’s not a priority, but strategically we thought it through and we would be quite open to them.
MM: I would add a couple of things to that. I think in principle, monopolies are bad for society because they lead to concentration of power and ultimately abuse of power. So there’s a level of once you get beyond, that there should be rules and regulations to stop that abuse of power from happening in a network, but I also think that, in the example of Tide, there could be different ways you see precipitation of value creation outside of the network effects that we talked about. So for example, with the providers that you deal with in those product categories, the way that they deliver services or the services that they actually deliver could be usurped. I’m thinking specifically around the accounting service providers platforms. I mean, who wants to pay a monthly fee for accounting services platforms? It’s actually a relatively simple and utilitarian service to do now. Why can’t it be usurped and subsumed into some other kind of a service and bundled up?
OP: Yeah, so maybe do two things on that. One is actually we can squeeze things around it because basically, we look at profitability across a connectivity chain and that’s why we can fundamentally, if we wanted to actually beat. We can offer services for free. Take invoicing as an example. They are providers of premium invoicing services. We don’t need that because we make money on the factory, right in the connectivity chain and because we have the customer with no marketing acquisition costs, no overheads to that, we can do that. So that’s already because it’s now a member’s interest what we do. Now, we had a debate that probably goes a little bit to what’s that is should we ask product providers to introduce customers to Tide? And we actually took the view as look, in the end, the platform and the member’s address is that it’s a rich and narrow since we had a lot of investors getting very excited about it. Oliver, your CPA will get a lot lower, and you will dominate it. But we argued against it because we took the view, and maybe we stand to to regret it, that actually we will win in the long run by having the best products on the platform, and if there can be reciprocal agreements, we will. But if they don’t want to do it, we don’t. And we’ve seen people change their tracks completely from a year ago. And maybe this is network effects to now, where they still think they can … they will build their own ecosystem and we say you do it. Do it. Try your thing and then maybe in two years’ time, we speak again. And a bit like what you say, maybe they are in a degree, some of them will make it…
MM: But you’ve got to curate the right products then within your walled garden network for that to happen, and I think the point I’m making is that those services that you’re creating could actually innovate very, very quickly and even disappear and get subsumed into other models. So it’s how well you create the product services that you provide within those verticals.
OP: Well actually one of the reasons why we don’t self-manufacture a lot of products was in addition to that, the obvious early start-up things like financial resources and focus and argument. We just raised £44 million, and we have a lot of financial power. The main thing is we actually believe product verticals. Now it’s hard to tell people that run product verticals this, but they should focus on creating great product vertical focus, especially when in the SME sector there’s new CPAs are so high on the product vertical side that’s actually very hard to run highly profitable business models and therefore we believe they will come around anyway because we cut out the CPA effectively.
However we do want them to innovate and we are actually very, very open. You asked about how prescriptive you are. We actually work very collaboratively with them. We say, look, we actually want you to try new innovations with our members because we believe a lot of the product level innovation needs to come from that one. Then we have, between how you cut it, between 20 to 50 products on the platform, depending on his credit one or end to end right now, however you cut it. And we actually want the product verticals partners to actually come with innovative solutions, provide them on the platform rather than us being too prescriptive. So we have a bit of a walled garden. The walled garden is really around two things, assuring quality, whatever that is and making sure we can weave everything together through the orchestrator connectivity chains. It’s less about stifling competition and innovation.
EP: If I may sort of echo this, in terms of product, I couldn’t agree more. I think a great product ultimately wins. I’d say in the beginning, it can be poor product that can accelerate network effects. You mentioned LinkedIn, you might remember monster.com. So that kind of became a large business driven by network effects before LinkedIn came. And so I think that’s actually a case in point of eventually poor product falling on its own sword despite the network effects that existed there. And so for me, ultimately, it’s not about sort of internalizing and externalizing. Ultimately it’s about category dominance. And so the way that we think about opportunities to invest is we don’t just look at a market that’s large. We’ll look at a specific value chain and where the business positions itself within that value chain, because you won’t be attacking a very large market. You might not be playing in the right part of it. The ones that the most painful was ones that has the least amount of competition or the one that really has a lot of meat to dig your teeth into and so far that is critical and once you figure out that part of the value chain, then you define what’s critical for you from a product perspective, what it is that you do yourself, what it is that you partner for and whatever eventually constitute your mode and you will eventually define your own category.
Ultimately I wouldn’t say that you necessarily get a huge premium for the network effects. It’s more your ability to have your cost base being linear, whether it’s your revenue is exponential. And so whether it’s network effects that that got you there or whether it’s your virality factors or your go-to-market, ultimately the great businesses will get the right price tag, regardless of what made them get there. — EP
BR: I have one final question, which is about valuing companies with network effects because if you think about our traditional tools for valuing companies like DCF for example. You take a very low terminal growth rate, right?
EP: You’re assuming that there is an EBITDA? :)
BR: Okay, good point. So this is a multifaceted question, right? Which is how do you value a business that has terrible unit economics but in theory once network effects kick in, will have wonderful economics. How do you value a business that’s got potentially exponential growth characteristics when our tools are not built for exponential growth characteristics? So I think I’d like to get all of the views because you’re investing in companies, you’re seeking investment, so how should we value companies that have network effects?
EP: If I may sort of attaining a theoretical point, you can have negative EBITDA and great unit economics. So for me as an early stage investor, I am conscious that my businesses are not profitable nor they should be at the stage where I invest, or maybe even the stage after, or a couple rounds after that. What we do look at in a great amount of detail is unit economics working… so at the unit level, how is my payback? If my payback is not great, is my net retention wonderful? So it means I’m overcompensating on the existing user base so that we’ll spend a lot of time on. Ultimately I wouldn’t say that you necessarily get a huge premium for the network effects. It’s more your ability to have your cost base being linear, whether it’s your revenue is exponential. And so whether it’s network effects that that got you there or whether it’s your, virality factors or your go to market you know, ultimately the great businesses will get the right price tag, regardless of sort of what made them get there.
BR: So you said that you don’t get a big premium for network effects?
We’ve been raising money, we find that the investors we spoke to honestly couldn’t give a c*** about network effects. All they care about is financial fundamentals and team. That’s it. All they care about is how much money are we currently making? What’s our attraction and how fast are we going to grow our revenue? And when are we going to make profit and how are we going to make profit, and do we have a good team and do we believe we can execute? That’s all they care about. — MM
EP: Not necessarily. What I’m saying is that it’s ultimately great businesses get great valuations and great prices. And then maybe network effects has got the merit. Maybe economies of scale and maybe virality. I think what you want to see is the exponential growth in their users or their revenue and linearity in their cost base eventually. I think at an early stage it’s not about profitability. I think it’s about sustainable unit economics. And then eventually every business ultimately becomes a financial asset that can be tradable on a stock exchange and face the harshness of the public markets, but typically you got at least 10 years before that happens.
MM: I think from our perspective, we find this quite interesting on the other side. So when we’ve been raising money, we find that the investors we spoke to honestly couldn’t give a crap about network effects. All they care about is financial fundamentals and team. That’s it. All they care about is how much money are we currently making? What’s our attraction and how fast are we going to grow our revenue? And when are we going to make profit and how are we going to make profit, and do we have a good team and do we believe we can execute? That’s all they care about. I mean, all the other things we were putting into our story in the early days, we started talking to investors became superfluous to how we actually ended up raising money. All comes down to the numbers when we’re looking for investors, for investment people, do they believe it’s a big market sector? Do they believe people are spending money or do they believe they will spend it with us and are we the company that can execute best?
BR: The bit where I struggle on that is, so I get everything you just said, except as a potential investor, I’d guess I’d want to also know when I’m mapping out the trajectory of your growth, I want to know is that underpinned by network effects? And when I look at the defensibility of your margins, for example, I want to know are they protected by network effects?
EP: There are other ways to build a defensible business. It’s not just network effects. I think sort of to your point, it’s more about whatever it is that constitutes your mode is by your ability to execute on it, and then it comes down to your team. It comes down to the sophistication of your go-to markets, etcetera. Right? So there are multiple ways to skin the cat.
BR: So the doctor Oliver Prill, economist.
OP: There’s Another Oliver Prill on LinkedIn that is not very keen on FinTech, and so we did that switch 15 years ago or something like that or 10 years ago. But just to give you my very quick answer to this. So I think people get in our space very mixed up between short term steering mechanisms and valuation. So short term steering to me is unit economics. We have a concept called structural profitability. So this is adjusting for no growth and no forward looking investment to have that break-even. Now as far as valuations are concerned, to be very honest, most VCs value through multiples and therefore it’s a question of which comparative multiple you do. I personally, because I’m also let’s say not an insignificant shareholder, I need DCF and my personal answer would be …by the way, I retro-engineer it.
So every valuation we have, I say look at our five year plan and then say how big does a terminal value need to be to justify that? And you know, do I sleep safe at night? And I can tell you with our current valuation I can. I would argue in the long run, your network effects need to basically reflect in the cash flow. Otherwise, they’re not worth anything financially. And the way to look at it is they should basically have that. Now you can argue that WeWork never had that and probably the valuation was way off. Now with an Uber, the interesting debate to be defensible on it is, is it that they had network effects, but they didn’t really properly leverage them. That’s why they got correction, or were they actually overvalued by the private market? So it’s actually an interesting one, which way it is, where they too slow to leveraging them or where they actually are not valuing them. I think that’s an interesting one. Maybe I know that you’re running out of time. Maybe an interesting one to do sometime.
BR: Yeah. Maybe we’ll do a part two on private market valuations versus public market valuations. Okay. So, thank you very much for that debate. That was excellent. Thank you. Thank you for your time.
All: Thank you.