Benoit Felten of Diffraction Analysis Talks Global Strategies - Community Broadband Bits Podcast 266

When policy and decision makers discuss how to improve connectivity in the U.S., they often compare Internet access in other parts of the world to connectivity in America. We can learn from efforts in other places.

Benoit Felten, CEO of Diffraction Analysis, has analyzed business models, approaches, and infrastructure development all across the globe. His company has studied infrastructure and Internet access from short-term and long-term perspectives through the multi-faceted lens of international economies. Benoit joins us for episode 266, his second appearance on the Community Broadband Bits podcast.

In addition to development of infrastructure, Christopher and Benoit get into competition, quality of services, and how it varies from place to place. Benoit has recommendations based on his years of analysis from different communities and cultures around the world. Be sure to also check out episode 21, in which Benoit and Christopher discuss Stokab.

This show is 40 minutes long and can be played on this page or via Apple Podcasts or the tool of your choice using this feed

Transcript below. 

We want your feedback and suggestions for the show-please e-mail us or leave a comment below.

Listen to other episodes here or view all episodes in our index. See other podcasts from the Institute for Local Self-Reliance here.

Thanks to Arne Huseby for the music. The song is Warm Duck Shuffle and is licensed under a Creative Commons Attribution (3.0) license.


Benoit Felten: Japan and Korea would be forward-thinking businesses, then Europe would be short-term businesses but forced to look at the long-term through policy, and then the US would be short-term businesses, laissez-faire, do what you want.

Lisa Gonzalez: This is episode 266 of the Community Broadband Bits podcast from the Institute for Local Self-Reliance. I'm Lisa Gonzalez. Benoit Felten is back on the show to talk more about connectivity from an international perspective. He last visited with Christopher way back in 2012 for episode 21. This time they discuss several models that his company, Diffraction Analysis have studied in areas other than the US. Learn more at the company website Before we start the interview, we want to remind you that this commercial free conversation is not free to produce. Please take a moment to contribute at If you've already contributed, thanks. Now here's Christopher and Benoit Felten from Diffraction Analysis.

Christopher Mitchell: Welcome to another addition of the Community Broadband Bits podcast. I'm Chris Mitchell with the Institute for Local Self-Reliance and today I'm speaking with Benoit Felten, the CEO of Diffraction Analysis. Welcome back to the show, Benoit.

Benoit Felten: Thanks for having me.

Christopher Mitchell: We last talked about Stokab, I think in Stockholm. You are the CEO of Diffraction Analysis which does telecommunications research all around the world and I often think of you as my go-to person on how things work outside US and sometimes inside the US. Let me ask you, Benoit, when you hear people saying, "The United States sucks at broadband and Europe is so amazing." How do you react to those monolithic statements?

Benoit Felten: Yeah, well I think that's generally true. I mean, the problem is always that broadband is as good as where you measure it, right? Of course there are places in the US where you can measure broadband and it's really good. The other thing that very few people realize is that independently of the broadband infrastructure, the US remains, by far, the largest hub for Internet content. Whenever anyone in the world wants to access content, nine chances out of ten, it's going to go through the US. Which means that if you measure broadband in the US, actually you get a home corner advantage despite the infrastructure. Actually the fact that you get better broadband in Europe shows how bad the broadband infrastructure is in most of the US, because you have that home corner advantage. Having said that, I mean, there's no magic about this, the quality of broadband is directly linked to the quality of the broadband infrastructure. I think generally in the US, in the last 20 so years, there's been a prevalent world view that the private market will provide and it hasn't. Generally, this laissez-faire thing, that's probably culturally much stronger in the US than anywhere else in the world, has failed you. In a lot of other countries in the world, policies that have been put in place have changed things. It's not true absolutely everywhere, but essentially in our world view, and based on the research that we've done there, there are three models that have worked.

Christopher Mitchell: Before we get into those models, I'd be curious, there are places in Europe with subpar access, right? I was recently speaking with a friend who has family in southern France and they have very good access but I remember speaking, and maybe it's changed, but 10 years ago, I was speaking with someone from the French equivalent of their federal communications commission and he was saying that they didn't even have a plan to cover the 10% most rural areas of France at that time.

Benoit Felten: Yeah, so they have a plan now but that doesn't mean it's done.

Christopher Mitchell: Okay.

Benoit Felten: In fact, our new president has, if I may be so blunt, gloriously put his foot in it by coming in and saying we need to accelerate all of these broadband deployment plans, which, of course, everybody claps at. Then it turns out he doesn't have a clue what is required and basically what he was pushing for is already in place, which means that it's not going to make any difference. Yeah, I mean, the challenges of rural broadband are more or less the same everywhere, then it's a matter of demographics and geographical layout. Some countries are advantage like the Netherlands, because basically it's all flat and it's all mud so basically it's very easy to dig and you don't have to go through any solid rock anywhere. There's geography, Norway is a nightmare. It's solid rock all over the country so an average home build in urban Norway is already more than twice the price of what it would be in the US on average. That's just because you have to dig with heavy machinery just to get to the house, wherever that is. Rural broadband is a general issue all over the world and all over Europe. I can tell you what we think a good solution might be but I don't think a lot of countries are anywhere close to that.

Christopher Mitchell: Okay.

Benoit Felten: Basically the prevalent model for rural broadband is, "We'll throw money at it." That works well until the government says, "Oh my god, that's a whole lot of money," and then suddenly it stops. That's kind of where we are. There's no working solution for rural broadband today and also there's a lot of ambiguity in what we call rural. In some instances you might be talking about the last 25%, in some instances you might be talking about the last 5%. Again, that's down to geography and demographics and how forward thinking the operators may have been.

Christopher Mitchell: Yeah, I can imagine that rural Switzerland is more challenging than rural Netherlands.

Benoit Felten: Yeah, although that doesn't necessarily mean you get better broadband in rural Netherlands than rural Switzerland, because it so happens that Swisscom is one of the forward thinking operators. Even though they're absolutely not deploying fiber to the homes everywhere, they're thinking about how to improve broadband everywhere, which a lot of big common operators aren't really when it comes down to it. The main reason we can see that things are not working in the US is that actually urban broadband is comparatively crap. We're not even getting to the rural.

Christopher Mitchell: Right, right, that's a good reminder.

Benoit Felten: Right, so having said that, you have countries like, obviously the Nordics and Europe and France, now, and Spain, very much so. Spain is very much ahead of the curve, so is Portugal. A country like that, you don't necessarily get fiber broadband or fiber quality broadband everywhere in urban areas, but mostly that's done. Or if it's not completely done, it's going to be finalized within the next couple of years. By the way, the policy approaches have been different in these different countries. Then you have a few examples of the market will provide that actually worked. Korea and Japan actually are in as much as there is state involvement in the equity of the incumbent operators. You could argue that there is some state involvement, but having said that, it wasn't a heavily subsidized public plan that led to fiber being deployed. It's very much a cultural thing where these kinds of infrastructure industries like rail or telecom have always been in the mind frame of providing for the next 25 years. In 1999 when NGT started deploying fiber, it wasn't necessarily because the government was saying, "Ooh, we need very fast broadband," because back then, five [Mbps] broadband was considered fast. It was more like, "Is copper going to be good enough for 20 to 25? No, so we need to be slowly rolling out what we will need already." By now they're 95% coverage and they've done this over close to 20 years. It's actually, from an investment point of view, it's not a massive burden because it was spread out across a long period of time. Now obviously if you look at that amount, you turn to an investor and say, I need to invest that in the next three years, your investor's going to say no.

Christopher Mitchell: Right.

Benoit Felten: If you plan it and if you think of it as network renewal, which is how they looked at it, then it's actually not that painful an investment, but you have to plan it ahead of time.

Christopher Mitchell: NTT, of course, is the Japanese incumbent operator. Now, so you mentioned that there is some state investment in NTT, but that it's primarily a cultural thing that they're focused on long-term investment rather than the shorter term returns that we see in the United States.

Benoit Felten: Yep, absolutely. Not just in the United States, by the way, I mean, the reason that European policy has had to be a lot more involved in this is that that long-term culture doesn't exist here in Europe either. The thing is, that, again, it doesn't exist in the US but very few companies have that long-term perspective and the policy isn't getting involved, it's like one step further behind. If you wanted to see it like that, Japan and Korea would be like forward thinking businesses, then Europe would be short-term businesses but forced to look at the long-term through policy, and then the US would be short-term businesses, laissez-faire, do what you want, which is short-term by definitions.

Christopher Mitchell: You spend a lot of time in Hong Kong and you recently had spent some time in China and Shanghai, I believe, how did those cities compare? What's some of the policy, maybe starting with Hong Kong?

Benoit Felten: I rarely use Hong Kong as an example for the simple reason that it's a very small market, it's seven million people, and it's a hugely dense urban market, which means that the cost of deployment are relatively low. There is no policy involvement in Hong Kong. Hong Kong is a capitalistic society in the sense that basically a few billionaires own most of the real estate and if you own the real estate, you own more or less everything. They own most of the large utilities and companies and everything and so it's basically their decision. What happened in Hong Kong was, it was a perfect example of competition actually doing its job. In 2003, a very small company that was doing IP telephony at the time, started to deploy fiber into the home. They became Hong Kong Broadband Network and basically they understood that this was a long-term business, probably way better than the incumbent did, at the time. They started deploying with private investment that backed them for the seven or eight years where they were losing money, because they were investing so much. After eight years, when it started turning around, the profits they were making just became insane because the paradox of this is that fiber networks are actually a very, very sound investment. Compared to highways or office buildings or anything like that, they require very little maintenance. The cost to build is actually not all that high if you compare it to these other infrastructures, and the expectation of revenue is very high, but like any infrastructure investment, for the first 10 or so years, you're not going to make a dime. The problem is that, and this is what Hong Kong Broadband Network understood very well, the problem is that if you see this as anything else than an infrastructure project, you're going to fail, because your shareholders are going to expect profit within the first three to five years and that's just not going to happen. At some point, and this has happened in various places around the world, at some point, they're going to pull the plug. You're going to be arguing all you can saying, "Well, give us another few years and you're going to see a profit." They're going to say, "This is not what I paid for." This is one of the fundamental problems of where we're at. If these things were viewed as infrastructure investments, there would actually be no problem building a business case, but they're viewed as telecom investment and telecom, by definition, is a short-term perspective. That's where the entire problem is for this and this is true the world over.

Christopher Mitchell: Now when you talk about Hong Kong Broadband Network, were there others that -- You say that it's a good business and Hong Kong Broadband Network has certainly proved very savvy with a really great strategy, but I presume that there's a few others who thought that they had good strategies and lost to Hong Kong Broadband Network and then presumably went out of business.

Benoit Felten: There was one competing broadband operator in addition to the incumbent and the cable operator. The competing broadband operator is dead as far as I can tell and the cable operator is dying.

Christopher Mitchell: It's interesting to me, I mean, one of the things that we think of when we talk about Seoul, Tokyo, is having a lot of different choices, we can talk about whether or not that's correct or not. Now in Hong Kong-

Benoit Felten: That's not true.

Christopher Mitchell: Okay, so that's good to address then.

Benoit Felten: That's not true in Hong Kong. Again, as I told you, this is an unregulated market, as unregulated as they get, so there's no incentive for anyone to share their infrastructure. It is actually turning into a duopoly now. It was more of a, kind of, competitive market until now, but now with the smaller guys just dying off, it's turning into a monopoly, a duopoly network.

Christopher Mitchell: Do you envision Hong Kong Broadband Network taking advantage of its status then as basically almost a monopoly provider?

Benoit Felten: I'd say yes to a small extent. Basically, it gets to a point now where for HKBN acquiring PCCW's core set of customers, becoming more and more expensive and also because PCCW was very expensive and still is, their customer set is very different from Hong Kong Broadband Networks. Essentially within the areas where HKBN has very high penetration, it's very hard for PCCW to get in because they're too expensive. Within the areas where PCCW has very strong penetration, it's hard for HKBN to get in because people want more services than what HKBN was offering. The kind of cheap no frills offer that they have is not what people want. They want football league access and things like that, which HKBN doesn't have. I think, essentially, market shares are probably going to end up being fairly stable in the next few years, which does mean that HKBN can probably slowly increase it's tariffs, but they can't increase them too much because they need to keep that gap between them and PCCW from a pricing point of view. Anyway, to go back to my initial point, one of the reasons that I rarely use Hong Kong as an example, I do use HKBN as an example of a smart strategy but I don't use Hong Kong as an example of a market because that's actually not the kind of outcomes that we tend to want for larger markets. Again, remember, this is a seven million pocket of a market, it's completely disconnected from China. It's its own country effectively from that point of view and it has virtually no policy. For example, if you live on one of the islands, apart from Hong Kong island, you're screwed. You might get very old DSL that PCCW has no incentive to upgrade or you might get nothing and have mobile access if you're lucky sometimes, and that's that. No one has any incentive to deeply there because the density is too small and you'd have to pull undersea cable to connect the islands. They have, effectively, the same rural problems that you can find anywhere else, although rural Hong Kong sounds weird to a lot of people, it does exist. Because of the lack of policy involvement, these problems will not get solved.

Christopher Mitchell: Let's move over to Shanghai then. I mean, one of the things that I've been interested in is that, as far as I understand, hundreds of millions of people in China have fiber connections or soon will per state policy, and yet my understanding is your connectivity in Shanghai was not super reliable.

Benoit Felten: Again, this is a bad example from an international point of view. The first thing that you have to remember is that all three telecom operators in China, all three big telecom operators in China are 100% state owned. Whatever competition exists is actually between state entities. That's the first thing to keep in mind. The second thing is, fiber was deployed because the government said, "You're going to deploy fiber." When the government says something, you do it. I'm not even sure there's a business plan anywhere, or if there is, it's like a post-PAC business plan where you look at how much money you're going to make or lose, but there's no business decision so to speak. Okay? That's the second thing. The actual quality of the broadband infrastructure is fine. I could get 200 [Mbps] up and down to other Chinese cities, but the thing that very few people are aware of is that the great firewall of China is just one way in which the Chinese government blocks access to Internet content. Actually, the more devious ways that they underprovision international transit dramatically, it's probably a hundredth of the capacity that would be needed to deliver proper service. Basically you can ping a server in Beijing from Shanghai and get 200 [Mbps] and then you'll ping a server in LA and, if you're lucky, you'll get 800 kilos. It's nothing to do with the infrastructure, or rather it's something to do with upstream infrastructure but not access infrastructure. Again, we're back to what I was saying at the beginning. These are the kinds of problems that are actually solved natively in the US because there's no international transit issue for 80% of the content you might be willing to access. This is something that's often a little hard to understand for westerners in general where these problems are solved without them knowing about it. There are countries in the world where, especially like a lot of the Pacific Islands have the cost of international transit is like 100 to 300 times what it would be in LA. Now, when you have that kind of cost, then anyone watching a YouTube video is actually a measureable immediate cost to your business model, which means you have to charge them one way or another. In these countries, usually you have, no matter how good or bad the infrastructure may be, you have to throttle usage because otherwise the economics just don't work.

Christopher Mitchell: If I had to guess, I would guess that companies like YouTube and Netflix can solve that relatively easy with local cashing servers. It's probably a lot of, not necessarily the top five sites on the Internet like Facebook and others, or at least if it is today then it won't be in a few years because it's a relatively easy problem to solve if the companies that are delivering this content want to solve it.

Benoit Felten: It's a matter of volume, so if you're talking to Fiji, which is a few million people and most of them below the poverty level and very few of them actually connected to the Internet. Does Google have any interest in solving that for Fiji? Probably not. It becomes a financial decision at some point, and for Facebook, even assuming every Fijian went onto the Internet and used Facebook, adding four million Facebook users is trivial. It happens organically every other day or something like that.

Christopher Mitchell: They'll do it with satellites soon enough.

Benoit Felten: I'm probably exaggerating slightly, but you see my point. I mean, these problems are solvable but actually I think they will only be solved through policy, again, because someone will say, "Well, if these companies don't want to host and cache in our countries, then we need to organize that and make it happen on our own. We can cache without these companies agreeing to it effectively just by replicating the content that is downloaded the most but that costs money and so you have to get it organized and usually a single operator doesn't have the critical mass to do it only for his customers and therefore, blah, blah, blah. You get back into these sharing of resources aspects that have provided better broadband for a broader subset of the population.

Christopher Mitchell: Do you have a lot of choices in those areas, all of Japan and South Korea?

Benoit Felten: In Japan, there's a theoretical unbundling policy for fiber. I say theoretical because the economics are so bad that no one's actually using it. One thing that's important to understand, and unfortunately it's a complicated thing to describe, but whenever you implement network sharing at a passive level, which means the ability for a competing operator to use your infrastructure as if he had deployed his own and have full control over that infrastructure, you're hugely dependent on points of concentration. Where you enable access to that infrastructure is a key factor of the economic model. One of the elements of this is that if you deploy a point-to-multipoint fiber network, which is what is being deployed in Japan, you either have full access to the entire end-to-end connection to the customer in which case you might be able to connect thousands of customers from a single point, but that's a discontinuous point, so that means that multiple operators have to be able to access that at the same time so usually these models are very complicated to set up. What you do instead is you have access at the splitter and the splitter connects 32 customers or 64 customers, and that's actually too small for you to connect that point with your own fiber. The cost of doing that for potentially up to 64 customers, assuming you have 100% market share, is just not workable. In the same way that sub-loop unbundling, as this is called in copper, has never worked. There are numerous countries where policies are being put in place forcing the incumbent operators to open up their VDSL copper, so what you call fiber to the cabinet in the US, open up these cabinets for competitors but actually no one's buying because the cost of connecting those cabinets is too high. Then you have to buy from the incumbent operator the cost of going to those cabinets and suddenly you can't make money anymore because you're paying too much of a recurring cost to make any margin. The fact is that in Japan, there is also an active resale product, so basically competing operators to the incumbent can buy capacity and serve the customer with their own brand. Now, these are lower margin products. The service is actually on the active layer, their service is actually run by NTT. It's not exactly like white labeling but it's not very far from that. It would be comparable to what an MVNO is to a mobile operator.

Christopher Mitchell: Right, a company like Ting that's using the Sprint network and just providing customer service, effectively.

Benoit Felten: Yeah, well, a little bit more than that but, so it's a similar model to that. Now, you have a few competing operators in Japan, so NTT has competition from KDDI and a couple of other smaller ones, but these don't have 100% coverage so there's probably half the country where NTT's infrastructure is going to be what you're using anyway. Now, don't get me wrong, I think that's better than nothing. It could probably be optimized, but at least customers do have a certain amount of choice. In Korea, it's a pure infrastructure competition. You have essentially two very large players, the incumbent KT, and the competing operator SK Broadband. Both Korea Telecom and SK have their own independent fiber networks. As far as I know, there's no sharing of infrastructure or anything so you're kind of in a duopoly system as well.

Christopher Mitchell: We think of, I think, these Asian areas as being some of the most highly advanced networks and in no case is it resulting in robust competition. Now, in most of these areas you can get very high quality connections but there are limitations on each and to some extent you get a sense that US policymakers are really just trying to invent something that no one else has done before and not heeding lessons from areas we consider to be some of the most highly connected where you tend to get it from one to two operators, it sounds like.

Benoit Felten: It's been my experience that American policymakers and, in general actually, American decision makers in businesses in the areas that I cover, have a hard time even accepting the idea that any other country may have lessons worth giving. I remember this huge and very interesting report that Yochai Benkler did, was it 10 years ago?

Christopher Mitchell: Almost, yes.

Benoit Felten: He got slammed left, right, and center for even daring to suggest that there were models outside the US that might be worth getting inspired from. My company has kind of given up making business in the US for that reason, because the value that we bring is the experience of dozens of other countries where interesting things happen in circumstances that might be very comparable to what you might see in US cities. It's extremely hard to drive that message through, nobody wants to listen and even at municipal level, most people are just reluctant to agree that there might be lessons to be learned from outside the US. That's the way it is. I've kind of made my peace with it but I can see how frustrating it might be for you trying to push cities into making right decisions and trying to find interesting examples outside the US that might be replicable. I was very frustrated with, in the early days of Google Fiber because I got the impression that they were just reinventing the wheel. If they looked at what was being done, I swear they could've saved themselves a whole lot of bother and frustration and time and money. Because none of what they were doing was unique to the US and yet they were convinced they were doing something revolutionary.

Christopher Mitchell: Well, not only that, I mean, when I talked to some of the people that have been involved with this, you get the impression that Google hired major contractors that said, "We're going to reinvent the wheel and we're going to do it without using any curved surfaces." Because they were belittling of anyone that had any experience, it was stunning frankly. It's amazing how smart people hiring smart people can do so many dumb things.

Benoit Felten: I think there were three working models. I don't mean workable, I mean working.

Christopher Mitchell: In real life.

Benoit Felten: In real life to improve broadband infrastructure. The challenge of extending those models to rural areas, they're still out there. There's one model out of the three that I think works even in rural. One is to say you want infrastructure competition that's the closest to what you might have in the US if there was actual competition. The point is that you cannot get competition in the same territories if one of the players has a massive advantage in deployment costs. The fact is that the incumbent has a massive advantage in deployment cost because all of the underlying infrastructure, the ducts, the poles, that you need to deploy better fiber broadband, is already in place so you're shaving off maybe 40-50% of your deployment costs just from that. Which means that there's no competitor can match that unless there's a policy in place to address that. There are a few countries in Europe that have done that. It's actually, conceptually, very simple, and the principle is that any preexisting passive infrastructure has to be shared and accessible at the same cost bases by everyone. Portugal did this, they said, "Okay, every duct and pole that exists, whether it's electricity, telecom, anything else, is registered. We look at how much space is available in there, we establish a rental cost for this that is low enough that anyone can affordably deploy a network." It took three to four years to actually set it up, get it in place. Which seems like a long time, but you have to think about the fact that you have to build a national register. You have available infrastructure and space in that infrastructure. It's actually a massive undertaking. Once you've done that, anyone can enter the market and affordably deploy. In a country as small as Portugal, you have three quasi national fiber operators competing, and so it works. The incumbent, of course, has a very strong incentive to start deploying and deploying fast because they know that if they don't do it, someone else is going to do it.

Christopher Mitchell: Now, did the incumbent just decide to do this or was there a government edict?

Benoit Felten: No, they were forced to participate. Obviously they didn't want to at first. Over time actually I think they realized that the model was beneficial for them as well because it meant that in areas where others were deploying, where they'd never deployed the proper infrastructure, they could use that infrastructure as well. It lowered their deployment costs significantly as well and it sped up their deployment. Lithuania did slightly different but roughly the same. This is actually, in my view, one model that should be top of mind for policymakers in a lot of developing countries, because in developing countries you get a huge amount of real estate is built every year as lower middle class people kind of get up in the world and they want to move into proper apartments and proper houses and so there's a huge property boom. It would be absolutely trivial and virtually costless to just connect all of these homes with ducts and then have these ducts open on the market for anyone who would want to deploy. That's model number one.

Christopher Mitchell: Model number one is sharing infrastructure, basically.

Benoit Felten: Model number two is also a form of sharing infrastructure but it's a different form of sharing infrastructure.

Christopher Mitchell: Okay.

Benoit Felten: Model number one is actually infrastructure competition but there's a level of sharing the underlying network elements, okay? Model number two is actually a model of proper infrastructure sharing in the sense that basically you either only deploy one infrastructure nationally and everybody can purchase and access that. Or each operator deploys their own infrastructure but they all do it together at the same time, which means that you only dig once and they all share the cost of that dig. If you do that, then effectively, the cost of deployment is hugely reduced for everyone which means you can go further and still get a certain level of competition. Now, these two models, they work, but they kind of break down when you get into rural because the density becomes so low that only the operators who have the most money can still afford to play, and therefore, basically, pass the first maybe 50% or 60% of population, you get only one player and so you're back to monopoly. Even that monopoly will not go all the way to 100%, it'll probably stop around 65-70% of the population.

Christopher Mitchell: That's actually, oddly enough, where the United States, we have a benefit in that. The United States can just say, "We're just going to give money to this firm to do that." Now, in Europe I think that's harder to do because of the anti-state subsidization rules for different countries, but in the US I think when it comes to rural in part because of our ability for the government just to give money away in ways that is harder to do elsewhere. Our history of cooperatives, I think rural is more or less solved in the US once we get our heads around it. I think we can just focus on the cities for the majority of your three models.

Benoit Felten: I think there is a question about whether you want a monopoly anywhere. I think that there's a real challenge in addressing that point, a service monopoly, I mean. The third model is a structural separation model. We have one example worldwide which is New Zealand, which went down that route. Where basically not only do you have a single infrastructure nationally, but that single infrastructure is not owned by a telecom operator, it's owned by an infrastructure company that manages and activates that network but they don't sell. Only the retail service providers buy from them the wholesale access and then they resell it to the end customer. The beauty of that model is that because you get demand certainty in that infrastructure, the economics become unbelievably better. As a consequence, the prices can go down and the reach can go further out. We actually published a white paper last year where we built that model comparing a pure retail operation to a pure wholesale operation and we came to the conclusion that by standard market expectation in terms of return on investment, you could probably cover 95% plus of the country without any subsidies in that last model. That's not the direction that the telecom deregulation went, right? It all started in the US actually, because when the deregulation, when the liberalization happened and the old Bell monopoly was broken down, instead of realizing that the infrastructure was actually a natural monopoly but the services above that should be competing and therefore slicing it horizontally, if you will, with the infrastructure being owned by one regulated entity and then free competition on top of that infrastructure, they sliced it vertically, geographically. Which actually didn't make a whole lot of sense because you still had local monopolies and then unbundling came and created some competition and then unbundling was killed and that competition died and then cable started doing broadband and they said, "Oh, well, it's okay we have competition," but the fact is that there's no healthy model for competition when you look at it that way. New Zealand's a really interesting example. We're hoping that a few other countries go down that route. There's some signs that some of them might.

Christopher Mitchell: In the time we have left, I'm curious, can you just contrast that with Britain where I think people might think Openreach is kind of that approach with that infrastructure play?

Benoit Felten: Yeah, it's not actually, because the key aspect is that the infrastructure company has to be able, has to have full freedom to reinvest it's benefits in infrastructure. Because it's the long-term thing I was talking about earlier, you have to be in this for the long-term. In the UK, what happens is that Openreach is a separate entity but it's owned by BT.

Christopher Mitchell: British Telecom.

Benoit Felten: Whatever profit it makes, go back to the British Telecom group, for them to invest in whatever they want and whatever they want recently has been football rights and mobile spectrum, but it hasn't been infrastructure at all. The consequence is that actually the UK is slipping down the ranks in terms of the quality of their broadband because basically Openreach invested in the cheapest possible upgrades to the existing DSL network. While some might consider that acceptable in terms of performance, it's certainly way below what other neighboring countries are now providing to their citizens.

Christopher Mitchell: Thank you, Benoit. This has been very interesting and I think to anyone who is not familiar with you or Diffraction Analysis, I hope will get a better sense of the value you provide, your perspective that could be useful in any city around the United States in terms of lessons learned, but also that you're not just knowledgeable but able to translate it into very normal language anyone can understand, so thank you for that.

Benoit Felten: If I can just add something, is that we don't only focus on policy, we're also very interested in the operational challenges and the marketing and the selling and all kind of very practical aspects of building a better broadband network. Yeah, I hope some people find this interesting and I hope we can talk again soon.

Lisa Gonzalez: That was Christopher and Benoit Felten, CEO of Diffraction Analysis, talking about connectivity and business models in Europe and Asia. We have transcripts for this and other Community Broadband Bits podcasts available at Email us at with your ideas for the show. Follow Chris on Twitter, his handle is @CommunityNets. You can also follow stories on Twitter where the handle is @MuniNetworks. Subscribe to this podcast and all the other ILSR podcasts including Building Local Power, and the Local Energy Rules podcast. You can access them on Apple Podcasts, Stitcher, or wherever else you get your podcasts. Never miss out on our original research, subscribe to our monthly newsletter at Thank you to Arne Huseby for the song Warm Duck Shuffle, licensed through Creative Commons, and thanks for listening to Episode 266 of the Community Broadband Bits podcast.