5G threatens the mobile competition model (Part 2)

The first question of course determines the importance of the problem : if owning pre-existing fiber only provides a small advantage over competitors who don’t, then sure, some will be there a little faster, but competitors will catch up soon enough. If it’s a massive difference, then the first mover advantage will be impossible for others to gap.

I should stress that what follows is by necessity a high level abstraction : the existence of not of open deep fiber in a given area, population spread, highways, industrial sites, all these elements will drive different economics for small cell 5G deployment. Hence the second question which I will answer below.

The question of the cost advantage revolves essentially around one key aspect : how much does an FTTC/FTTH and a 5G small cell deployment overlap. As mentioned in Part 1, if there is no fiber deployed, the playing field is somewhat level, down to investment capacity rather than established advantage. The FTTH Council recently published a study undertaken by Comsof on this very topic. Comsof is a Belgium based software and consulting firm specialising in tools to enable the modeling of fiber deployment costs.

What this study tells us is that if an FTTH network already exists, the deployment of 5G small cell anyhaul is 65 to 95 % cheaper than if the fiber needs to be deployed. The variation depends on the topology choices of the FTTH deployment. Of course, anyhaul deployment is not the only cost component of a small cell 5G network deployment and operation, so this number needs to be put in perspective.

Talking to various 5G vendors, I managed to get a general sense of how much the anyhaul costs weighed in the overall business model. The consensus is that in a greenfield deployment, it's around 20% of the 5 year total cost of ownership (both capex and opex). I can't quote any publicly available source on that, so I'm happy to be pointed to such sources if they exist. If these numbers are correct, it means that a preexisting FTTH network would deliver a 5 year TCO between 13 and 19% lower than if the fiber needs to be deployed from scratch.

TCO does not translate directly into margin, but that gives us a sense of the massive advantage a player who already owns the fiber would have over a player who needs to deploy it. And that's not even accounting for time to market which would obviously another crucial advantage for an established network.

So how do you refine that number and assess the impact on a market by market basis. It's not simple, and with business partner Phil Marshall of Tolaga Consulting we have been working on a methodology to assess this. The goal of this approach is to be able to assess the number of viable 5G small cell networks on an area by area basis in a specific geography. By geotyping areas of a given country based on the key criteria that determine the need for 5G small cell coverage (population, roads, industrial zones, etc.) and the cost of infrastructure deployment (available spectrum, existing infrastructure, layer 0 resources (ducts and poles), cost of trenching, etc.) we are able to build a model that tells us when a given market player's economics break down in deploying 5G.

This model can then be played around to assess the impact of infrastructure sharing and other remedies on 5G small cell coverage and competition. We believe that this is something that governments, especially those who seem gung ho about 5G should be assessing, ideally before allocating spectrum : you don't want to sell expensive spectrum to players who won't be able to make use of it because they lack the infrastructure to deploy.

This is not an insolvable situation, but like any such conundrum, it starts with assessing the situation as accurately as possible. Once that is done, I can envisage several types of approaches to remedy the issues :

1. The market will provide. If there is a void in fiber anyhaul for 5G, economic theory suggests that private players will step in to fill that void. This may be partially true in some markets, especially where wireline players with an affinity for wholesale already exist. This is a clear opportunity that companies like Crown Castle are already pursuing in the US, but also for existing neutral fiber networks in countries like the UK and France. However, there are some caveats : first, the mobile industry needs to be open to a neutral fiber solution and feed information on where to deploy ahead of time. If these resources are to be shared, market players will need to agree on their needs ahead of time. Second, it should be recognized that mobile anyhaul alone may not make a business model sustainable for a neutral wholesaler in challenging geographies. In those areas, intervention may be needed to either incentivize market players to pay more (through coverage obligations maybe ?) Third, and perhaps most crucially, few players emerge from scratch as wholesale fiber operators : it takes time, skill and funds to become such a player. In markets where they don't exist, that's an additional challenge.

2. Force existing networks to open. In the same way that copper unbundling created a competitive market in wireline in the early 00s, existing FTTH networks could be forced to unbundle dark fiber or at the very least lease capacity suitable for mobile anyhaul. This is the path that South Korea seems to have followed although details are sketchy. It makes sense if you have a pre-existing network even if (as anyone who's worked in wireline wholesale in the last 20 years can tell you) figuring out fair prices for wholesale is tricky (and usually makes everyone involved unhappy).

3. Public investment in a neutral 5G anyhaul network. This is similar to the policies implemented in Columbia and Mexico (although Mexico goes further in sharing RAN elements as well). It seems to have worked brilliantly for Colombia (although not at the same level of density than what would be required for small cell 5G). For Mexico, it's too early to tell. This is probably the best solution, especially for emerging markets where no fiber network exists today. But this also takes time, which means that governments should be looking into this already.

There are other sharing mechanisms that I think will not deliver. MVNO set-ups in particular, where an operator who cannot invest in its own Small Cell infrastructure would lease access on another operators network would be a fool's gamble. It's hard enough for anyone to figure out where the additional margin needed to finance 5G investment will come from, and MVNOs make piddly little margin as it is...

In conclusion, there are solutions that could be implemented, but for players to move decisively, governments need to be aware of the issue and measure its extent. Tier 2 market players need to admit that they may have a problem and find ways to address it as well. And sure, 5G is years away still, especially small cell implementation, but the remedies outlined above will also take years, so there really is a time crunch to all of this.