A simple altnet model
Across the UK, Openreach, Virgin Media and over 100 altnets are deploying fibre to the premise. The economics of this deployment depend heavily on the cost of capital, since significant upfront investment is required. Another key variable is the cost-per-premise-passed (CPPP), which is generally substantially higher in rural areas. And of course, the average revenue per user (ARPU) is a critical factor.
To understand how these factors interact, we have built a simple model (embedded below). The model takes these input variables to calculate how many operators are viable in a given area. In low cost areas, multiple networks can build, and still hope to recover their cost of capital. However, in high cost areas only a single network (or none at all) might expect to recover its cost.
The coloured table below shows how the number of viable players changes, depending on CPPP and cost of capital. For example, in the top left corner we see that in a region with CPPP of £200, if operators have a 6% cost of capital, then four networks will be commercially viable. However, if the cost of capital rises with increasing interest rates, you travel down the table, and fewer networks may be viable.
The blue numbers above can be varied - click on them and input your own assumptions to see the impact on the number of viable networks. For example, if you set long term FTTP penetration to 90% instead of 80%, then commercial deployment - that is, at least one network being viable - extends to include homes with £1000 CPPP at a cost of capital of 9%. Increasing wholesale ARPU to £16 has a simlar effect. (Simply refresh this page to reset the model to its defaults).
A note of explanation on 'ARPU reduction from competition'. Operators worry about overbuild because it divides the available customer base between multiple players, and this is undoubtedly a major issue. But competition may well have an impact on pricing and/or marketing costs. To allow for this, we assume that each additional player beyond the first reduces ARPU by 5% - so ARPU is 10% lower in an area with three players, compared to one with only one FTTP network. You can vary this assumption, but the impact is primarily on the lowest cost regions, where intense competition is more likely.
This model highlights a number of issues:
- Overbuild per se isn't a problem - it depends where it happens. In areas with CPPP of around £300, multiple deployments are viable
- Rising interest rates may have a material impact. A 2% increase in cost of capital could reduce the number of viable networks by 1 across much of the country, as well as narrowing the range of commerically viable deployment
- The impact of price competition could similarly reduce the number of viable networks by 1, but this impact is limited to the lowest cost regions
One note of caution on the model's results. In reality, regions don't have homogeneous CPPP. A region with (say) £700 CPPP might broadly only support one network. But if it is a mix of materally lower and higher cost premises - say village houses and surrounding farmhouses - then it might be economic to overbuild low cost premises, but not the whole region. This form of partial overbuild is likely to be important in regions with diverse CPPP.
Technical notes
This is a simple model, and as such makes a number of simplifying assumptions. For example:
- The cost-per-premise-passed is assumed to be the same for all players in a region. In reality, those with existing infrastructure or more efficient operations may have a lower cost to deploy
- The available customers are assumed to be shared equally between all operators. In practice there will be variation between players
- Retail operations are ignored - in practice, the costs of retail operations are assumed to match the difference between retail and wholesale ARPU
- There is no ramp-up of penetration - operators are assumed to achieve there long run penetration immediately
- The cost of connection is assumed to match the connection fee, so that both can be ignored
For comments or questions on this model, please contact Rob Kenny