Overbuild is a widely used term in the context of broadband deployment. Put simply, it refers to two separate entities (usually ISPs) offering internet services at the same geographic location, to the same people, through their own separate infrastructure. Given the relatively high costs of rolling out physical broadband infrastructure to an area, and the resulting fight for take-up between any potential competitors, this is something broadband providers in particular try to avoid unless they are sure there is still money to be made. FTTP deployment is the prime example of this taking place today.
However, despite the simplicity of the concept there is a lot more difficulty in calculating overbuild, especially if you want your numbers to actually mean something compared to someone else’s estimates. The reason is that overbuild is relative, and it is hard to decide what it is supposed to be relative to.
To explain this, let’s take a look at the diagram below, where each square represents one household in a country of 15 households in total. 10 of 15 households are connected to broadband, i.e. they form the market. 5 of these households in the market are signed up to ISP 1, 7 to ISP 2, 1 to another ISP.
Thus, the overbuild is 3 households (marked in purple). But how do we represent this as a proportion or percentage of households? This is where it gets tricky and inconsistencies arise.
While we would most commonly refer to overbuild as a matter between two entities, in this case ISP 1 and 2, the providers do not have identical footprints and so we must decide who is overbuilding on whom. From the perspective of ISP 1, 3/5 households are overbuilt by ISP 2, or 60%. But ISP 2 has a footprint of 7 households, hence from their perspective overbuild is 3/7 households or roughly 42%.
If you are not an investor or employee of either ISP and have no affiliation with them, you might not like either of these perspectives and instead want to calculate overbuild from the combined footprints of both ISPs, in which case overbuild would be 3/9 or 33%. In other words, out of the combined footprint of ISP 1 and 2, 33% of households are covered by both providers and therefore overbuilt.
That’s not it, though. What if you wanted to calculate overbuild for a region or country? In this case, you want to look at overbuild from a wider perspective, where you have the choice of looking at either all households in the country or all households considered within the market. That gives you two competing measures of overbuild, at either 3/10 (10%) or 3/15 (20%) of households.
As you can see, and as is often the case with data, a little bit of mathematical gymnastics produces a wide array of different “answers” to your question. Overbuild, while a simple concept, can be surprisingly tricky to calculate.