The UK today has more broadband operators deploying more networks than at any time in the past. The introduction of infrastructure competition, the separation of Openreach and the influx of investment supported by increased demand means many consumers are now able to access at least a gigabit of downstream bandwidth and some have a choice of suppliers.
In the coming years we will see these footprints expand, at least for FTTP/B/H[1], and as they do so they will increasingly intersect and overbuild offering even more choice to many. Not for all however; we are expecting 17.6% of UK households to have access to only one gigabit capable network and 3% without any by 2030.
We have been forecasting availability coverage in the UK for some years and the latest version, laid out in this paper, is the first to be completed in the newly competitive market.
General and specific outcomes
We are projecting that the majority of UK postcodes, around 75%, will have the choice of up to three gigabit networks. There are more than 20% where the current plans and operators in the market mean four or more networks in a postcode with a small number left behind.
Tracking the deployments and identifying at a point in time exactly where they will be is challenging and likely to be wrong in many instances. This output is based on a model and as such does not allow for all the elements that direct deployment decisions or particular outcomes.
While we generally accept the aspirations of the operators that have stated their intentions it is unlikely that all the plans will be executed or that all the operators will survive in their current form. Reality and externalities will push plans and investment away from the original path.
However the overall outcomes, total coverage, places with multiple networks and those that are challenging and unlikely to get single or multiple gigabit offerings should be reliable.
We have tested the model by running it against historical data. The current version of the model predicts the deployment of FTTP to a postcode by any operator with an accuracy of just over 70%. Individual operators predictions vary from around 15% for smaller and newer operators to well over 70% for the more established players where we have good training data.
The next iterations of the outputs will focus on improving and reporting the results for single operators as we collate more real world input data. As deployments continue we can get a better sense of what factors are influencing their decisions, or not.
Postcode versus higher geographies
While we generate outputs at postcode level that is not necessarily the best way to use forecast outputs. Outputs from models will always be estimates and as we highlight below those outputs have varying probabilities of being correct at any particular point in time. So a general view, aggregated or averaged at a higher geography can make the data easier to interrogate or view in a map or other visual aide without significantly compromising the ‘accuracy’ of the view.
One consequence of models is that we can get unexpected or unrealistic results. The maps above, at LSOA level, highlight one outcome in particular. The areas with 6 or more gigabit networks arise as different operators start to overlap more and more in the outcomes from their stated aims and targets.
This will not happen, at least not to the extent that this model predicts. As time progresses the effect of competition will start to reduce the coverage of particular operators and even threaten their existence. Understanding where this is possible or likely is useful input even if the number of networks is overstated. At the other end of the scale highlighting where there will be much lower and even no fixed gigabit coverage is useful for tracking intervention areas as well as those where more adventurous operators may be ablet to gain good market share.
Methodology
The foundational logic in the forecasting of operator deployment is to score locations based on their attractiveness - that is how profitable it would be for them to deploy in that location. There are a number of inputs that we use to determine whether a location is well suited for operator deployment in general, which make up the General Attractiveness Score (GAS).
There are also inputs which are unique for each operator derived from their stated strategy, budget, current and historical footprint and progress which combined with the GAS provide a mechanism for projecting where we believe each operator will deploy in the coming months and years.
We forecast a select number of operators individually, with the remainder collated into a single ‘other’ operator field. The list for this version is in Appendix A below. As we generate updates to these forecasts those lists will evolve as more detail and timeline emerges on each operator.
After each postcode has an indexed score from the inputs, weighted by their relevance to each specific operator, the forecast flags are generated by ordering the postcodes by their attractiveness to each operator and matching to their stated target numbers and locations where available.
There are also inputs generated specifically for each individual operator:
Catchment Area - Geographic boundaries based on operator announcements. The operators are only forecast to deploy inside the areas which they have announced.
Upper Limit - The number of premises the operators have stated that they will have passed by time t and on to 2030. In many cases this is extrapolated from announcements made by operators for an earlier date.
Once all operator specific attractiveness scores have been generated for all postcodes in their respective catchment areas, postcodes are then listed in order of attractiveness with a running sum of premises generated alongside, and a flag for forecasted presence is allocated to each postcode, until the upper limit for that time step is reached.
This allows us to generate expected deployments for each operator. These results are summed for the outputs at Local Authority level that accompany this paper, which can be accessed through a UK Plus subscription.
[1] Fibre to the premises or building or home – various acronyms generally the same outcome and supply