We have updated our forecasts, including new methodologies for postcode level in the UK.
They fall into two categories. Premises passed, how many home plus business premises a deployer/operator has ready and lit for service; and Operator/ISP footprints which track where we expect each of them to deploy to next by postcode.
This is an update to our forecasts from December 2022 - https://www.point-topic.com/post/uk-fixed-gigabit-broadband-availability-2022 - these will now be updated quarterly. If you are currently a subscriber contact us for details and deliverables.
Full data outputs are available to Thinkpoint subscribers. See the product page for more details.
Operator Footprints – postcode predictions
This category of forecasts uses several variables to arrive at a score that rates each postcode according to how ‘attractive’ it is to a particular operator.
In order to understand where and when an operator will deploy we use the following:
This builds an ‘attractiveness’ score for each UK postcode for each operator in the analysis. The last two elements make this score unique for each operator.
Each operator is also assigned a catchment area. The pool of potential areas available to an operator is constrained by their public statements which offer a view of where an operator is planning to build. Openreach for example publishes a long-term plan of where (at exchange level) and when they will deploy FTTP. The same is true to a more limited extent for most of the operators in the model.
The model then allocates the requisite number of postcodes/premises to the operator so that they move along their deployment curves as required to meet their targets.
We also use external sources, for example the roadworks databases, to check if an operator has made an application or plans to dig in an area. This can be useful for guiding expectations in the shorter term.
The model is recursive. Each period the calculations are re-run based on the interacting behaviour of the operators. This influences Overbuild Sensitivity in particular. If the model places one or more operators in a postcode at time t+1 that then influences the deployment decision of the other operators depending on how much attention they pay to competition and overbuild in their footprint.
The model output provides an individual operator forecast and can be aggregated to provide a compound forecast that counts the number of operators present in a postcode at time t, in 6 month increments to 2030.
Individual operator footprints
This approach allows us to publish expected postcodes for each operator. Subject to the limitations of our input data we can review each operator over time.
See below animation:
We can also review how they interact and where there will be less or more competition.
This version of the model forecasts that Gigaclear will face the following competition at the end of 2026, by postcode count.
Similar outputs can be generated for each operator as well as sub setting the lists to review which combination of Operators will be where. For example Gigaclear will face Zzoomm in 537 postcodes by the end of 2026.
Premises passed
Overall, the UK is progressing rapidly with deployments. Accelerated in the last few years by regulatory changes, where infrastructure competition is now encouraged and BT and Openreach were structurally separated as well as a (pandemic enhanced) appetite for more and ‘better’ broadband.
More suppliers with plenty of outside investment have added to the historic core of Openreach and Virgin and have driven gigabit capable broadband services further and faster than might have been possible otherwise.
With (as of writing) almost 60% of the UK with at least one FTTP option and almost 80% with gigabit access available the frantic last few years have meant significant changes to the UKs coverage with more to come.
The pace of that change and how it relates to the targets that the companies and indeed government have announced is where we produce one set of our forecasts.
Compound outputs
One view is to sum all the operators present in a postcode to give a general measure of what is expected locally.
Previous models naturally predict multiple operators present in a postcode. There were instances with postcodes with up to 6 operators. This latest version has postcodes with up to 10 operators present in a postcode.
While this may seem unrealistic and is not sustainable in reality, it is a consequence of the number of companies deploying in the UK and the updates to the models reflecting early (relative) avoidance of overbuild.
Overall, the UK is heading for 30 million plus premises with one or more gigabit suppliers before the end of the decade. Real pressure on operators will kick in from 2027 as we see more and more instances where there are more than 4 operators projected in a postcode.
There are three primary reasons:
We are seeing less overbuild early in the series as operators avoid each other, establishing their footprints where they can build a customer base with less competition. However with the number of operators present later in the series (we do not allow consolidation in the models) that overbuild returns and increases. This comes from real world data where operators are proving more sensitive to the presence of others than previously expected.
Our model fitting growth curves, particularly for Openreach, indicate faster deployment with a longer tail of slower growth. This is a consequence of improving the fit of our curves on roll back testing.
There are higher gross targets from the operators than before. If you sum the totals to end 2030, including Openreach, there will be almost 80 million total premises passed not accounting for overbuild. Meaning an average of 2.5 operators in each UK postcode.
Single operator statistics and forecasts can be harder to come by and to generate. Point Topic’s time series of who, where and when in terms of broadband deployments is at the core of our output. Tracking the changes and checking them against expected and projected growth allows us to produce regression-based analysis that outlines the challenge that remains for each.
Currently, this forecast project focuses on 15 individual broadband operators:
These forecasts will now be released quarterly and will regularly update and add to the list above. Nexfibre for example was too late to make it into this round but a view will be generated (even with a short time series and less populated validation set) in early 2024. Netomnia is another large operator that will be added.
Looking at Openreach where we have complete data lets us look at the mature end of the market.
They are experienced and fully established. As you would expect they largely live up to their promises. They have been fairly consistent and perhaps more importantly have a relatively easy path to meet their targets.
The average gradient of the curve that remains between the most recent data used in this set of forecasts (June 2023) and their stated end points is the best indication of how difficult it will be to hit those numbers.
In Openreach’s case even though they have yet to pass the point of inflexion (the steepest part of the curve where the most rapid deployment phase is expected to be) the model of their deployment leaves enough room to slow down towards the end of the forecasting period (2030). This is likely to be the case in reality as they start to encounter more challenging deployment conditions in the harder to reach areas that are slated for the end of the decade.
Most operators however provide targets in shorter time horizons. The model assumes that they continue to exist and expand after this date up to the end of 2030.
To display this we have taken the stated targets (as of June 2023) and using our data on premises passed have calculated the percentage change (from existing RFS base) per half year required to achieve that target.
The current expectations for selected operators, ordered by the relative remaining growth rates according to our data as at June 2023.
These will be updated in Q1 2024 which will include an update on targets as operators re-state their expectations.
This is an update to our forecasts from December 2022 - these will now be updated quarterly. If you are currently a subscriber contact us for details and deliverables.
Full data outputs are available to Thinkpoint subscribers. See the product page for more details.