The UK is in a period of consolidation in terms of networks and ISPs. The explosive growth triggered by regulatory changes combined with COVID and plenty of cash has resulted in more than one hundred full fibre operators around the country.
The heat from the initial deployment dash is starting to dissipate and for some while now the consolidation of these networks is on the rise in the afterglow.
We have put together some of our datasets to provide a guide to some of the key metrics that drive acquisition and valuation and here we present some of the outcomes.
We currently (Feb 2024) track 127 active operators in our ‘availability’ database. Including postcode level presence and technologies. We also collate data on tariffs, speeds, and generate forecasts for expected deployment.
Remember that our data will have estimates, not all ISPs publish their subscriber numbers for example and our coverage data can lag the reality on the ground by several weeks. As a result these outcomes should be treated as a guide but we do have some confidence in the outcomes after comparing with real datapoints, like ARPU, published by many companies.
Some key takeaways:
You are not alone – almost every (at least 98% of networks in our data) gigabit network is overbuilt by another with an average of 11 competitors overbuilding 37% of a network
The average speed delivered (according to speed tests) over gigabit capable networks is 184Mbps.
Adoption is increasing. More than 20% take-up rate today starts to move ahead of the average network but our data indicates that less than 2% of subscribers are taking a top tier option
Average ARPU is £35 a month for gigabit networks
Across the smaller players there is an average of 200,000 potential subscribers to be gained on top of their existing client base while Virgin still has headroom of almost 7M to be won over, against 4.3M across the Openreach/BT gigabit capable footprint.
Footprints and overbuild
Coverage and competition are two of the more straightforward measurements that drop out of the data. In the UK at the end of 2023 if gigabit networks were laid end to end, with no overbuild, the UK is well over 40M households passed by a gigabit capable fixed network (fttp/b and cable). Enough to cover Germany.
However with overbuild the picture is closer to 24M households passed in the UK. Still significant and enough to blanket South Korea, if they had not done it already.
The average gigabit operator in the UK is overbuilt by a gigabit capable competitor in 37% of the footprint. This ranges from 11% for KCOM, with a legacy monopoly still having an impact, to more than 50% for Virgin Media and that is increasing rapidly as competitors eat into the revenue dense footprint that VMO2 has established.
We estimate that an operator will face an average of 14 competitors over their network. BT for example, at the top of the list, currently faces 91 gigabit operators and ISPs across their FTTP deployment and while there are still a few operators who face only one or two competitors they are across small areas.
Gigabit network adoption
Take-up rates in the face of so much competition are another important data point. What percentage of the possible base actually subscribe to an operator?
In the UK there were just over 11 million households that subscribe to a service from a gigabit capable network. Which means a national take-up rate of around 46% across gigabit networks and is 40% of the total broadband subscriptions. The picture varies widely however with a range of over 60% of potential targets adopting across an ISP’s network down to 10% and below in a few instances.
There is another metric that helps to refine our view of adoption, of the customer base who (and how many) are adopting an actual gigabit capable tariff. We turn to the tariffs themselves and speed test data to help us understand these splits.
Tariff tiers and prices
There are not many actual gigabit tariffs (1000Mbps or more downstream) offered in the UK. Many ISPs top out at 900Mbps, although some still label this as ‘gigabit’.
We collect data from ISPs in the UK every quarter. Here is a selection of ISPs and their reported FTTP tariffs and downstream speeds offered from the end of 2023. Some clusters start to emerge.
1. Up to ~210Mbps – The lowest tariff tier with a range of offers and some vertical clusters around 100 and 150 Mbps. Otherwise a slightly chaotic picture reflecting the scrappy and heavily competitive nature of the market at this end. Although in general the prices are lower than (2) there are plenty of mixed offers (more upstream being one) that drive up the subscription price at this level.
2. Greater than 210Mbps and less than 420Mbps. The cable cluster and where Gfast tariffs max out – Although Gfast came to market too late and is no longer being deployed it still has an effect. Originally pitched against the VM max tariffs at the time (~300Mbps) it is still a popular offering for FTTP ISPs to try and attract subscribers.
3. Greater than 420Mbps and less than 800Mbps. Another group around 500Mbps and a noticeable general increase in price. A natural break when a gigabit is the commonest top tier available
4. Over 800Mbps. At 900Mbps another cluster where Openreach based packages dominate. We report here the quoted maximums from the ISPs rather than the actual experience for the customer.
Adoption profiles and tariff groups
Turning to the speed tests, we have data from the Thinkbroadband speed test network. For this analysis we are taking the average at postcode level for each ISP over the last 6 months. Around 7 million tests covering almost 40% of UK postcodes. This helps us understand, using density plots or histograms, what proportion of users are subscribing to which tariff.
Caveats apply. The sample universe is self selecting, a user chooses to do a speed test and we do not use data covering the whole UK. There may also be a ‘wifi effect’ with users testing without plugging into the fixed line directly which can lower the speeds recorded.
For this example we adopt a general split of low, medium and high tiers according to the general picture in the UK. From the chart above we group by clusters 1 for low, 2+3 for medium and 4 for high.
It is possible to modulate these limits as well as split the groupings out further for individual ISPs by examination of their speed test density plots and tariffs reported. These two examples give an indication of the alignment. For this analysis where there are multiple tariffs from an operator in a cluster we take the average. A more refined approach could look in more detail at histograms for example to allocate a subscriber split to each individual tariff.
Community Fibre is one of the few operator/ISPs to offer a multi-gigabit tier and while it isn’t obvious from the density plot above there is evidence of some adoption at this level. It shows up more clearly in the histograms.
Household subscription revenue
Now that we have the numbers for subscribers, prices for tiers and density plots to split the subscribers into those tiers we can multiply these together to get a value for total monthly subscriber revenue estimate for the consumer retail offerings across gigabit capable networks.
Subscription revenue = (price of low tariff number of users on low tariff) + (price of medium tariff number of users on medium tariff) + (price of high tariff * number of users on high tariff) .
This provides an estimate of household subscriber revenue. Given the number of estimations and approximations underlying this output it will not be precise but it does provide insight.
Top of the list is Virgin Media, with the largest gigabit capable network in the UK for the moment. From our models we estimate they generate around £250M a month from their consumer retail base. Only a couple of other networks exceed £10M a month but we believe Hyperoptic are likely to join the club this year with estimated revenues of over £9M a month at the end of 2023.
The larger the network the higher the revenue of course. So this number needs a bit more analysis.
Average Revenue per User - ARPU
ARPU simply expressed and usually calculated as revenue/number of users.
We produce an estimate of subscription revenue which is likely to be on the low side. There are a couple of variables we are not allowing for in this example:
1. No bundles v standalone distinction. We have incomplete evidence on the variation between ISPs in terms of the mix between bundles and standalone, which could be significant.
2. Local promotions or free upgrades. Discounts and in-contract changes and promotional deals often drive early adoption to relatively higher tiers
3. Legacy subscribers. Still on tariffs that have been discontinued for new subscribers. Typically on lower speeds they may or may not be paying less than the equivalent new tariffs.
These outcomes can be compared directly to statements from a number of the ISPs/operators from their publications. Here we show a selection from our full analysis and data tables prepared for our subscribers.
Selected estimated household subscriber ARPU end 2023
ISP/Operator | ARPU |
BrawBand (CityFibre) | £34.70 |
brsk | £37.13 |
Gigaclear | £44.10 |
GNetwork | £27.40 |
Hyperoptic | £32.08 |
KCOM | £39.94 |
Virgin | £46.76 |
Zzoomm | £36.85 |
We can generate the majority of these outputs from our UK datasets for the majority of ISPs and operators. The above is a small sample of what is possible. Our subscribers can access this data and analysis either on request or in our forthcoming site upgrades next quarter.
Adding up the evidence
These outcomes form only part of the value of a network. For subscribers we publish a longer list, see below.
Valuation metric | Source | Description |
Coverage area | Direct from our data | The geographic reach of the network by count of premises (households and businesses) passed by a ready for service deployment |
Subscriber base | Direct from our data | The number of active subscribers or users on the network. |
Take-up rate | Direct from our data | A metric derived from the first two inputs. Subscribers/coverage. |
Average speeds delivered | Direct from our data | The average internet speeds provided to customers, typically measured in Mbps (Megabits per second). This is a crucial factor in customer satisfaction and network competitiveness. Where there are mixed networks the technologies are broken out (DSL, FTTC, Gfast, FTTP)… |
Technology used | Direct from our data | The type of technology used (fiber optic, DSL, cable, satellite, etc.) can affect the network's speed, reliability, and upgrade potential. |
Pricing (high/med/low) | Direct from our data | What are the average prices of the ISPs tariffs in the high, medium and low tiers |
Subscriber tariff tier mix (high/med/low) | Direct from our data | What are the tier splits in the high, medium and low tiers indicated by speed tests |
ARPU | Derived from our data | Generalised tariffs for high, medium and low cost from their offerings multiplied by the number of customers on each tier from speed test density plots |
Subscription revenue | Derived from our data | Generalised tariff for high, medium and low from their offerings multiplied by the number of customers on each tier from speed test density plots and confined to residential consumer subscribers. |
Current overbuild | Direct from our data - current and future overbuild expectations | Understanding the competitive landscape and how the network stands in comparison to its competitors. |
Future overbuild | ||
Competitor count today | ||
Future expansion path | Direct from our data | The network's potential for growth, including plans for geographic expansion. We report the gradient of forecast coverage curve, how difficult for them to meet their targets |
Average population density over network | Direct from our data | Number of premises passed per sqkm |
Headroom | Direct from our data | How much market is left to sell to in the current footprint. |