1.1 Broadband speeds are important to consumers, helping inform their choice of ISP and affecting their experience of the internet. However, there is a lack of robust information available on the actual speeds that consumers receive and how these relate to both the maximum possible line speeds and the advertised ‘up to’ broadband speeds they pay for.
1.2 To help address this issue we have commissioned a two-part survey to provide an improved level of understanding of current broadband performance across the UK.
1.3 It is our intention that this research will provide an evidence base from which to inform Ofcom’s policy making. Publication of the data will also provide information to consumers on how the performance of their broadband connection is likely to be affected by a range of factors, including where they live, the time of day they use their broadband and their choice of ISP.
1.4 We believe that the technical methodology we have selected, combined with our rigorous sampling approach, means that this research represents a step-change from any previous research into broadband speeds across the UK.
1.5 This initial report contains key findings from our survey and offers a high-level analysis of the performance data generated from the first 30 days of data collection (23 October to 22 November 2008). A more detailed report will follow in spring 2009.
1.6 The average speed (or more precisely the actual throughput download speed) received by panel members was 3.6Mbit/s in the 30 days from 23 October 2008. As we have weighted our panel carefully to reflect UK demographics, the market share of the leading ISPs and the distribution of headline speeds, our results offer a true reflection of the average speed experienced by UK consumers during this time.
1.7 This represents 49% of the average ‘headline’ speed (7.2Mbit/s) and 83% of the average maximum line speed (4.3Mbit/s)(-1-). Consumers on the most popular broadband headline speed package (‘up to’ 8Mbit/s) received an average actual throughput speed of 3.6Mbit/s (45% of headline speed), and they had an average maximum line speed of 4.5Mbit/s (56% of headline speed).
1.8 Speeds varied considerably between consumers: one in five people on an ‘up to’ 8Mbit/s package receive an average speed of less than 2Mbit/s.
1.9 Eighty-three per cent of consumers say that, overall, they are satisfied with their broadband service. Over a quarter of consumers claim that the speeds they receive are not what they expected when they signed up to their broadband service.
1.10 Twenty-one per cent of all consumers claimed to have some level of dissatisfaction with speeds, compared to 16% expressing dissatisfaction with value for money and 13% expressing dissatisfaction with the reliability of connection.
1.11 While 93% of users claim to be satisfied with their web browsing experience when using their broadband connection, satisfaction is significantly lower for all other types of internet use; ranging from 77% of those who listen to or download audio, to 60% of those who watch or download full feature films.
1.12 Despite its significance, 28% of consumers are unaware of the headline speed of the package they have purchased. The level of understanding of the factors which influence speed is also modest. For example, many consumers were unaware that the distance they live from the local exchange is a major determinant of speeds for DSL-based services.
1.13 Overall dissatisfaction with broadband is higher among rural users (12%) than among urban users (7%). This may be because they tend to receive slower speeds; rural consumers on ‘up to’ 8Mbit/s packages received average speeds 13% lower than their urban counterparts.
1.14 Much of this difference is likely to be explained by speed degradation among DSL connections caused by the fact that rural customers typically live further from their nearest exchange and therefore have, on average, a longer line length from the exchange to the premises. (Around 80% of UK broadband connections are DSL, whereby broadband is delivered via the copper telephone wire; and a characteristic of DSL broadband, in contrast to cable broadband, is that speeds degrade significantly with the length of the line).
1.15 Speeds vary significantly by time of day, generally becoming slower in the evenings. Across the UK, speeds were slowest between 5pm and 6pm on Sunday, indicating that this is when domestic use of the internet is at its highest. We found that for consumers on ‘up to’ 8Mbit/s packages average throughput speeds at the evening peak evening hours are over 30% slower than average throughput speeds during the off-peak hours of between 4am and 7am.
1.16 At a package level we have statistically robust data for the ‘up to’ 8Mbit/s product for the six operators with more than 5% retail market share in September 2007. We find that the average speed delivered by the slowest of these providers is around 70% of the fastest. This may be in part due to a different geographical profile of customers, resulting in different average line lengths. However, the fact that the worst-performing ISPs had much greater variation by hour of day suggests that degradation caused by contention is also a driver of this difference in performance.
1.17 Download speed is only one of many factors which affect performance. In this report we also analyse upload speeds, latency, packet loss, DNS and jitter; we find a similar pattern of poorer performance during the evening peak time. (See the glossary at the end of this document for definitions of key technical terms.)
1.- The ‘headline’ speed is the download speed at which an internet service is advertised. The ‘maximum’ speed represents the highest speed that the line is capable of, defined in this research by the highest speed ever achieved in the 30 days of data collection. See the Glossary for fuller definitions.
The full document is available below