The Cost of Britain’s Buffering

Key findings

This working note estimates the annual consumer welfare losses caused by peak-hour mobile network congestion. The analysis transfers willingness-to-pay parameters from a discrete choice experiment, calibrates speed degradation using Opensignal and Ofcom data, and aggregates losses across UK mobile data users.

1. London

   £150–260 million per year in speed-only welfare losses at central assumptions, after a 30–50% hypothetical-bias discount. A direct read of the Opensignal time-of-day data would push the upper bound to £265–370 million.

2. UK-wide

   £490–785 million per year, extrapolating the same framework using national operator speeds. The figure relies on an assumed rather than measured rest-of-UK congestion ratio, with a sensitivity range of approximately £478m–£1,010m.

3. Latency

   Latency degradation during congestion plausibly adds losses of similar order. No published time-of-day latency data exists for UK mobile networks, so the latency component is presented as illustrative and awaits empirical calibration.

4. Mechanism

   The losses are attributable to insufficient cell-site density, which is substantially constrained by local planning decisions on mast approvals. Rural welfare losses, though real, are a structurally different problem and are not addressed here.

Abstract

We estimate the annual consumer welfare losses from mobile network congestion in the United Kingdom using willingness-to-pay (WTP) parameters from a discrete choice experiment on internet access preferences. Adapting speed valuations from Czajkowski et al. (2024) and calibrating congestion levels using Opensignal/Ofcom crowdsourced data, we present two estimates differing in their geographic scope and the strength of the underlying calibration.

For London specifically – the only UK city for which Opensignal publishes time-of-day mobile speed data – we find annual welfare losses from speed degradation of approximately £150–260 million per year, after adjusting for hypothetical bias in stated preference estimates. The estimate is conservative relative to a direct read of the Opensignal data: a 46% peak-to-uncongested ratio (as implied by the reported 5pm trough) would push the figure to £265–370 million per year. This estimate rests on directly calibrated, measured congestion data and is the stronger of the two figures presented here.

If we extrapolate the same framework to the UK as a whole, we find annual welfare losses of approximately £490–785 million per year. This figure uses actual national operator speeds from Opensignal (January 2026) but relies on an assumed rather than measured peak-to-uncongested ratio outside London, where no published sub-national time-of-day data exists. Data limitations make this estimate correspondingly more uncertain, with a sensitivity range of approximately £478m–£1,010m depending on assumed congestion severity in the rest of the UK. The mechanism – peak-hour congestion driven by insufficient cell-site density under binding planning constraints – plausibly operates in other large UK cities, but the absence of comparable sub-national data prevents direct calibration. Rural welfare losses, though real, are a structurally different problem and are not addressed here.

Latency degradation during congestion plausibly adds further losses of similar order, but no published time-of-day latency data exists for UK mobile networks; we present illustrative latency calculations that await empirical calibration. Both the London and UK-wide welfare losses are attributable to insufficient cell-site density, which is substantially constrained by local planning decisions on mast approvals.

Model

1.1 Source of WTP estimates

We draw on Czajkowski, Zawadzki, Bernatek & Sobolewski (2024), “Assessing the substitutability of mobile and fixed internet: The impact of 5G services on consumer valuation and price elasticity,” Telecommunications Policy 48(10), 102869 (also circulated as WNE Working Paper 15/2024 (451)). The study estimates a random-parameters logit model in WTP space using data from a discrete choice experiment (DCE) with 5,204 Polish respondents conducted in 2018. Respondents chose between home fixed (HF), home mobile (HM), and mobile (MO) internet alternatives characterised by speed, latency, data transfer limits, and monthly cost.

The WTP-space specification yields coefficients that are directly interpretable as marginal willingness to pay in monetary units. The utility function for alternative j and individual n is:

where α_n > 0 is the individual-specific cost sensitivity (scale) parameter and β_nk are WTP parameters distributed across the population.

1.2 Key parameters

Table 1 reports the estimated WTP-space coefficients from Table 2 of Czajkowski et al. (2024). All coefficients are in units of 100 PLN/month.

1.3 Welfare formulae

For a representative consumer experiencing a speed improvement from s₀ to s₁ (Mbps):

For a latency reduction from ℓ₀ to ℓ₁ (ms):

These are in 2018 Polish prices. Both expressions follow directly from the log-specifications of the utility function.

1.4 Income scaling

We adjust for differences in income between the Polish survey sample and UK residents using GDP per capita at purchasing power parity (World Bank, 2018 values): Poland 2018 GDP/cap PPP = $33,891; UK 2018 GDP/cap PPP = $49,210; London GVA premium ≈ 1.70 (ONS regional accounts). For London, the income ratio is:

For the rest of the UK, the income ratio omits the London GVA premium:

We convert from PLN to GBP via the PPP rate (1.7 PLN per international dollar, 2018) and the UK PPP exchange rate (0.69 GBP per international dollar, 2018):

where η ∈ [0.5, 1.0] is the income elasticity of WTP. Our central case uses η = 0.75, reflecting the fact that mobile data is closer to a near-necessity than a luxury good in 2025 UK contexts (consistent with Glass & Stefanova 2012).

1.5 Population

London population ≈ 9.0 million. UK smartphone penetration is approximately 94–95% (Ofcom Online Nation 2024; Media Use and Attitudes 2024). We assume 90% of smartphone owners use cellular data regularly, giving:

This excludes the ~1 million daily inbound commuters who also use London’s mobile networks. For the rest of the UK, the population of mobile data users is:

London welfare loss estimate

This section presents the London-specific analysis. London is the only UK city for which Opensignal publishes time-of-day mobile speed data, which means the congestion calibration here rests on measured rather than assumed peak-to-uncongested ratios. The extension of this framework to the rest of the UK is presented in §3.

2.1 London mobile network performance – speed

We use operator-level data from Opensignal reports (London-specific where available; January 2025/2026 reports) and UK market share estimates.

With ~20% of London user time on 5G connections (Ofcom Mobile Matters 2025: 29% of urban connections are 5G, but time-on-5G is lower), the blended uncongested speed is:

Congestion

Opensignal time-of-day analysis for London shows 4G speeds ranging from 17.5 Mbps (5pm peak) to 38.3 Mbps (3am off-peak) – a factor of 2.2×. More recent UK-wide data shows operator speeds dropping 15–30% from daily peak during noon–midnight. We model three usage periods.

The usage-weighted actual speed experienced is therefore s̄_actual = 62.6 × 0.765 = 47.9 Mbps. The effective speed multiplier from eliminating congestion is 62.6 / 47.9 = 1.31×.

Calibration of the 65% peak-to-uncongested ratio

The Opensignal time-of-day ratio (17.5 / 38.3 = 0.46) implies a more severe peak-hour penalty than the 65% central assumption. Three considerations support the milder central estimate. First, 17.5 Mbps reflects the single trough hour (5pm) rather than the average across the peak period (12pm–9pm). Second, the 38.3 Mbps off-peak figure is itself congested to some degree – the framework already assumes off-peak speeds are at 95% of uncongested – so the directly-implied peak-to-uncongested ratio is approximately 17.5 / (38.3 / 0.95) = 0.43, not 0.46. Third, the log-WTP specification was estimated on Polish consumers with ~20 Mbps baseline speeds; applying it over a 2.3× ratio requires greater out-of-sample extrapolation than over 1.54× (at 65%). We therefore use 65% as a conservative central estimate. Table 4 reports London welfare losses under alternative peak-ratio assumptions; a direct read of the Opensignal time-of-day data (peak at 46%) would push the London estimate to £265–370 million per year as an upper bound.

2.2 London mobile network performance – latency

Operator-level latency data from Ofcom Mobile Matters 2024 (Opensignal crowdsourced data, October 2023–March 2024):

Latency congestion modelling

We use the same usage-period structure as for speed. Latency is more sensitive to congestion than throughput: when a cell is at capacity, packets queue and latency spikes. We assume:

The counterfactual uncongested latency is 18 ms (blended 4G/5G under uncongested conditions).

2.3 Results: London

Speed welfare gains

Applying the welfare formula with the income adjustment:

Latency welfare gains (illustrative)

Applying the latency formula with usage-weighted latency 27.6 ms and counterfactual 18 ms:

We also compute a “low congestion” scenario with peak-hour latency of 28ms and a proportionally reduced shoulder-period latency of 20ms (off-peak unchanged at 16ms), giving usage-weighted average 0.15 × 16 + 0.35 × 20 + 0.50 × 28 = 23.4 ms and ln(23.4 / 18) = 0.266.

Combined welfare losses (London)

Table 9 presents combined speed and latency welfare losses for London. The speed component is grounded in empirical congestion data; the latency component is illustrative pending primary data collection.

Our preferred estimate for speed losses alone uses η = 0.75 as central, with η ∈ [0.5, 1.0] as sensitivity, and a 30–50% hypothetical bias discount. This yields a headline range of approximately £150–260 million per year as the welfare cost of current peak-hour congestion in London. Even under the most conservative assumptions (η = 0.5, 50% bias discount), speed-only losses exceed £110 million per year. A direct read of the Opensignal time-of-day data supports a higher upper bound of approximately £265–370 million.

The illustrative latency estimates, if approximately correct, would add a further £30–110 million per year. We do not include latency in the headline figure given that the underlying time-of-day latency data is not empirically available.

Counterfactual interpretation

The estimates above are the welfare cost of current peak-hour congestion – i.e., the integral under the WTP curve between the current usage-weighted speed and a fully-uncongested counterfactual. They are the welfare cost of congestion, distinct from the welfare gain from feasible policy: planning reform plausibly closes some share of the gap but does not eliminate peak-hour congestion entirely. The £150–260m figure should therefore be read as an upper bound on planning-addressable speed welfare gains in London, with the realised gain depending on what fraction of capacity expansion is actually achievable under any given reform package.

As a sanity check, Analysys Mason (2012) estimated total UK mobile consumer surplus at £24–28 billion. The London speed-only estimate implies that London peak-hour congestion destroys approximately 0.5–1.0% of national mobile consumer surplus – concentrated in the city with the highest user density and thus the most to gain from capacity expansion.

UK national estimate

This section extends the welfare framework above to the UK as a whole, using national operator speed data from Opensignal. The central UK-wide figure rests on an assumption about rest-of-UK congestion severity that is not directly calibrated against published sub-national time-of-day data. As detailed in §4, readers should treat the figure as indicative rather than as a measured estimate of equivalent rigour to the London-specific analysis in §2.

3.1 UK-wide speed data

We use national operator speed data from Opensignal’s Mobile Network Experience Report (January 2026, based on crowdsourced data from October–December 2025).

The weighted UK national all-day average speed is 42.7 Mbps, substantially below London’s constructed estimate of 62.6 Mbps (see §2.1). We back out the rest-of-UK average speed as a residual:

where ω_L = 0.134 is London’s share of UK mobile data users.

3.2 Rest-of-UK congestion

London congestion severity is grounded in Opensignal’s published time-of-day speed charts for the capital, which show peak-hour speeds at approximately 65% of uncongested levels (see §2.1). For the rest of the UK, no comparable time-of-day data is published at the sub-national level. We assume milder congestion outside London based on the following considerations:

• Tutela (2020) measured Three UK’s peak-hour slowdown at 36% nationally, with other operators showing 15–25% drops.
• Rural and suburban areas – which comprise the majority of UK geography though not necessarily of usage – experience minimal congestion.
• Other large cities (Manchester, Birmingham, Leeds) face moderate congestion but less acute than London’s uniquely dense user concentration.

Our central assumption is that rest-of-UK peak-hour speeds are 80% of uncongested (vs. 65% in London), with correspondingly milder shoulder-period effects. This yields a usage-weighted speed fraction of 0.868 (vs. 0.765 in London).

3.3 UK-wide results

Table 11 presents annual speed welfare losses for London, the rest of the UK, and the UK total.

The preferred UK-wide estimate uses η = 0.75–1.0 with a 30–50% hypothetical bias discount, yielding approximately £490–785 million per year for speed losses across the UK. Of this, London accounts for roughly one-third (£150–260m) despite containing only 13% of UK mobile users, reflecting both worse congestion and higher incomes.

Per-user welfare losses at η = 0.75 with a 30% bias discount are £27/year (£2.25/month) in London and £10/year (£0.80/month) in the rest of the UK. The gap is driven by London’s more severe congestion (peak at 65% vs 80% of uncongested) and the London income premium (1.70×).

Sensitivity to rest-of-UK congestion

Table 12 shows how the UK-wide estimate varies with the assumed rest-of-UK peak congestion severity, holding London fixed. This is the most important source of uncertainty in the UK extension.

Data provenance

The UK extension rests on three tiers of evidence:

• Measured: UK national operator speeds (Opensignal Jan 2026); 5G connection shares (Ofcom Mobile Matters 2025); London operator speeds (Opensignal London-specific reports).
• Derived: rest-of-UK all-day speed (39.6 Mbps, backed out from national average minus London contribution).
• Assumed: rest-of-UK congestion severity (peak at 80% of uncongested). This is the key judgment call; it is informed by Tutela (2020) national peak-hour data and the logic that rural/suburban areas are less congested, but no published sub-national time-of-day data exists to calibrate it directly.

Data limitations on the national estimate

The UK-wide figure presented in §3 should be read with the following limitations clearly in mind. The London estimate in §2 rests on published time-of-day speed data; no equivalent data exists to extend the analysis to the rest of the UK with comparable rigour.

The mechanism documented above – peak-hour speed degradation from insufficient cell-site density, with planning constraints as the binding factor – plausibly operates in other large UK cities (Manchester, Birmingham, Leeds, Glasgow, Liverpool, Sheffield, and other metros) at smaller magnitudes than London. The methodology used to assign a figure to that extension, however, depends critically on an assumed peak-to-uncongested ratio for the rest of the UK rather than a measured one. Opensignal does not publish sub-national time-of-day speed data for any UK city other than London, and the calibration of London congestion rests entirely on the time-of-day series for the capital. Without comparable data for other cities, any extension is a function of an assumed congestion ratio rather than a measured one, and a range of equally defensible assumptions span an order of magnitude in implied welfare loss (see Table 12).

For purely illustrative purposes, a calculation under the assumption that other major UK cities (aggregate population ~22m) experience peak-hour speeds at 75% of uncongested – milder than London’s 65% but more severe than truly rural areas – would imply an additional welfare loss of approximately £190–270 million per year at central specification. This figure is offered only to indicate the order of magnitude that would follow from a particular extrapolation. Primary data collection (timestamped crowdsourced speed measurements in major UK cities outside London) would be required to convert this illustrative figure into a defensible estimate.

The framework deliberately does not extrapolate to rural areas. Rural mobile coverage gaps are real but structurally different: where the binding constraint is the absence of a cell (extensive margin) rather than congestion on an existing one (intensive margin), the DCE methodology calibrated on speed differences does not capture the welfare loss correctly, and the policy questions involved are not the same as those addressed here. Low population density changes both the economics of cell deployment and the appropriate analytical framework, and we leave that question to other work.

Two practical consequences follow. First, the £490–785m UK-wide range presented in §3 should be communicated with the caveat that the rest-of-UK component (£270m to £802m at η = 0.75, 30% bias discount) is an assumed-rather-than-measured number, and the choice of assumption is the dominant source of uncertainty in the figure. Second, where rigour is at a premium, the London-specific estimate in §2 – which rests on published time-of-day data – should be preferred as the headline number, with the UK extension framed as a plausible order-of-magnitude indication of what national congestion losses could be.

Assumptions and caveats

1. Stated preference and hypothetical bias.

   All WTP estimates derive from hypothetical choices, not actual purchase behaviour. The meta-analytic literature finds systematic hypothetical bias in stated preference studies: Murphy et al. (2005) report a median hypothetical-to-actual ratio of 1.35 (i.e., ~26% overstatement) for choice-based mechanisms; Schmidt & Bijmolt (2020, n = 77 studies) find an average bias of 21% for consumer goods, with larger bias for higher-valued products.

   Critically, Haghani et al. (2021) find that while hypothetical bias affects opt-in rates and total WTP levels, marginal rates of substitution between attributes – which is exactly what WTP-space coefficients represent – show substantially less bias. Since the welfare calculation depends on the WTP coefficients from Czajkowski et al.’s WTP-space model (i.e., marginal valuations of speed and latency), the relevant bias may be closer to 20–30% than to 50%. We report results under both a 30% discount (consistent with the DCE meta-evidence centre of gravity) and a 50% discount (conservative upper bound).

   As a revealed-preference cross-check, Nevo, Turner & Williams (2016, Econometrica) estimate ~$2/Mbps/month WTP for download speed from actual ISP billing data under three-part tariffs, in 2007 US prices and at the baseline mean speeds (5–10 Mbps) prevailing in their sample. The Czajkowski coefficient implies a marginal WTP per Mbps at speed s of ∂W/∂s = 15.03/s PLN/month. Evaluated at s = 5 Mbps, PPP-converted to international dollars and scaled to US incomes with η = 0.75, this gives $2.83/Mbps/month in 2018 USD. NTW’s $2/Mbps in 2007 USD, inflated to 2018 prices (CPI factor 1.20), is $2.40/Mbps/month. The two estimates agree to within 18% at the low baseline – a genuine triangulation, given that they come from different methodologies (RP vs. SP), different geographies, and different decades.

   At high baselines, however, the two estimates diverge sharply: Czajkowski’s log specification at s = 47.9 Mbps gives a marginal WTP of approximately $0.30/Mbps/month, an order of magnitude below NTW. Applying NTW’s $2/Mbps linearly to London’s 14.7 Mbps congestion gap would yield welfare losses on the order of £3 billion per year – an order of magnitude larger than the DCE-based estimate. This discrepancy is informative rather than damaging: NTW estimate marginal WTP at the low speeds prevailing in their 2007 sample, where the WTP curve is locally steep; linear extrapolation to a 50+ Mbps baseline ignores the diminishing returns that any plausible utility specification embeds. The log specification reproduces NTW’s estimate at NTW’s baseline and dampens appropriately at higher speeds. We treat NTW as confirming both (a) the Czajkowski coefficient at low speeds and (b) the necessity of diminishing-returns structure at high speeds, rather than as a directly competing estimate.

2. Preference transferability.

   The DCE was conducted on Polish consumers in 2018. London consumers face different outside options, have different usage patterns (more reliance on navigation, ride-hailing, mobile payments), and different baseline expectations. Income scaling adjusts for level differences but not structural differences in demand.

3. Latency congestion calibration.

   Opensignal publishes time-of-day speed data but not time-of-day latency data for London or any UK city. The peak-hour latency assumptions (35ms central, 28ms conservative) are engineering conjectures, not empirical measurements. No publicly available dataset reports diurnal mobile latency variation in a UK urban context. The speed congestion ratios (peak at 65% of uncongested) are better grounded in Opensignal’s published London data. The latency welfare estimates should accordingly be treated as illustrative.

4. Logarithmic specification.

   The log-log functional form captures diminishing marginal returns to speed and latency improvements. This is appropriate: the value of going from 10 to 20 Mbps exceeds the value of going from 100 to 110 Mbps. However, the extrapolation from the Polish sample’s median mobile speed (~20 Mbps in 2018) to London’s current speeds (~48 Mbps) requires that the estimated curvature is stable out of sample.

5. Income elasticity.

   We assume WTP scales with income raised to η. Setting η = 1 assumes proportional scaling; η = 0.5 assumes WTP is concave in income. This follows standard benefit-transfer methodology (Brouwer & Bateman, 2005; Bateman et al., 2004), where η ∈ [0.3, 1.0] is typical for health and environmental goods. For telecom specifically, no consensus exists. Rabbani et al. (2024) report large WTP discrepancies by income level (higher-income respondents value speed substantially more), consistent with η > 0 but without estimating the elasticity directly. Glass & Stefanova (2012) find that broadband demand has become increasingly inelastic, with small income elasticities, suggesting broadband access is now a near-necessity – which may imply η < 1 for quality improvements on the intensive margin.

6. Omitted dimensions.
   • Reliability: Network connection failures and timeouts are not captured by speed or latency averages. Boyce (2024) estimates WTP for broadband reliability at $10–38/month, but in a different context (rural US fixed broadband). We do not attempt to quantify reliability losses.
   • Business users and commuters: The ~1 million daily commuters into central London and business/enterprise mobile users are excluded. Their WTP may be higher than residential averages.
   • Data limits: The model includes a separate WTP coefficient for data transfer limits (β = 6.23 PLN per log-unit), which may be relevant if congestion also affects effective data consumption. We do not model this channel.
7. Second-order effects.

   We assume that behavioural responses to speed – where and when people use mobile data – are approximately zero. In reality, congestion may cause substitution toward WiFi or avoidance of data-intensive activities during peak hours, which would mean the welfare loss is partially internalised as reduced usage rather than degraded quality.

8. No UK-specific WTP research.

   Frontier Economics / LS telcom (2022), in a report commissioned by DCMS, stated explicitly that since the 2012 study there has been no new willingness-to-pay research, so consumer surplus cannot be estimated reliably. The 2012 study was Analysys Mason’s spectrum valuation, which estimated total mobile consumer surplus at £24–28 billion per year. Our speed-only congestion-loss estimate of £150–260 million for London represents approximately 0.5–1.0% of that total – a plausible magnitude for the cost of peak-hour speed degradation in one city. This estimate fills a gap that the UK government’s own commissioned research has identified, but should not substitute for primary UK research.

Policy implications

The welfare losses documented here arise from peak-hour congestion: an estimated £150–260 million per year in London at central assumptions (with the upper end at £265–370 million under a more direct read of the Opensignal time-of-day data), and approximately £490–785 million per year UK-wide under the central assumption that rest-of-UK peak speeds are 80% of uncongested. The UK figure rests on an assumed rather than measured congestion ratio outside London and should be communicated with that caveat.

The engineering solution is well understood – cell-site densification through small cells and additional macro sites – and the binding constraint is planning. Local authority approval rates for mobile mast installations vary substantially across London boroughs and councils nationwide, with conservation area designations, amenity objections, and prior approval processes creating significant delays and refusals. Even at the lower end of the headline range and after a 50% hypothetical bias discount, the welfare case for streamlining the approvals process for mobile infrastructure – particularly small cells deployed on existing street furniture – is strong on consumer-welfare grounds alone. Latency and reliability improvements, once empirically quantified, would add substantially to this figure.

These figures should also be read as upper bounds on the welfare gain from any given reform package: planning liberalisation can close some share of the congestion gap but is unlikely to eliminate peak-hour congestion entirely.

The mechanism plausibly operates in other major UK cities at smaller magnitudes than London, and the §3 figures are calibrated on that basis. Where rigour is paramount, the London-specific estimate in §2 should be preferred as the headline number. We deliberately do not extend the welfare argument to rural areas, where the binding constraint is coverage rather than capacity and the policy questions involved are different from those addressed here.

References

1. Analysys Mason (2012). Study on the value of spectrum to the UK mobile market. Report for Ofcom.
2. Bateman, I.J. et al. (2004). “Economic valuation with stated preference techniques.” Journal of Health Economics.
3. Boyce, B. (2024). “WTP for broadband reliability.” Telecommunications Policy 48(10).
4. Brouwer, R. & Bateman, I.J. (2005). “Benefits transfer of WTP estimates.” Environmental and Resource Economics 29(2).
5. Czajkowski, M., Zawadzki, W., Bernatek, G. & Sobolewski, M. (2024). “Assessing the substitutability of mobile and fixed internet: The impact of 5G services on consumer valuation and price elasticity.” Telecommunications Policy 48(10), 102869. DOI: 10.1016/j.telpol.2024.102869.
6. Frontier Economics / LS telcom (2022). Estimating the value to the UK of mobile connectivity. Report for DCMS/DSIT.
7. Glass, V. & Stefanova, S. (2012). “An empirical study of broadband diffusion.” Info 14(4).
8. Haghani, M., Bliemer, M., Rose, J., Oppewal, H. & Lancsar, E. (2021). “Hypothetical bias in stated choice experiments.” Journal of Choice Modelling 41.
9. Liu, Y.-H., Prince, J. & Wallsten, S. (2018). “Distinguishing bandwidth and latency in households’ WTP for broadband.” Information Economics and Policy 45: 1–15.
10. Murphy, J.J., Allen, P.G., Stevens, T.H. & Weatherhead, D. (2005). “A meta-analysis of hypothetical bias in stated preference valuation.” Environmental and Resource Economics 30(3): 313–325.
11. Nevo, A., Turner, J.L. & Williams, J.W. (2016). “Usage-based pricing and demand for residential broadband.” Econometrica 84(2): 411–443.
12. Ofcom (2024, 2025). Mobile Matters reports. Opensignal crowdsourced data.
13. ONS. Regional gross value added (balanced), London vs UK average.
14. Opensignal (2025, 2026). UK Mobile Network Experience Reports.
15. Rabbani, M.G. et al. (2024). “Consumer WTP for broadband attributes.” Telematics and Informatics 93: 102173.
16. Schmidt, J. & Bijmolt, T.H.A. (2020). “Accurately measuring WTP for consumer goods.” Journal of the Academy of Marketing Science 48(3).
17. SIMSherpa (2025). London operator speed data compilation.
18. Tutela (2020). “UK mobile network congestion analysis.” Peak-hour speed measurements by operator.
19. World Bank. GDP per capita, PPP (current international $), 2018.