Slide1: Future Mobile Networks: Business Models Presentation held at 1st Caspian and Black Sea Regulatory Conference, Istanbul 25 - 27 May 2006 Erik Bohlin, Chalmers University of Technology Contact: Erik Bohlin (erbo@mot.chalmers.se) Contribution & proposal builds upon results developed in the EC/DG JRC/IPTS Project Future Mobile Services (see http://fms.jrc.es) The usual disclaimer apply - this contribution is of the authors, and does not necessarily reflect the view of the European Commission.


Aim & Contribution: Aim & Contribution The aim is to provide a financial simulation of future mobile services Provide an interrelated set of technology, geography, costing and demand levels, based on scenarios and sensitivity analyses Contributes with an early financial analysis of 4G networks Outline of presentation Trends in mobile communications Overview of model Main assumptions and results (incl. sensitivity analysis) Conclusions Appendix showing more detailed assumptions


Slide3: Source: ITU World Telecommunication Development Report, 2002; ITU World Telecommunication Indicators Database and ITU projections ITU Fixed line estimate 2002


Slide4: Global Mobile user population growth 6 5 4 3 2 1 0 Billions of mobile services users worldwide 1985 1990 1995 2000 2005 2010 2015 2020 2025 2030 NB Confidence in estimates Reduces with time 1.5Bn users, 2004 800K users, 2000 With globalisation, OECD cost levels of services and handsets will slowly be rebalanced, being set by average world prices Saturation is set by affordable price for the majority of users – numbers which may reach 65% of a global population of over 8.2Bn by 2030 (US Census Bureau, 2005) if world usage goes towards average OECD levels of saturation today: 5.3Bn users


Slide5: 4 Generations of Mobile Radio are appearing


The basis for the business modelling: What is 4G?: The basis for the business modelling: What is 4G? Important characteristics used in our models: Licensed AND unlicensed spectrum Increased data usage Multi technology usage Decreased cell sizes Introduction of ad-hoc (mesh) network capabilities Leading to reduced density of radio equipment Increased bit rates Increased importance of software (e.g. SDR) New pricing schemes In our models we have used two radio access points: APs and UAPs UAPs are defined to function as ordinary base stations with backhaul connections APs are defined to function as repeaters and signal amplifiers, relaying radio signals to allow extensions of hops.


Model characteristics – 4G network overview: Model characteristics – 4G network overview All 4G network elements have mesh capabilities Multihop paths between handsets, APs and UAPs, and handsets have relay functionality (and with SDR) Redundant paths are often available BN


Model characteristics - overview: Model characteristics - overview The simulated 4G network is divided into a number of corresponding ”network units” Urban, suburban, and rural The ”network units” have distinctively different characteristics E.g. Cell size, average distance between users, number of users in each network unit, etc. Each network unit is composed of a number of base stations, UAPs, etc., using various technologies The characteristics of each ”network unit” are defined regarding e.g. number of users coverage, average distance between users, number of users in each network unit, etc. leading to different investment costs By spreading the network units like a puzzle over the simulated geographical area, the investment costs can be approximated e.g. 10 rural units, 16 suburban units, and 24 rural units to cover a certain area


Network unit – overview: Network unit – overview The network units differ in several ways, e.g.: Size Population density Technology choices Cost structures Urban Suburban Rural Different wireless technologies used


Model characteristics - overview: Model characteristics - overview The model used is schematically presented to the right By using researched estimates for input data reasonable cost structures can be approximated The aim is to evaluate 4G business models by calculating ARPU levels needed to cover OPEX and CAPEX for the different socio/economic scenarios developed in the project Technology limitations/Capabilities Population density CAPEX data Base station costs Handset subsidies Equipment replacement Installation Etc. Population data Geography data OPEX data Marketing Backhaul Site Rent Etc. Factors from scenarios developed in previous Work Packages


Network units – assumptions on population / km2: Network units – assumptions on population / km2 A geographical area is divided into three subareas, differing regarding e.g. technology needs, population density, service usage, etc.: Urban areas Suburban areas Rural areas Population density data for real countries have been used to construct the network units Singapore and Hong Kong have been used to model Urban areas The Netherlands have been used to model Suburban areas Sparsely populated countries as Sweden and Estonia have been used to model Rural areas Population Land area (km2) Population/ km2 Urban 6 000 Singapore 4 353 893 683 6 377 Hong Kong 6 855 125 1 042 6 579 Suburban 500 The Netherlands 16 318 199 33 883 482 Rural 30 Estonia 1 341 664 43 211 31 Sweden 8 986 400 410 934 22


Network unit – assumptions on roll-out : Network unit – assumptions on roll-out The network units are constructed with the following characteristics: (note: rounding errors appear in table)


Eurolandia – assumptions and network units: Eurolandia – assumptions and network units A fictive European country, Eurolandia, has been constructed for the financial simulations of a 4G network Population: 46 400 000 Population figure is the average of France, Germany, Italy, Netherlands, Spain, Sweden & UK populations 50 % are modelled to live in urban areas 35 % are modelled to live in suburban areas 15 % are modelled to live in rural areas Network Units needed to cover Eurolandia population:


Assumptions - CAPEX: Assumptions - CAPEX APs & UAPs Through usage of mesh networking capabilities fewer APs and UAPs will be needed, leading to lower equipment costs per subscriber 10% replacement investments annually for UAPs and APs are estimated Installation costs are not yet included in the model SDR Handsets Additional cost for a SDR Handset with mesh networking capabilities is estimated to maximum 160€ The 160€/subscriber are considered operator investment costs in the model The additional handset cost will most likely have to be subsidized by the operator in order to acquire subscribers With increased intelligence in the handsets, handsets adjust depending on which technology the base station it temporarily communicates with uses


Assumptions - CAPEX: Assumptions - CAPEX Equipment investment costs are expected to decrease rapidly Starting cost estimates (year 0) for network equipment: UAP = 12 000€/unit AP = 10 000€/unit SDR = 160€/handset An additional 100€ per subscriber in acquisition cost, same during whole period Cost reductions during period according to figures to the right Network Coverage 36% coverage in year 0 64% in year 1 83% coverage in year 2 90% coverage in year 3


Assumptions - OPEX: Assumptions - OPEX Data Backhaul Data backhaul costs are estimated to be 700€ per site per month Only UAPs are assumed to have backhaul connections, APs are considered to be repeaters Site Rent Annual rental cost for AP and UAP sites are assumed to be 3000€ in Urban Areas (or 250€/month) 1500€ in Suburban Areas 1000€ in Rural Areas Rental costs include electricity. Urban area cost figures are based on research made by Björkdahl & Bohlin (2004) Maintenance Each maintenance personnel is estimated to handle service of 50 APs/UAPs Each maintenance personnel is estimated to cost 100 000€ annually (including material, salary, vehicle, etc.) Maintenance cost figures are based on research made by Björkdahl & Bohlin (2004)


Assumptions - OPEX: Assumptions - OPEX Marketing Costs Marketing costs are assumed to be ~1.9 € per inhabitant in country/operator/year The figure is based on advertising figures for telecom operators in Sweden in 2003 & 2004 (not including customer acquisition costs!) Administrative Costs Administration costs are assumed to be 10% of other costs The low-cost telecom company Tele2 is used as a benchmark. Tele2 had administrative expenses equivalent to 10% of other costs in 2003 and 9% in 2002 (according to annual income statements)


Slide19: Favour the take up of services Unfavourable Favourable Inhibit take the up of services Economic conditions Social/ cultural/ political conditions Scenario assumptions - economic and sociological conditions Scenario 1: Smooth Development Scenario 2: Economic stagnation Scenario 3: Constant Change


Assumptions - CAPEX: Assumptions - CAPEX The network is built in 4 years in all scenarios (year 0-3) All UAPs and APs installed first 4 years (covering 90% of the population) Investments in mesh capable handsets are made in line with customer diffusion each year Replacement investments (10% annually) are made each year Differences in maximum diffusion levels 90% in Scenario 1 50% in scenario 2 75% in scenario 3 Differences in starting years Scenario 1 starts 2010 Scenario 2 starts 2015 Scenario 3 starts 2012 For easy comparison, scenarios have been set to ”base line”, Year 0-Year 11


Assumptions - scenarios impact on subscriber diffusion: Assumptions - scenarios impact on subscriber diffusion Diffusion more rapid in urban areas Slowest in rural areas Overall diffusion reaches 90%, 50% & 75% respectively in year 10-11 Subscriber diffusion – Scenario 2 Subscriber diffusion – Scenario 1


Results - CAPEX development for Scenarios 1,2 & 3: Results - CAPEX development for Scenarios 1,2 & 3 CAPEX development – Scenario 1 CAPEX development – Scenario 2 CAPEX development – Scenario 3 Initial network investments in year 0-3 Highest number of users added in years 5-7 Replacement costs significant late in period


Results - OPEX development for scenarios 1,2 & 3: Results - OPEX development for scenarios 1,2 & 3 Total monthly OPEX/subscriber curve bathtub-shaped Initial OPEX/subscriber very high in scenario 2 due to low number of subscribers in first years Total yearly OPEX/subscriber Follows same pattern as monthly OPEX/subscriber Total yearly OPEX per subscriber Total monthly OPEX per subscriber


Results - CAPEX & OPEX comparisons : Results - CAPEX & OPEX comparisons Accumulated CAPEX in each scenario Scenario figures similar first 4 years (build-out phase) Scenario 1 highest (due to high number of subs.) Accumulated OPEX in each scenario Similar to CAPEX developments Accumulated CAPEX Accumulated OPEX


Results and assumptions for ARPU calculations: Results and assumptions for ARPU calculations Interest rate used to discount figures: 12% Interest rates are currently very low (Euroswap, STIBOR, etc. 2-3%) A risk premium of 9-10% has been added Both CAPEX and OPEX have been discounted to NPV NPV has been annualized to calculate ARPU levels needed Monthly ARPU needed to cover CAPEX+OPEX 15.8€ in Scenario 1 18.8€ in Scenario 2 16.6€ in Scenario 3 Monthly ARPU needed


Sensitivity analysis – Different cost development scenarios: Sensitivity analysis – Different cost development scenarios Three different price scenarios have been simulated Base scenario: High starting cost, low cost reduction Sensitivity check scenarios: Low starting cost, very slow cost reduction Medium starting costs, slow cost reduction UAPs SDR APs APs – cost development curves SDR – cost development curves UAPs – cost development curves


Sensitivity analysis –Changed cost curves and license cost: Sensitivity analysis –Changed cost curves and license cost Changed cost development patterns has small impact on monthly ARPU levels needed Maximum of 1.1€ difference (Scenario 2) If considering both diffusion patterns and cost development patterns impact increases Maximum of 3.5€ difference between lowest and highest ARPU needed Adding a license fee has strong effects on ARPU levels needed Impact depends on diffusion. strongest impact in scenario 2 A license fee of 652€ per capita (3G license cost in the UK) increases the ARPU needed by 130% in scenario 1 200% in scenario 2 150% in scenario 3 ARPU needed with added license fee* ARPU needed with different cost curves *Based on High starting cost, rapid cost reduction dor components


Conclusions: Conclusions Our simulation indicates minimum ARPU levels of 15-19 € monthly needed for 4G business cases Depending on diffusion levels and equipment cost developments Excluding spectrum license costs Impact on 4G business cases indicated by sensitivity analyses: The cost of spectrum licenses High The 4G diffusion levels Medium Equipment cost development curves Low Implications is that spectrum management methods may be far more critical than market uptake for profitability


Slide30: Scenarios – assumptions on economic development EU Economic output Mean Disposable income 2005 -2020


Slide31: Smooth development – EU economy provides growth and positive progress in development, but in a fair and managed way that brings prosperity across all 25 members and any new accessions EU Economic output Mean Disposable income 2005 -2020


Slide32: 2 Stagnation – economic stagnation with progressive deflation as in Japan from late 1980s, continuing to 2004, with enormous overhang of loans on assets, especially in private housing 2020 2015 2010 2005 Scenario


Slide33: 3 Change – Constant changes, up and down with ad hoc growth and recession, often in parallel in different geographic markets, but moderately positive overall –with geographic and sector pockets of stability, and some of failure, all randomly distributed