- Technology, Media & Telecom
- Pricing Analysis
- Data Science and Quantitative Modeling
- Europe and Central Asia
- Burkina Faso
- Congo, Dem. Rep.
- Ivory Coast
- South Africa
- Middle East
- Saudi Arabia
- West Bank and Gaza
- North and Latin America
- Asia- Pacific
- Hong Kong
- New Zealand
- South Korea
- Sri Lanka
Technology, Media & Telecom provides operators’ Key Performance Indicator (KPI) across 106 countries on a quarterly basis and covers operators’ performance in the following areas:
Our database includes operator metrics dating back to 2000.
COVERAGE (106 countries)
Quantum-Web’s methodology to provide statistics and forecasts are defined by the following principles:
The primary source of our statistics and indicators are quarterly financial and operational reports of operators, vendors and officially-recognized international sources such as The World bank, IMF, Eurostat, OECD and national telecom regulatory bodies. This approach presents the most current and accurate national and international data available by cross-validation of our statistics and the used for our forecasts.
Quantum-Web uses a variety of Econometrics techniques and quantitative methods to estimate market supply and demand of telecoms industry, services and applications.
Telecommunication forecasting supply and demand is crucial for planning, operating, and developing telecommunication networks and enables telecoms and media companies to rapidly understand the key trends in the sector and anticipate and capitalise on market shifts.
Principles rules to develop our telecom forecasting are:
Keep the econometric model simple.
Use all the available data, e.g. case of fixed line forecast,using socio-economic and wireless data.
Use theory and not the data as a guide to selecting causal variables.
Consider the demand function for fixed broadband. On the basis of economic theory, we postulate that the demand for fixed broadband is a function of number of households and businesses, GDP per head , fixed network coverage, Investment in fixed network (as a measure of the supply side), and other relevant variables (e.g., urban population, wireless broadband penetration). From time series data, we estimate an appropriate model of fixed broadband demand (Linear or Non-Linear model), which can be used for forecasting demand for fixed broadband in the future. Of course, forecasting errors increase rapidly if we go too far out in the future.
To produce telecoms market forecasts we use two types of models:
- Linear Models
- Non-Linear Models
Single or Multiple Linear Regression Models
Regression models (Single-equation, Simultaneous-equation regression models) which are used when two or more variables are thought to be systematically connected by a linear relationship. Typical example of Linear Regression models deployment in our forecasting process is the estimation of “Wireless Traffic” which systematically and directly is in relation with number of wireless broadband connection.
The emphasis of these methods is not on constructing single-equation or simultaneous-equation models but on analysing the probabilistic, or stochastic, properties of economic time series on their own under the philosophy let the data speak for themselves. Unlike the regression models, in which Yt is explained by k regressor X1, X2, X3, . . . , Xk, the ARIMA models allow Yt to be explained by the past, or lagged, values of Y itself and stochastic error terms. For this reason, ARIMA models are sometimes called atheoretical models because they are not derived from any economic theory and economic theories are often the basis of simultaneous-equation models.
A typical example of ARIMA model deployment in our forecasting process is the estimation of “National Churn Rate” which is not related to a specific and measurable exogenous variable to explain directly the variation of Churn Rate at the national level.
VAR methodology superficially resembles simultaneous-equation models in that we consider several endogenous variables together. But each endogenous variable is explained by its lagged, or past, values and the lagged values of all other endogenous variables in the model; usually, there are no exogenous variables in the model.
A typical example of VAR Models deployment in our forecasting process is the estimation of “Roaming Revenue”. Roaming Revenue is a very important for Telecom pricing managers’ decision making. This example illustrates roaming revenue forecast using the VAR model.
Roaming demand can be measured in terms of:
number of leisure and business visits from an origin country to a destination country
The average amount of data usage per day, number and duration of call per day by visitors from the origin country in the destination country.
Duration of travel by roamer in the destination country.
Destination country, where the roamer travel to and the roaming rates applied by roamer operator.
Holt-Winters’ method is suitable for time series which contain both trend and seasonal variations. The method gives larger weights to more recent observations, and the weights decrease exponentially as the observations become more distant.
A typical example of Holt-Winters models deployment in our forecasting process is the estimation of short terms (typically 8 quarters ahead) “Mobile Subscription” and “Number of Fixed Broadband Connections”.
Growth studies in many branches of Telecoms have demonstrated that more complex nonlinear functions are justified and required if the range of the independent variable encompasses early adoption, growth, maturity and decline stages of a product or service. Then a function with a sigmoid form, ideally its origin at (0,0), a point of inflection occurring early in the growth stage and either approaching a maximum value, an asymptote, or peaking and falling in the decline stage, is justified. Examples of such models include the Logistic, the Gompertz, and the Bass-Diffusion.
A typical example of Non-Linear models in our forecasting process is the estimation of “Telecom Traffic”. We calculate how the average monthly fixed broadband traffic is related to the stand-alone, Double, Triple and Quad play broadband packages. The non-linear models often accurately describe how the average monthly fixed broadband traffic is related to the diffusion of stand-alone and bundled packages among households in a given country.
To see a sample of our data please contact Andrea Riddling on (email@example.com), +44 (0)20 3286 9570. For a sample