The Canback Global Income Distribution Database (C-GIDD) is the only database with comprehensive and detailed GDP and household income data for all 211 countries in the world. Data are available not only at the national level, but also by subdivisions (states, provinces, etc.), in turn divided by individual cities, other urban and rural entities. In total, we cover 693 countries/subdivisions and 1,022 cities from 1997 till 2017.

C-GIDD was introduced as a commercial, web-based service in 2008 and is now on its seventh update cycle.

The real benefit of the database is that it quantifies how many households are at a certain income level. That is, how many households are in a given income bracket or socioeconomic level. And it does this with a uniform global standard based on purchasing power parity, so that, e.g., Mumbai can be compared to Shanghai and Rio de Janeiro. Click here for a description of C-GIDD.

Moreover, we have complemented the income data with social, demographic and psychographic data so that we have a comprehensive picture of the world’s consumers. There is also a benchmark products module consisting of 20 products and services—such as cellular phone penetration or insurance sales— that can be used for comparisons.

It took 15 years to develop C-GIDD’s logical foundation, create the underlying mathematical algorithms and to populate it with data. To illustrate its uniqueness and utility, consider the following:

The first task in analyzing a market opportunity is to understand how large the available market is. In emerging markets, the available market is largely defined by how many people can afford the product or service. If people cannot afford the product, metrics such as high purchase intent are meaningless.

C-GIDD allows us to quickly quantify the size of the available market. The database covers 211 countries and territories at the provincial and city levels (all the world’s cities with more than 500,000 inhabitants). For each of these units, we have time series data on how many households belong to a certain income bracket (in purchasing power parity or market exchange rate terms).

C-GIDD also contains data on how many households exist within a certain socioeconomic level, using the AMAI socioeconomic level classification (ABCDE levels) as a global benchmark. We like the scheme the AMAI uses in Mexico because it is rigorous and climate/culture independent. It also corresponds well to SEL schemes in other countries.

The second task in analyzing a market opportunity is to understand how large the addressable market is. Among those consumers that are available, how many are addressable in the sense of having an interest—latent or overt—in the product? We have expanded C-GIDD to help answer this question as well. C-GIDD includes 160 social, psychographic and demographic indicators collected from a variety of sources.

This, for example, allows us to quantify how many women live in Guangzhou who have at least a bachelor’s degree, are between ages 25 and 35, are married and have children, view themselves as adventurous and like to try new things, prefer to shop in hypermarkets, and prefer magazine articles to TV commercials as their source of product information.

These market-level data from C-GIDD can be merged with consumer survey data to predict where a category or brand is heading, and what the fundamental purchase intent drivers are.

The third task when predicting markets is to understand the actual market. That is, what are people buying today and why? C-GIDD is again helpful because select data from it can be merged with category data from, for example, ACNielsen, IRI or Euromonitor or other sources. The merged database can then be used to calculate demand drivers and price, income or marketing elasticities.

In sum, C-GIDD is central to how we look at markets, consumers and marketing. It allows us to quickly analyze markets within and between countries. Most importantly, it gives us an intellectual foundation and analytical rigor that allow our clients to make true comparisons between categories and countries today and into the future.

The article "Where in the World is the Market?" describes how C-GIDD can be used in consumer marketing.

We complement C-GIDD with proprietary databases:

• The retailer database maps the individual location and sales/size of hypermarkets, supermarkets, and warehouse clubs, and aggregate sales/size of modern convenience stores and traditional stores/free markets in Brazil, Russia, China, India, Mexico and the Philippines. Among other things, we match these data with affluence data from C-GIDD and our clients’ distribution footprint to identify expansion opportunities. We complement the database with a statistical model (Traill) that predicts the spread of modern trade in emerging countries.

• Our Cellular Operator Analysis Tool (COAT) contains data from 36 countries with 74 operators and allows us to model expected operator profitability and growth based on factors such as an operator’s market-entry order (1st, 2nd, or third), its ARPU strategy, number of subscribers for voice and data, prepaid %, cumulative investments (marketing and infrastructure), etc. The database is analyzed with a structural equation model.

• The US recession prediction database predicts the likelihood of a recession 5-12 months out. We use this econometric analysis to enhance our short- and medium-term demand projections in various categories. We publish a recession probability estimate every week using the term-spread method at http://recession.canback.com.

Over the years, we have collected a library of demand elasticities for numerous products and services. This allows us to quickly build predictive demand models with short- to medium-term horizons using the Golder Tellis approach. The graphic to the right shows Golder Tellis’ base specification, where each driver (in green) has an elasticity that can be calculated. We complement this base specification with consumer behavior and product/service-specific variables that show our clients how (and by how much) their actions can influence demand.

In short, we overlay historical demand data for the category or brand at hand with chosen elasticities from the library. This makes it possible to quickly estimate future market sizes without going through an elaborate data collection effort. An example is a model we built for cellular services in Brazil and other countries (see the sanitized Excel version here).

For most tasks, we do not build our own databases. Rather, we use external databases and merge the data with, for example, C-GIDD. Identifying and leveraging external databases is in itself a skill that is hard to build. We have immense experience in using disparate data sources and know the benefits and drawbacks of them.

For example, are prefecture-level GDP numbers reliable in China? More so than Chinese disposable income numbers?  Do Euromonitor category sales data suffer from autocorrelation when used in pooled time series cross-sectional analyses? How comparable are FAO and USDA food consumption databases? Can Nielsen data be used to reliably estimate price elasticity by store format? What is the best way to merge three databases where one reports bimonthly sales starting with January-February and the other starts with December-January, while consumer spending and confidence is reported quarterly?

Having performed these kinds of tasks thousands of time, we now have an experience base second to none and can build integrated databases quickly to attack most market assessment problems.