How Income Distribution Shapes Category Demand

This newsletter explains the income distribution concept, shows why it is important, and discusses applications. Understanding income distribution profoundly changes a company's view of its current and future markets.

Income distribution determines market size, growth prospects, premiumization opportunities, and much more. 20 years ago such data did not exist. Today, our TelluBase dataset offers them across and within subnational geographic units for all countries in the world.

What is Income Distribution?

Income distributions show how many people there are at given income levels. To make it useful, there are three conditions:

• Incomes are expressed in absolute (e.g., dollar) values and not in percentages of a country's average income. Often, there are statements such as "the bottom 30% of the population earns 10% of the income." But that was more or less true 50 years ago, and it is more or less true when comparing rich and poor countries. Thus, looking at income in absolute terms is the only way to gain meaningful insights.
• Income and its distribution is defined the same way across countries. Too often, definitions differ between countries. One country may use gross income, another disposable income. The distribution metric usually differs too. Some countries use consumption expenditure, others use pre-tax income.
• The highest incomes should be captured correctly. This is difficult to do, leading to underestimates of affluent people.

The graph below shows an income distribution from the United States.

Our world-leading TelluBase product covers income distribution and other metrics for 218 countries, 2,600 cities, and 2,400 subdivisions (states, provinces, etc.) from 2000 till 2050.

Why is Income Distribution Important??

Companies typically model markets based on averages. This reduces predictive power and in the end leads to mis-allocation of resources. We estimate $6 billion globally in savings for large FMCG companies from implementing a robust income distribution paradigm, and even larger revenue benefits of $9 billion.

The three graphs below illustrate the importance. They show how the predictive accuracy when estimating new car sales increases substantially when income distribution is used.

• The left graph reflects how most of us think about markets. It shows unit sales against population. This is the same as sales per capita. While we may not think about it, when we discuss per capita sales we do this with very low accuracy.
• The middle graph shows sales regressed against average income. The explanatory power increases, but it is not good enough.
• The right graph shows a better model. It uses the top 20% of global income earners (with big differences between countries for what the local percentage is). Now we have a goodness of fit four times better than in the first graph, and an astonishing R2 of 0.85.

Below are some other categories.

What Are Some Applications?

There are numerous applications for the income distribution approach. Here are five examples:

1. Quantifying market sizes. Often, market sizes are not known between and within countries. All channels may not be not covered or data may be missing. The income distribution approach allows companies to create complete, harmonized, and accurate market sizes.
2. Predicting category and brand demand. Predicting demand is often based on historical data. But we know much more. We have good estimates of future population and income growth, we know antecedent categories, and we can pool countries. The income distribution approach makes this a high accuracy exercise.
3. Improving pricing decisions. Income distributions can be applied to price segments, brands, and items (and not only categories). A key determinant of good pricing is the work effort required, which is price divided income per time unit. Thus, price decisions are directly linked to the relevant income brackets.
4. Finding latent market opportunities. Companies typically know how much they sell, but not what the latent market is. By using "envelope analysis" based on income brackets (and some other methods), it is possible to show what demand could be by income bracket, rather than what it is.
5. Converting survey micro data to macro-level insights. Consumer surveys are not easily scaled to the national level and back down to cities or subdivisions. By linking them to age × income brackets it is possible to get granular estimates. It is also possible to see how segments grow over time in the future.

Making the understanding and analysis of income distribution part of the corporate fabric is well worth the effort. Tellusant has the tools and data that facilitates the transition to a more efficient future both as strategy consultants and as cloud-based services providers.

TelluBase: Unique subnational income distributions on a global scale. 218 countries—2600 cities—2400 subdivisions

PoluSim: World-leading strategic forecasting product (5-10 years out) for categories and brands, own and competitors

PACE (Pricing Against Consumer Economics): Revolutionary latent market-opportunity quantifier

We are more than happy to discuss this further. Please contact me with a LinkedIn message or by email.

Staffan and The Tellusant Thought Team

Canback, Burgin-Young, Shen, and Ezepik