De-averaging

De-averaging as a strategic concept is closely associated with management consulting firms, particularly McKinsey & Company. The core insight is deceptively simple: averages lie. When data is aggregated across products, customers, geographies, or time periods, the resulting averages can mask dramatic differences that have profound strategic implications. A business that appears moderately profitable on average may actually consist of a highly profitable core subsidizing a deeply unprofitable periphery. Without de-averaging, management makes decisions based on a misleading composite picture.

The strategic power of de-averaging lies in its ability to redirect resources toward their highest-value uses. One of the most common applications is customer profitability analysis. In most businesses the profit distribution across customers is far more uneven than the revenue distribution — a minority of accounts carry the result, and some accounts subtract from it. This pattern, far more extreme than the simple Pareto principle suggests, only becomes visible when you de-average customer-level data. Without this granularity, companies invest equally in all customers, over-serving unprofitable ones and under-serving their most valuable accounts.

McKinsey's work on corporate resource allocation shows why de-averaging matters at scale. Its research found that most large companies allocate capital across their business units in nearly the same proportions year after year, even though returns differ widely from one unit to the next, and that companies which reallocated resources more actively delivered higher returns to shareholders. When capital follows historical allocations and blended corporate performance rather than unit-level, forward-looking returns, profitable units can be starved of investment while underperforming units keep receiving funding.

In retail, Walmart's success has been partly attributed to its sophisticated de-averaging of store-level and product-level data. Rather than managing categories based on chain-wide averages, Walmart analyzes performance at the individual store level, adjusting assortment, pricing, and inventory to local demand patterns. This granular approach, enabled by its early investment in data systems from the 1980s onward, helped Walmart compete against retailers that managed their businesses on regional or national averages.

Practitioners should apply de-averaging across multiple dimensions simultaneously for maximum insight. A product that appears average in profitability might be highly profitable in one region and deeply unprofitable in another. A customer segment that appears average in retention might contain one cohort with 95% retention and another with 50%. The key is to disaggregate along the dimensions most relevant to strategic decision-making: customer segments, product lines, geographies, channels, and time periods. The more granular the analysis, the more actionable the insights, though diminishing returns eventually set in as data becomes too sparse for reliable conclusions.