ScholarSignal
Menu
Sign inEvaluate an idea

Evidence · Research methods

Empirical generalizations in marketing: what the research shows

An empirical generalization is a pattern that repeats across many conditions, such as categories, countries and time periods, and can be stated simply. Marketing's most durable examples include the negative binomial pattern of repeat purchasing, first shown in 1959, and the Bass model of new-product diffusion. Proponents argue that patterns replicated across many data sets deserve more trust than single significance tests, which flexible analysis choices can easily push below p < .05.

Definition

An empirical generalization is a regularity that holds across different circumstances and can be described simply, with a formula, a graph or a sentence. It comes with scope conditions: stating where a pattern stops holding is part of the generalization. The approach puts many sets of data first and theory after, and treats replication under deliberately varied conditions, rather than one well-controlled study, as the way knowledge accumulates.

What the research finds

Many sets of data first, theory after

A 1995 special issue of Marketing Science set out what an empirical generalization is and why marketing needs them, arguing that regularities can come before the theory that explains them. Replication, on this view, is not about repeating a study identically but about finding the range of quite different conditions under which a result holds, along with its predictable exceptions.[1][2][3]

The oldest examples still hold

In 1959 the number of purchases of a brand per consumer was shown to follow a negative binomial distribution, so aggregate repeat buying follows from a simple regularity. A decade later, the Bass model described how first purchases of new durables grow through innovation and imitation, and it fitted the sales histories of eleven appliance categories. The Dirichlet model later tied such patterns together: big and small brands differ mainly in how many buyers they have, not in how loyal those buyers are.[4][5][6]

Patterns travel to new categories, within limits

Nearly 50 years of replication of the Dirichlet model underpin how the double jeopardy pattern is drawn, and a simple straight-line version works only under limited conditions. In U.S. cigarette purchasing, an addictive product, buying frequencies followed the usual pattern, loyalty was high, and double jeopardy was present but weak.[7][8]

Managers need effect sizes, not just directions

An editorial argues that marketing needs a knowledge base of quantified effect sizes so that marketing actions can be translated into financial returns. A meta-analysis of 863 elasticities from 89 studies illustrates this: the average link between market share and financial performance is .132, lower than for other marketing metrics, and weaker for business-to-business firms, service firms and U.S. markets.[9][10]

Generalizations now cover many fields

Well-supported generalizations exist for reference prices: consumers compare prices with an internal standard shaped by past prices. Studies of every film released in the U.S. (2,098 films) and Germany (1,360 films) produced generalizations about film supply and demand that correct earlier selection and endogeneity problems. Across seven industrialized countries, accountability, innovativeness and customer-connection skills were consistently related to the marketing department's influence.[11][12][13]

Behavioural effects get conceptual replications

Replication is not limited to buying patterns. A replication supported the IKEA effect, the higher willingness to pay for products people assemble themselves, and pointed to psychological ownership as the mechanism. Three replications of a classic music-in-advertising effect found only weak effects for low-involvement products and none for higher-involvement ones, and a French replication showed that intrusive product placements were liked less by viewers under high mental load.[14][15][16]

Single significance tests are easy to game

Simulations and experiments show that flexibility in collecting, analysing and reporting data can, in many cases, make a false effect reach statistical significance more often than a true null is correctly found. Replication across independent data sets is the natural counterweight.[17]

How strong is the evidence?

Much of the work on lawlike buying patterns comes from one school of researchers and from consumer panel data in repeat-purchase categories, so claims that these patterns hold everywhere go beyond what has been tested. Descriptive models say what happens more reliably than why, and the debate between descriptive models and more elaborate causal models remains open. Conceptual replications do not always succeed, as the music-in-advertising studies show. Which patterns from digital platforms have replicated across platforms and data sets is still largely open, and replication assumes the underlying data come from real people rather than from AI-generated responses.

References

  1. Frank M. Bass (1995). Empirical Generalizations and Marketing Science: A Personal View. Marketing Science.
  2. A. S. C. Ehrenberg (1995). Empirical Generalisations, Theory, and Method. Marketing Science.
  3. R. Murray Lindsay, Andrew S. C. Ehrenberg (1993). The Design of Replicated Studies. The American Statistician.
  4. A. S. C. Ehrenberg (1959). The Pattern of Consumer Purchases. Applied Statistics (Journal of the Royal Statistical Society, Series C).
  5. Frank M. Bass (1969). A New Product Growth for Model Consumer Durables. Management Science.
  6. Andrew S. C. Ehrenberg, Mark D. Uncles, Gerald J. Goodhardt (2004). Understanding Brand Performance Measures: Using Dirichlet Benchmarks. Journal of Business Research.
  7. Cullen Habel, Larry Lockshin (2013). Realizing the value of extensive replication: A theoretically robust portrayal of double jeopardy. Journal of Business Research.
  8. John Dawes (2014). Cigarette brand loyalty and purchase patterns: An examination using US consumer panel data. Journal of Business Research.
  9. Dominique M. Hanssens (2018). The value of empirical generalizations in marketing. Journal of the Academy of Marketing Science.
  10. Alexander Edeling, Alexander Himme (2018). When Does Market Share Matter? New Empirical Generalizations from a Meta-Analysis of the Market Share–Performance Relationship. Journal of Marketing.
  11. Gurumurthy Kalyanaram, Russell S. Winer (1995). Empirical Generalizations from Reference Price Research. Marketing Science.
  12. Michel Clement, Steven Wu, Marc Fischer (2014). Empirical generalizations of demand and supply dynamics for movies. International Journal of Research in Marketing.
  13. Peter C. Verhoef, Peter S.H. Leeflang, Jochen Reiner, Martin Natter, William Baker, Amir Grinstein, Anders Gustafsson, Pamela Morrison, John Saunders (2011). A Cross-National Investigation into the Marketing Department's Influence Within the Firm: Toward Initial Empirical Generalizations. Journal of International Marketing.
  14. Marko Sarstedt, Doreen Neubert, Kati Barth (2016). The IKEA Effect: A Conceptual Replication. Journal of Marketing Behavior.
  15. Ivar Vermeulen, Camiel J. Beukeboom (2015). Effects of Music in Advertising: Three Experiments Replicating Single-Exposure Musical Conditioning of Consumer Choice (Gorn 1982) in an Individual Setting. Journal of Advertising.
  16. Felipe Pantoja, Patrícia Rossi, Adilson Borges (2015). How Product-Plot Integration and Cognitive Load Affect Brand Attitude: A Replication. Journal of Advertising.
  17. Joseph P. Simmons, Leif D. Nelson, Uri Simonsohn (2011). False-Positive Psychology. Psychological Science.

Bibliographic details only. Follow the DOI to read each study at its publisher.

Related topics

A synthesis of 17 studies, written by ScholarSignal. Last reviewed 2026-09-26.

Building on this evidence?

ScholarSignal checks a new research idea against the closest existing work and an editor rubric, and tells you whether the contribution is new enough for the journals you target.

Evaluate my idea for free