Showing Prior Choices Concentrates Consumer Demand

CommunityMaecker, O.; Grabenströer, N.S.; Clement, M.; Heitmann, M. · 2013International Journal of Research in Marketing
Topicssocial influence·sales charts·herding·market concentration·gini coefficient·replication·online experiment

Online sales charts increasingly show shoppers how many people chose each product. Demand is more concentrated when a few products take a larger share of all choices. Imagine your team is deciding whether to show prior choices beside the songs in a music storefront. How would that change demand, and how well could you forecast market share?

When shoppers see what others already chose, they crowd toward popular products, which concentrates demand and makes market share harder to forecast.

Data chart

Showing prior choices concentrates demand and interest

Percent rise in how unevenly choices spread across a category's 30 products when shoppers see prior choices, compared with seeing none. Unevenness runs from 0, every product chosen equally, to 1, one product taking every choice.

Music demand+50%Movie interest+19%Scarf consideration+15%
  • A higher score means choices are less evenly spread.
  • Each purchase stage was tested in a different category.

Showing prior choices concentrates actual music demand, interest in movies and consideration of scarves.

Online choices across 30 products per category; each purchase stage was tested in a different category. Exact concentration scores with prior choices hidden and shown: music .147 and .220; movies .088 and .105; scarves .155 and .178 (Table 1, p. 430). Bars show the percent rise between each pair, rounded to whole numbers, not a rise in total demand. The earlier music experiment reported .2 and .45–.55, read from its graphs; these are a separate study benchmark.

Data chart

Midrange songs can gain or lose share when choices are visible

Largest observed fall and rise in a midrange song's market share when shoppers see prior choices rather than none, in percent change.

40%Largest fall in a midrange song's share65%Largest rise in a midrange song's share
  • A midrange song can lose up to 40% of its share.
  • A midrange song's share can rise by up to 65%.

Showing prior choices can move a midrange song's market share sharply in either direction.

Observed extremes for intermediate songs in the artificial music markets, compared with choices made without prior-choice information: decreases by up to 40% and increases by up to 65% (p. 430; signed values −40% and +65%). The paper does not define how many songs count as intermediate and does not report these extremes for movies or scarves.

Action guide

  1. Favor sales channels that show ranks for your strongest products.High-quality products with high current or expected market share are likely to sell more there, because showing prior choices concentrates demand rather than lifting all products equally.
  2. Ask your insights team to revisit market share forecasts.Showing prior choices shifts demand toward already popular products and away from less popular ones, which forecasts built on private choices do not capture.
  3. Plan for more uncertainty around your midrange products.Showing prior choices can move their market share sharply in either direction, so a single correction to forecasts built on private choices will not suffice.
  4. Have your team size the effect in your category first.Showing prior choices changes what shoppers choose in every category tested, but the size differs by category and from an earlier music experiment.

Evidence

  • Showing prior choices raises a 0 to 1 concentration score by about 50% for music demand, 19% for movie interest and 15% for scarf consideration, compared with showing none.
  • When prior choices are shown, the top ten products gain market share, on average about 2% for music, 4% for movies and 8% for scarves. The bottom ten lose about 20%, 1% and 5%, respectively, compared with markets where prior choices are hidden.
  • Products in the top and bottom ten percent when prior choices are hidden stay there when choices are shown. Quality still matters, and showing prior choices appears to amplify its effect on demand at those extremes.
  • Among midrange songs, showing prior choices can cut market share by up to 40% or raise it by up to 65%, compared with markets where prior choices are hidden.
  • An earlier music experiment found concentration roughly 125% to 175% higher with prior choices shown than hidden, read from its chart, against about 50% higher in this experiment. The direction matches, but the effect is smaller here.

Key takeaway

When shoppers see prior choices, they crowd toward popular products, which makes market share harder to forecast.

Limitations

The experiment covers online choices about music, movies and scarves in a panel from Western Europe, and only music involved actual demand. Movies and scarves involved stated interest or consideration, and the study did not track retail sales or profit.

Source

Maecker, O., Grabenströer, N.S., Clement, M., & Heitmann, M. (2013). Charts and demand: Empirical generalizations on social influence. International Journal of Research in Marketing, 30, 429–431. https://doi.org/10.1016/j.ijresmar.2013.08.001.

https://doi.org/10.1016/j.ijresmar.2013.08.001

Exact values behind the rounded figures: concentration is the Gini coefficient of choices across the 30 products in each category, where a higher value means fewer products take a larger share of choices; values without and with prior choices are .147 and .220 for music (about 50% higher), .088 and .105 for movies (about 19% higher), and .155 and .178 for scarves (about 15% higher); average top-ten market-share gains of 2% for music, 4% for movies and 8% for scarves, and bottom-ten changes of −20%, −1% and −5%, respectively; midrange music market-share falls of up to 40% (−40%) and rises of up to 65% (+65%); earlier music experiment about .2 without and .45–.55 with prior choices (roughly 125% to 175% higher), versus .147 and .220 in this replication.