How Online Reviews Move Sales

ContentYou, Y.; Vadakkepatt, G.; Joshi, A. · 2015Journal of Marketing
Topicsewom elasticity·online reviews·meta-analysis·review volume·review valence·social media

Electronic word of mouth (eWOM) covers online reviews, blogs, forums and social posts. Volume counts the messages. Rating or tone measures how positive they are. Imagine planning next year's eWOM budget. Which products and sites come first?

Sales response to electronic word of mouth varies by product, site and competition.

Data chart

Sales response to more online reviews and posts varies by product and site

Predicted percent change in sales for a 1% rise in the number of online reviews and posts, by product or site type.

0Product durabilityNondurableDurable-.01%.52%Ability to try before purchaseEasy to tryHard to try.21%.62%Product consumptionPublic usePrivate use.06%.52%Site expertiseGeneral sitesSpecialized sites.05%.69%Site ownershipRetailer sitesIndependent sites-.03%.47%
  • Hard-to-try products have a larger sales response to message volume.

Sales respond more to message volume for durable, hard-to-try and privately used products, and on specialized and independent sites.

Dot labels are rounded to two decimals. Exact predictions from Table 4 (p. 32), allowing for the other study characteristics: durable .516%, nondurable -.007%; hard to try .624%, easy to try .210%; private use .519%, public use .057%; specialized sites .691%, general sites .045%; independent sites .469%, retailer sites -.026%. The comparisons are differences between predicted responses, not tests of each dot against zero. Rating or tone is omitted because durability and site expertise were not estimable for that measure. Competition has no paired predictions because it is measured continuously.

Data chart

Online rating and message responses exceed short-term advertising

Average percent change in sales for a 1% rise in review or post rating or tone, in the number of online reviews and posts, or in short-term advertising spending.

Short-term advertising benchmark: .12%Online review and post rating ortone.42%Online review and post count.24%

Online rating or tone has a higher average sales response than message volume, and both exceed the two short-term advertising benchmarks.

Bar labels are rounded to two decimals. Exact average sales responses across 51 studies (339 volume and 271 valence estimates): .417% for rating or tone and .236% for message count. Table 6 (p. 36) reports short-term advertising benchmarks of .12% (Sethuraman, Tellis and Briesch, 2011) and .22% (Assmus, Farley and Lehmann, 1984), from separate studies. Only rating or tone exceeds short-term personal selling (.31%); price has larger absolute responses (-1.76% and -2.62%). These values measure responsiveness, not returns.

Action guide

  1. Treat averages across studies as context only.Before approving eWOM spending, ask your insights team for estimates on your own products and sites.
  2. Prioritize products whose sales respond most to eWOM.Focus on hard-to-try and privately used products, and on volume for durable goods.
  3. Choose the sites that carry your eWOM.Favor independent over retailers' own review sites for both measures, specialized over general sites for volume, and community sites for rating or tone.
  4. Pay attention to online complaints.Have your measurement team report negative and average ratings separately before approving rating or tone assumptions.
  5. Avoid relying on eWOM alone in intensely competitive markets.In volatile markets, pilot leaning more on advertising and promotion; the study did not test that plan directly.
  6. Use volume estimates when studies conflict.Ask your insights team how far its estimates for rating or tone depend on model and data choices.

Evidence

  • On average, a 1% rise in rating or tone is linked to about a .42% rise in sales, against about .24% for the same rise in volume.
  • Both averages exceed the short-term response to advertising, but only rating or tone exceeds the .31 short-term response to personal selling.
  • For hard-to-try products, a 1% rise in volume is linked to about .62% higher sales, against about .21% for easy-to-try products. Privately used products show a stronger response to both measures, and durable goods to volume, the only measure estimated for them.
  • Independent review sites show a stronger response than retailers' own sites on both measures. Specialized sites show a stronger volume response than general sites; rating or tone was not estimated. Community sites show a stronger response to rating or tone but no reliable volume difference.
  • Sales respond less to both measures in industries with more competitors, and industry growth has no established link with either.
  • Estimates for rating or tone vary with the model used and its adjustments for reverse influence and study differences; volume estimates are steadier. Models using negative instead of average ratings estimate a weaker response.

Key takeaway

On average, sales respond more to rating or tone than volume, and the response varies by product, site and competition.

Limitations

The 51 studies measure sales response, not profit or the best budget split. Their averages do not show what a particular investment will earn.

Source

You, Y., Vadakkepatt, G. G., & Joshi, A. M. (2015). A meta-analysis of electronic word-of-mouth elasticity. Journal of Marketing, 79(2), 19–39. https://doi.org/10.1509/jm.14.0169

https://doi.org/10.1509/jm.14.0169

Exact values behind the rounded figures: average sales responses of .417% for rating or tone and .236% for message count; predicted volume responses of .516% for durables and -.007% for nondurables, .624% for hard-to-try and .210% for easier-to-try products, .519% for private and .057% for public use, .691% for specialized and .045% for general sites, and .469% for independent and -.026% for retailer sites.

Read the paper ↗