An endorsement for endorsements

Virat Kohli is the new pot of honey for marketers. After his recent centuries, the cricketer is being chased by brands willing to pay as high as Rs 3 crore for his endorsements.

Obviously marketers want to cash in on him while his bat talks. But what is the economic impact of such celebrity endorsements? Especially as marketers are increasingly banking on the pull of a celebrity — a 2010 global study found that 20 per cent of advertisements feature an endorsement and, in some countries, it is as high as 45 per cent.

Also, is it worth the potential risks of embarrassment for the company if the celebrity gets involved in a scandal? Studies showed that negative publicity around Tiger Woods' marital infidelity decreased the stock market performance of firms that he endorsed.

A recent National Bureau of Economic Research (NBER) working paper by Craig L. Garthwaite — You Get a Book! Demand Spillovers, Combative Advertising, and Celebrity Endorsements ( http://www.nber.org/papers/w17915 ) — tries to answer these questions and more.

Garthwaite estimates the direct sales effect of an endorsement, the impact on other unadvertised products sold under an umbrella brand, and also whether it results in market expansion or business-stealing — getting a consumer to shift from a rival product.

Endorsements are not a new phenomenon. As Garthwaite points out, the use of a celebrity to sell a product easily dates back to Pope Leo XIII's 1899 endorsement of Vin Mariani, an alcoholic drink containing Bordeaux wine and coca leaves that reportedly inspired the original Coca-Cola recipe.

There's also a whole body of research on the impact of endorsements. Garthwaite cites a 2009 survey by MEC Global which found that nearly a quarter of consumers surveyed said they had bought a product because of a celebrity spokesperson. But significantly, half of all consumers in the survey reported that they only noticed an endorsement if it was for a product category in which they were already interested.

Research has also shown that endorsements have increased stock market performance for the advertising firm (Agrawal and Kamakura, 1995) and, to a more limited extent, higher sales of the advertised product (Elberse and Verleun, 2011).

But Garthwaite says there are empirical issues with many of these studies. The biggest being the potential endogeneity in the timing and selection of an endorser.

The Oprah Way

For his research, Garthwaite studied the effect of Oprah Winfrey's Book Club, which she started in the mid-'90s to get “America reading again”. Over the Club's 15 years of existence, Winfrey endorsed 70 titles that sold millions of post-endorsement copies.

And to study the effect on sales, Garthwaite uses Nielsen's BookScan panel, a point-of-sales dataset that represents nearly 75 per cent of total US book sales.

At the outset, Garthwaite does point out that Oprah never charged money for her book endorsements — but this works in the study's favour.

Also books as a category are experience goods — the quality of the story and writing are only revealed after the purchase.

First, Garthwaite tracks Oprah Winfrey's value as an endorser — there is no doubt that she commands significant influence. Time magazine named her along with Mahatma Gandhi and Einstein as big influencers of the 20th century. And repeated her in its 21st century list where she rubs shoulders with Bill Gates and Nelson Mandela.

From a commercial standpoint too, Oprah's value as an influencer is proven — for instance, when in her show she spoke for 42 seconds about holiday smoked turkeys from Greenberg Smoked Turkeys of Tyler, Texas, the firm in the next two weeks sold 22,000 turkeys to new customers, generating over $1 million in revenue. Eight years later the company was selling 200,000 turkeys a year, a 33 per cent increase over its average annual pre-endorsement sales. Before Oprah talked about it, the firm was only averaging 5,000 new customers a year.

Garthwaite's study is pretty comprehensive — he looks at the impact of endorsements on book sales genre by genre. And finds that the popular adult fiction category actually saw no impact, but the harder books benefited.

Overall, he finds that there is no evidence to support the popular belief that Oprah's Book Club increased the population of Americans who read regularly. On the contrary, it's the existing readers who read the books she suggested.

Cause and Effect

Garthwaite finds three main effects of the book endorsements. First, consumers immediately increase purchases of an endorsed product. This estimated sales effect peaks at an approximately 400 per cent increase during the week following the endorsement. In subsequent weeks, the effect declines, yet remains large and is statistically significant more than half a year later.

Second, endorsements generate meaningful spill-over benefits for co-branded goods. What this means is, if one book of an author is endorsed, all other books by the author also benefit. “From this pattern of spill-over demand, we can infer that an endorsement for one product can provide important information to consumers about related goods or services,” finds Garthwaite.

Finally, the author finds that despite causing large changes in purchasing behaviour, endorsements result in more business-stealing than market expansion. In the publishing sector, aggregate book sales do not increase following a club endorsement - suggesting that non-readers are unaffected.

COMMENT NOW