What Is Viral Marketing?

One of the reference texts is the essay "Viral Marketing" that Steve Jurvetson and Tim Draper, investors in Hotmail, published in 1997. They describe how Hotmail added an advertising line with a link to every email its users sent: "Get your free email at Hotmail". Each message thus carried an advertisement and an implied endorsement by the sender, because the recipient could see that someone they knew already used the service. In the 1998 version, the authors put the result at over 12 million users in the first year and a half, with less than 500,000 dollars spent on marketing.

The example shows the two forms the term usually takes. In the first, spreading is part of using the product: whoever uses it shows it to others. In the second, a brand creates a piece of content, such as a video or an image, in the hope that it will be shared far beyond its own audience.

Social media marketing is something else: planned, ongoing work on social networks with the brand's own profiles and ads. A viral campaign can use social networks, but equally email or any other channel through which users pass content to each other. Social media marketing has its own entry in this glossary.

There is data on which content gets shared more. Jonah Berger and Katherine Milkman analysed every New York Times article over three months and published their results in 2012. Positive content was shared more than negative content, but valence did not explain everything. Arousal mattered: articles that evoked high-arousal emotions, positive ones such as awe or negative ones such as anger or anxiety, were shared more. Articles that evoked low-arousal emotions such as sadness were shared less. Content that was surprising, interesting or practically useful was also shared more.

These factors raise the probability that something gets shared. They do not guarantee that it becomes a phenomenon.

Anyone planning a viral campaign needs to know how often it works. Goel, Anderson, Hofman and Watts measured this in a study published in Management Science in 2016. They analysed around a billion diffusion events on Twitter: news stories, videos, images and petitions. More than 99 % of cascades were tiny and ended within a single generation. Moderately popular content occurred about once in a thousand events, viral hits closer to once in a million.

The second finding is less comfortable for campaign planners. Even among the most popular content, person-to-person spreading was usually low. Popularity depended mainly on the size of the largest single broadcast, that is, on whether an account or a media outlet with a large audience published it. What looks viral often travelled far through one big megaphone.

The practical consequence is to plan the reach you can control and to treat spontaneous spreading as a possible bonus, not as the aim of the campaign.

Faking that spontaneity has legal limits. The Unfair Commercial Practices Directive 2005/29/EC, which the member states have transposed into national law, lists among the practices considered unfair in all circumstances: paid editorial content that is not identified as such, falsely representing oneself as a consumer, and submitting or commissioning false consumer reviews or endorsements, or misrepresenting consumer reviews or social endorsements, in order to promote products.

• Build spreading into the use of the product where possible, as Hotmail did with every email sent.
• Plan initial reach through owned or paid channels, because popularity depends mainly on the largest single broadcast.
• Create content that surprises, is useful or evokes strong emotions, knowing that this only raises the probability.
• Label all paid content as advertising, including when a third party publishes it.
• Set campaign goals that still make sense if the content does not go viral.

• Promising or budgeting for a viral result.
• Copying a success story and expecting the same effect, when viral hits occur about once in a million.
• Faking enthusiasm with accounts posing as consumers or with commissioned reviews.
• Crediting person-to-person spreading with reach that came from a single post with a large audience.
• Measuring only how often content was shared and not whether it met the campaign goal.