This is the first part of a two-part review of the academic paper ‘Seeding Strategies for Viral Marketing: An Empirical Comparison’ by Hinz, Skiera, Barrot and Becker. It is tough going to read it but the information held within it is solid so we have been through it (with strong coffee in hand) so that you do not have to and extracted the key actionable information for video marketers and social media marketers:
In part one, we are going to give you an introduction to the main theories and discussions in the paper. In part two we will talk in more depth about practical application of the theory.
If you choose to read the paper yourself (just Google the above title and click on the PDF result), beware of the complex formulas, which might not be your thing! We have summarised the main points in lay terms so that you can get an idea of the kind of discussions taking place in the field.
Viral marketing campaigns
Individuals rely on professional and personal recommendations when making decisions about, well, anything. Social media facilitates this process perfectly, which is why many marketing budgets have been shifted away from mainstream methods towards online methods. A seeding strategy involves initiating a marketing campaign by carefully selecting a target group of seeds who it is felt will maximise the effectiveness of the ‘diffusion’ of a campaign. A basic example would be a company who have a video to promote a new product; they would select well-connected individuals on social media platforms to launch the online spread of the video.
Theories about viral marketing were traditionally based on epidemiology – the study of the patterns, causes and effects of disease (and other health related incidents but ‘disease is snappier!).
With viral marketing campaigns, your customers are effectively doing the work for you. This makes it a comparatively inexpensive option. However, if you want to reap rewards similar to Hotmail (who spent a minuscule $50,000, over 18 months and generated 12 million subscribers!) you need to think carefully about the following:
- the content and the message of the campaign – it needs to be memorable
- the structure of the social network you have decided to target
- the behaviour of your target market and what incentives might motivate them
- the seeding strategy – which is what will determine who those initial target consumers are
Choosing the initial seeds
The fourth point is the one we are interested in here. What is particularly interesting are the elements of control in the hands of the marketer when considering seeds. The marketer should be selecting these initial consumer seeds on the basis of two main factors
1) social characteristics relevant to their campaign and their goals
2) knowledge of metrics and stats relating to consumer networks
Making these factors a primary consideration, marketers should be able to develop formulaic approaches to seeding their campaigns and the seeds to target in the first phase of their campaigns. However, it is not that easy.
It has been argued that there is a great need for more extensive research and experimentation of seeding in order to understand the true role of these initial consumers or ‘hubs’.
Previous research
- A good strategy can be up to 8 times more successful than other strategies – so there is a good deal of value in working to find the right seeding strategy. There is much research to still be done but what has already been done has shown that seeding strategies can have a big influence on the success of a viral marketing campaign.
- Previous research found that better-connected seeds, or ‘hubs’, actively use their reach but have less influence on peers. This could be due to the fact that a hub is unlikely to be connected on a more personal level with their hundreds or thousands of connections. It could be argued that users with fewer connections have more influence over them because they are connected more intimately somehow.
- People with less connections are referred to as ‘fringes’. There are also ‘bridges’ who connect two otherwise unconnected elements of the social network.
Key new findings – participation
- The findings of this new research contradict this theory that individuals with fewer connections have more influence. It has found that better connected individuals are best to seed to for another reason: they are more likely to participate. Leskovec, Adamic, and Huberman (2007) have evidenced the fact that ‘hubs’ do send more messages.
- It appears that the definition of a hub should be based on messaging behaviour, as well as number of connections.
It is unsurprising that viral marketing is a hot topic as it appears that marketers are getting closer to formulating ways to make campaigns go viral – an extremely powerful tool. However, there is still some distance to travel. It is not unreasonable to suggest that there is some truth in all of the research. The interesting thing about this paper is that their experiments include a look at which combinations of seeds are most likely to yield the best results.
The social network as a crucial determinant
The paper identifies that the social network remains a crucial determinant of the best seeding strategies, mainly because a social structure is much easier to observe and analyse than the intensity and frequency of actual communication. So information about mutual relationships, even without the more detailed information around actual interaction, is still useful when deciding a seeding strategy. The paper states that robust results to support this have been obtained when experiments have been carried out which have involved control of the level of communication activities.
What this means for marketers
Once you have done a great job on creating content you need to identify the right people to market your content to so that it then disseminates further. Luckily there are some fantastic tools for doing this now available. what I am talking about is identifying social influencers in your niche.
- Little Bird – http://getlittlebird.com/ – it’s still in a private beta mode but it is well worth a try and if you happen to be a large business or a mid to large marketing agency then you may be able to afford the $250 per month for the full product. When I tested this I entered a keyword “video marketing” and was then given various lists of the key influencers in the video marketing space – it was incredibly accurate. The tool analyses all of the interconnecting relationships that people have and gives you all of the data in an easy to use format.
- Followerwonk – https://followerwonk.com/ – now part of the SEOMoz suite of tools with a requirement that you pay the $99 per month for their pro membership to use the tool. This is frustrating as they used to be independent but it is what it is. This tool is less complex than Little Bird and just sorts people based on things like number of Twitter followers and also Kred rank.
- Twitter – The new browse categories function – https://twitter.com/i/#!/who_to_follow/interests – allows you to enter a keyword and see the people related to that keyword. This looks relatively simple as i think that it is just returning results based on Twitter profile data without consideration of inter-relationships.
Here is a slideshare summary of the research by one of the authors
Next time…
In the next installment we discuss the four empirical studies that were carried out for this research and summarise the analysis regarding the reasons for differing levels of success of different seeding strategies.