A new piece in Nieman Journalism Lab talks about influence and the networked nature of Twitter. It refers to a small experiment by the authors of the book ‘Connected: The Surprising Power of Our Social Networks & How They Shape Our Lives’ – Nicholas Christakis and James Fowler on trying to see how many books they sell via tweet recos by allegedly influential tweeters like Alyssa Milano, Tim O’Reilly and Pew’s Susannah Fox. It seems they sold just 4 books with those 3 tweets from those twitter stars – just 4.

That led to Christakis noting, “these links â?? these Twitter links â?? are weak”. Sounds more like Woody Allen/Kramer’s ‘these pretzels are making me thirsty’, but I digress.

The following is a relevant excerpt from the post that I’m going to add my views on.

…binary that might actually be relevant in that regard, Christakis suggested: influencer versus influence-ee. â??If weâ??re really going to advance this field, we need to figure out how to identify not just influential people, but also influenceable people,â? the professor noted. â??We need not just shepherds, but sheep.â? And â??if weâ??re going to exploit online ties,â? Christakis said â?? say, by creating communities of interest around news content, and potentially monetizing those communities â?? then â??measures of meaningful interactions will be neededâ?: We need metrics, in particular, to determine â??which online interactions represent real relationships, where an influence might possibly be exerted.â?

If you want your work to have impact, then targeting a bundle of closely connected networks â?? with news, with links, with messages â?? may make more sense than going for numbers alone. Spreading a conversation is not the same as affecting it. â??Iâ??m not saying that Twitter is useless,â? Christakis said, â??but I think that the ability of Twitter to disseminate information is different than its ability to influence behavior.â?

A couple of observations.

1. The first para in the excerpt above talks about finding ways to monetize communities and how influence could be exerted to achieve that. The question then is, ‘Is monetizing the holy grail of influence’? I suppose I’d get a resounding yes from brands and marketers, but, in the path to monetizing we may be missing a lot of nuances that cannot be measured at least by current systems. For instance, in the Alyssa Milano example, I understand that the post does refer to appropriateness of influencer with regard to the product/service being peddled, but what about people who bought the book not from Alyssa’s tweet, but via one of the many (possible) retweets? Shouldn’t Alyssa be the source of those sales as well, as per this experiment?

So, we perhaps need a consistently measurable system that helps us trace the origin of a sale (purchase) to the root, something like a mail trail, so that influencers can be judged better; and not half-heartedly as in the experiment cited.

2. That ‘spreading a conversation is not the same as affecting it’ and ‘ability to disseminate information is different than its ability to influence behavior’ are fascinating thoughts. My question simply is, ‘Why not?’. This stems from my first point, in any case, but why is spreading a conversation not affecting it? Or, why is dissemination not equal to influencing behavior? Is making a piece of information available not a significant part of affecting that information’s worth? Consider the situation when an influenc-ee did not get to see that piece of information – he’d have missed it completely, or at that point in time. So, I believe there is a lot of value in spreading a conversation too – it may not result in ‘affecting’ it as marketers want it (the holy grail!), but if you extend my first point here, the spread is an integral part of the process – a critical one that can eventually lead to a change in behavior.

Similarly, if we do not have adequate systems to trace change in behavior via an information dissemination process, how could we even start differentiating both? To me, dissemination is the first and most important step in changing behavior. The influencer’s message may not change it, but subsequent retweets (if I may simplify it!) and even repetitive retweets at various points, by others, in time, act like placing a billboard in front of an audience again and again. If one of those messages change behavior (say, towards a purchase) and we do not have a way to track that to the source, we shouldn’t be complaining about both being different! They may not be at all and could well be part of a larger process that we seem to be missing given our current measurement limitations.

3. The other thing missing in that experiment is ease of enabling a change in behavior. Digital actions, by default, could be easier than something that requires a person to do things in real life. So, Facebook’s ‘like’, Twitter’s ‘retweet’, Email’s ‘forward’, paypal’s ‘donate’, FourSquare’s ‘check-in’, ecommerce’s ‘add to cart’/’buy’ and so on – they are an extension of what we’re doing online and that ease, coupled with a compelling influencer’s voice could result in them being done faster. It is unfair to compare them to say, volunteering for an activity, or landing up for a concert.

In the experiment cited in the post, the task is strictly online – an Amazon affiliate link to buy the book. So the assumption was that it is relatively easy to order the darn book online and hence people will/may do it if allegedly influential tweeters spread the message. What was perhaps missing was who spreads the message. I bet the sales would have gone through the roof if Oprah had spread the message on Twitter, but again, Oprah’s twitter stream is like a billboard in the heart of NYC – it would come at a cost, most probably. I’m sure Christakis’ conclusions would be different if Oprah was substituted with Alyssa and the book’s title would have simply come alive!

Stock art source: this post from Marcel LeBrun’s Media Philosopher blog.



5 thoughts on “The holy grail of online influence vs. our current measurement limitations

Leave a Reply

Your email address will not be published. Required fields are marked *