1. CONNECTORS, NETWORKS and INNOVATION SUCCESS
by Werner Iucksch
It may not look like it, but one of the most controversial concepts around social
networks is that of the importance of individuals that became known as
“connectors”. In “The Tipping Point”, Malcolm Gladwell argues that this special
kind of people are able to pick‐up ideas, values, trends from one cluster of people
and transport it to another, spreading ideas in society. As connectors are linked
to a high number of people, they could initiate big movements of behaviour
In “Linked”, Albert‐László Barabási writes that his research found that some
places in cyberspace (that he calls “hubs”) concentrate a vast amount of links and
traffic of information. The author suggests that this result reinforce the theory of
the “power of connectors”.
This would confirm a golden piece of information to marketers. Imagine the
economies that would be possible by marketing to a selected group of people,
subsequently letting them use their networks to spread the news and drive
adoption of products or services.
This concept is now widely accepted around the business world, Procter and
Gamble even sells the access to such group of people, but if that’s what it takes to
spread the word and drive adoption of new product/services, why most of them
continue to die off?
One of the most respected network researchers, Duncan Watts, from Columbia
University, is also interested in this question. He found some interesting data and
formulated some hypotheses on why that happens, when he conducted an
2. enhanced version of Stanley Milgram’s “Small World experiment” (the one that
argues that any two people in the world are separated by no more than 6 other
After starting more than 60,000 e‐mail chains to find 18 targets around the
globe, Watts noticed that only 5% of the messages passed through highly
connected people. From this data it is possible to hypothesize that although
some people may have many more connections than others, it doesn’t mean that
they have much the power to influence behaviour, for they are not relied upon
for given tasks. Of course this hypothesis had a great impact:
If this is the case, it’s fair to say it takes more to the adoption of new
ideas/products/services than simply having them spread around by connectors.
Innovation diffusion is a new science, but some studies point the way. It’s helpful
to think about “critical thresholds”, a concept that basically estates that
contagion of individuals with ideas will only happen if a certain threshold of
influence is achieved.
For example, someone may begin to use a product because it is interesting,
someone else close to him notices that it solves a problem s/he has. A third
friend notices that two of his colleagues are using it and is no longer shy to try
the innovation. This way they may build an “infected social cluster”, which can be
connected to other clusters (groups of people they know) in such a way that it
will convert enough individuals to cascade the usage of the product to more and
more clusters, until it’ll get to a point where critical mass is achieved to unfold a
global infection (or global adoption of the innovation).
This is, however, an extraordinarily difficult process to manipulate or to predict.
As we hear so often, movie studios still invest millions of dollars in projects that
don’t take off (e.g. Waterworld, Battlefield Earth), music history is full of groups
heavily promoted that never lived up to the “hype” (e.g. any one hit wonder) and
books as successful as Harry Potter get rejected several times before being
published. It can be argued that professionals involved were not so competent,
made bad decisions, but there’s an additional component that is a pretty cruel:
Another experiment is very telling of our social nature. Some years ago, aprox.
14,000 people were recruited to evaluate a number of songs from unknown
musicians via a website, downloading whichever they wanted. Half of these
people only had information about the name of the band and the song, not being
able to have any interaction with each other (thus each person was not
influenced by anyone else). The other half was sub‐divided in 8 groups. In
addition to the name of band and song, each person in a sub‐group had access to
a real‐time “quality rating” and number of downloads of each song by the other
sub‐group’s members (so there were eight independent social sub‐groups, each
with their own rankings, quality levels and number of downloads.)
If the publishers that rejected Harry Potter were so incompetent, the book’s
quality should be evident to anyone since the first draft. Likewise, every group in
the experiment should have evaluated each song similarly, given the size of the
sample. However, the song ranking of the first half (the one that only knew the
name of the bands and songs) was radically different from that of the sub‐
The sub‐groups results were yet more interesting. They were remarkably
different amongst themselves. The top song in one group was the 40th in another
group and so on. The reason: the top song in a group was well evaluated early in
the process and as other respondents were able to see the ratings of other
people, they ended up being influenced by the opinions. In other groups, another
song was an early leader, so they had a different result. This is something that is
observed in other areas of science and is known as “Power Law”, or as the
expression goes: “the rich, get richer”.
Well, if connectors do not equal success and if it is so hard to predict the
behaviour of networks, what can marketers do?
4. From what was learned, speed, control and flexibility seem to be essential. One
idea to take advantage of them is to launch many products/services at the same
time, measure their adoption in the first moments of their lives. As soon as it’s
possible to identify which one has most potential to spread, rearrange the budget
in benefit of this option, making people aware of it, facilitating its growth. Also
check if some other product/service is doing well within important niches,
calculate its potential and invest accordingly. Terminate the rest.
In communication, rather than having a big launch of one finalised version
campaign, it may be the case of spreading several different messages and
interaction possibilities, with common values and strategic direction, in a smaller
scale. After some time, evaluate how each one is doing (WOM, sales, brand value,
etc.) and invest heavily in whatever is considered more important at the
moment. The concept is similar to that of beta testing many versions of software
before rolling out a full commercial version.
If this is done, connectors can be useful. Although they are not necessarily
influential to everyone, they can help to seed a variety of messages quite fast.
With a variety of messages, it is more likely that there will be some that are
relevant by themselves (without the need of a connector vouching for it), gaining
space because they reach the appropriate audience and achieve critical mass fast
enough and will live longer, because resources will be reallocated to keep them
This is an idea. Science still is learning how networks operate and businesses are
still struggling to make the most of social networks and digital platforms. There
are no proven models yet; there is more in developing a successful
product/service than speed, flexibility and control; the way brands are thought is
changing with the behaviour of individuals. There are lots of threats, as well as
lots of opportunities. What an exciting time to work with communications.