Viral marketing - week 4
html text version help for online - opera inline translator
Google Translate v. 7.22 extension for Opera
Viral Marketing and How to Craft Contagious Content
University of Pennsylvania
Lecture 22 - How Social Networks Spread Information and Influence
Lecture 23 - What are Social Networks?
Lecture 24 - How Networks Shape the Spread of Information
Let's begin to think about how networks
shape the spread of information and influence.
People often talk about network effects, but network effects are slightly
different than social networks.
What network effects mean is that
a product or service is more valuable, the more other people that are doing it. So take, for example, a fax machine.
If you were the first person to buy a fax machine, it probably wasn't very useful cuz
there is no one else to send a fax to. In fact, even if there were 100
people using a fax machine, it's still not very useful.
Unless those 100 people were
people that you knew well and were the ones you wanted to send a fax,
it wasn't very helpful. And so the value of the fax machine
depends on the number of other people using it. The same with eBay for example. eBay and other services
marketplaces where buyers and sellers come to meet are the more useful,
the more people that are there.
Does this help or hurt diffusion? Well, at the beginning, it often hurts.
At the beginning of a fax machine,
it's hard to get it off the ground, to get it to catch on, because no one
wants to be the first person to buy it. Who wants to be the first person or
even the 100th person on Facebook?
If no one else is there, it's not very valuable. So, at the beginning, network effects
often slow diffusion, but later on, they can actually speed it up because
once enough people are in there, once enough people are doing something,
it'll catch on even faster. Slow at the beginning,
but helping later on.
But even beyond the number of others, the
pattern of social ties matters as well. As we've started to talk about, social
ties are the bridges that help us get information and influence off an island.
Imagine, for example,
that you live in Singapore and you wanna get a new product
to catch on in the US. You share that information with someone
you know in the US, that'll help. But what if you don't
know anybody in the US? Well, you might go to one of your
friends who knows somebody in the US and try to use them to help spread the word. That particular friend, or the few friends you might have that know
someone in the US, will be very useful. But what if instead you wanna get that
information to someone in London? Well, there, there might be a different
set of of people in your social network that know someone in London and
will help the information get there. It's not just about how popular those
people are, or how many friends they have, position matters as well.
Does that specific person happen to
know someone else in the social network that can get your information there? It's a little bit like bridges between
different disconnected islands. Cars can cross the different bridges,
but only certain cars can go over those bridges, and if there's no bridge between
two islands, it's going to be difficult for the information and
the influence to cross over.
You can see this concretely when you look at smoking over time. Over the number of years, smoking has been decreasing in
its prevalence around the world. And when researchers looked
at people stopping smoking, they found that it tended
to happen in clusters. If you take a picture of a social network
and you look at the number of smokers, it wasn't just that individual
smokers stopped smoking. Whole clusters of people stopped
smoking at similar points in time. If one friend stopped, other people that were connected to
them became more likely to stop. Smokers moved to the periphery of the
network, and there were three degrees of influence, as we discussed at
the beginning of this section. People are not only influenced
by people they knew, but by others they knew who knew
others who knew others, and so on. And so
influence can spread through a network, not just from one person to the other.
One great example of this was a study that
was done a number of years ago called the Small World study. It was done in the early 60s and researchers were interested in seeing
whether information could go from any person in the world to any
other person in the world.
They sent people a number of packages, people who lived in Nebraska and Kansas. And they asked them to get
that package to Boston. If they knew someone, the person that
received the package who lived in Boston, they could send to that person directly. But if they didn't know someone, they were
supposed to send it to whomever they knew who could get the package
there as quickly as possible. Any friend or relative who might
speed the package along its way. And the researcher wondered, well,
would any of the packages get there? And if so, how long would it take? Well, it turned out that many of the
package didn't reach their destination. It's hard to know whether people
weren't motivated to share them or they just didn't know someone. But those that did took about five or
six hops, and that's where we get the famous phrase,
six degrees of separation. Given the number of friends that
people have though, why were there so many degrees of separation? You can imagine, well, if each person has
100 friends, and each of those friends has 100 friends, and so on, that number
should get very big very quickly.
It should take less than six ties to get that package to its destination. So why was it so slow? Well there are two key reasons, the first
we talked about previously is homophily, people tend to be friends
with similar others. But what that means is that they tend
to be friends with a lot of their friends' friends. Well you may have 100 friends, and
your friend may have 100 friends, some of their friends
are the same as your friends. The number of ties don't just multiply. It takes longer for things to get from
one side of the world to the other. And the second is that we don't
always know the fastest route. Just like we talked about previously, it's not easy necessarily to know
what our network looks like. That in part is the power of social media. Now social media technologies like
LinkedIn can tell us who we know who works at a company, or the fastest path
to get that information there. It becomes much faster to
reach particular individuals. But importantly, 48% of completed
chains went through three people. You might wonder, who are these people? Are they more infuential or not? If we had a new product, should we target these people in terms
of getting that product to catch on? And that's what we'll talk about next.
Practice Quiz 4
Lecture 25 - Sprinkler vs. Waterfall Strategies
Say, you've just finished
making a great new website. You've built a contagious message, so that
people will share it, but you only have enough money to diffuse it to a certain
set of your target demographic. Should you spend all your
money targeting one group? Should you spread it out? And within that group,
who should you target and why? Should you pick people that are popular,
maybe that have a lots of friends? Or people that have the right
positions in the social network and might connect you to other
individuals in different networks.
One way to think about this is
the sprinkler or waterfall strategy. A waterfall strategy is when you
concentrate all your marketing efforts in one place, or one region.
If you're the United States, for example,
you might pick a given city to start in. Concentrate all your resources there and hope to spill over into
the nearby cities moving on. A sprinkler strategy is when you spread
out your resources to different areas. Rather than concentrating
100% in one city, you might spend 10% in each
of 10 different cities. Which of those things might
be more effective and why? We can think about reasons for both. The benefit of the sprinkler
strategies are spreading it out, you're starting many seeds in
many different communities. That could be good, because it
could get it to catch on faster in different networks, but
a waterfall strategy could also be good if people need multiple doses of
influence before they adopt something. And so one way to think about whether
to use a sprinkler strategy or a waterfall one is to understand
whether that product, that you're hoping to catch on is a simple or complex contagion.
A simple contagion is something that you only need one dose of influence before you're willing to adopt. Take, for example, a newspaper article that someone might send to you online.
You didn't need to hear about that article from five or six people before you're
willing to open the link.
One person's enough for you to check it out. A movie takes a couple more doses of influence. You might not just go see a movie, because one person said something about it.
But if a few people say something about it, you'll check it out. And on a much more complex
end of the spectrum say, someone suggested a new open heart procedure.
You wouldn't just need one person to tell
you about it before you'd try it, you wanna do lots of research before you're willing to try that risky procedure.
So complex contagions require more doses
of influence before you're willing to adopt, whereas simple
contagions take fewer.
Take, for example, a pickled pickle. You only need one dose of influence
to try a brand of pickles.
But if someone said try pickled broccoli, you might need to hear about it a few people before you're
willing to try it out.
So again, simple contagion are things
that require only one dose of influence, whereas complex require multiple doses
before people are willing to try it. Complex things tend to be more costly, whether in terms of time,
effort or energy. The more money you have
to spend doing something, the more you wanna gonna hear
about it a few times before do it.
Well, simple versus complex contagions
have an important implication for whether you should use a waterfall
of a sprinkler strategy. When you have a complex contagion, you
need to use more of a waterfall strategy. When people need multiple doses of
influence before they're willing to adopt, they're gonna need to hear about that
product or service for multiple people before they feel comfortable,
which means you need to make sure that multiple of their peers have heard
about it to get them to change their mind. That suggests you need to concentrate your
resources on geographic, demographic or social network region online. Making sure that indi people and individual area a subset of a network
have heard about it multiple times. If you've got a simple contagion, if
people only need one dose before they're willing to do it, well,
then you can spread your resources out. People don't need to hear
about it multiple times. And in fact, multiple times will waste
some of your marketing resources. It's much better to spread them out, so
that each person hears about it once. They hear about it from one other person
that causes them to adopt it and that increases the likelihood it will catch on
more quickly in the broader community. The more doses required for
product adoption, the more you wanna concentrate
your resources in one area. One geography, one social group or
one set of people that has similar tastes.
Lecture 26 - Social Ties and Active Sharing
Imagine you're trying to get a new job. You've sent a bunch fo resumes out and
you've gone online. You've done everything you can, but you still haven't found
something you're happy about. So you start working your social network. Who do you think'll be most
useful to getting you a job, someone you know well or someone you know less well?
It turns out that all our social ties
don't find out about something when we do. In fact, information doesn't flow
like water through an open channel between a social network. It requires active sharing. And so rather than thinking about social
ties as bridges between islands, it's better to think of them as drawbridges for
sharing information that can be either open or closed, depending on whether
someone wants to pass something on. But importantly the flow also
depends on the type of ties. There are two different types,
strong ties and weak ties. How many people do you think you
talk to at least once a week every week of the year?
Well, it turns out you talk to about ten people, and these people are considered
your strong ties. You can think about
the same thing on Facebook. People have on average about 130 Facebook friends. By how many they interact with regularly? Well, by analyzing the photos and looking
at who people show up with repeatedly over time, researchers find you show up
between four to six friends repeatedly.
Again, those are your strong ties. If you do this same thing with phone records, if you take a look at different
people's phone records, you see that 80% of calls are to four people in their social network. And in all these cases these people
are strong rather than weak ties. Strong ties are people you talk to often, you've known for a long period of time. They could be good friends,
family members, or other people that you know quite well.
Weak ties tend to be more casual relationships. People you talk to infrequently,
acquaintances, or someone you bump into at the water cooler.
So which of those types of ties would be more effective in helping us get a job? The strong ties, the people we know really well or the weak ties the people we don't know as well?
Well, a number of years ago a sociologist studied exactly this. He looked at how people get jobs and he found that a lot of times people get jobs through their social ties.
But he was interested
in more than just that. He was interested in what type of ties
people tend to get their jobs through. You might expect that our strong ties will
be more useful in helping us get a job.
They know us much better and they care about our interests more. Take your parents for example. Your parents know a lot about you and they wanna do everything they
can to help you get a good job. In fact if you ever been set up on a date. You know that your parents are very
motivated to help you get things done in your life. You may not expect that our friends, the people that know us well
are similar to family members. They want to help us get a job and so they work hard to make it happen. You might think our weak ties could sort of care less. They don't know us as well and they don't
care as much as about our interests. They shouldn't be very helpful in helping us get a job.
But if you dug further, scientists realized that weak ties
can actually be quite useful. In an article titled, the Strength of Weak Ties, he found that weak ties are actually much
more helpful in getting people a job.
And the reason, well counterintuitive, makes sense once you hear it. Weak ties have access to a lot different
information than our strong ties do.
If you remember that idea of homophily, the fact that we tend to be friends with similar others. Our friends tend to know us very well, but they also tend to pull from
the same wells of information. Professors tend to hang out with
professors, accountants tend to hang out with accountants, and art historians tend to hang out with other art historians. But because of that, the information
they get is often redundant. Your friends, your colleagues,
the people that know you very well have access to the same information that you do. Whereas, your weak ties tend
to be different from you. You don't know them as well, but they tend to drill in different wells of information. And because of that, they have access to
information you might not have already. And as a result, they can be much more
effective in getting you a job And second, simply put,
there are more of them. There are so many more weak ties that we have than strong ties. That even if each weak tie on average
has a lesser effect per individual, because there are more of them they add up to have a big impact.
Different types of ties not only shape whether or not we get a job. They also shape what we wear. Think for a moment on what we share with
our stong ties versus what we share with our weak ties. What we share with people that we know
very well versus what we share with people that we don't know as well. It turns out that if we
don't know people as well, if someone is a weak tie,
we're not gonna share as much information with them that we might
share with a strong tie. If we don't know someone as well we are a little more worried about how we might look, how sharing a particular piece of information might make us come off. So we tend to share things
that make us look good and tend to avoid sharing things
that make us look bad. We don't share that embarrassing
news story we read online or we don't tell that story about how we
screwed up or picked a bad restaurant. Whereas with our strong ties we
are more willing to be honest. We are more willing to disclose
embarrassing things that we might not want those weak ties to know. Because strong ties know what's better,
we're less worried how this one piece of information we share with
them might make us look. But that is an important implications for getting information off violence because
weak ties tend to bridge between islands that become very important
in getting things to catch on. If people just a product or service among
their strong ties, they're never gonna spread to different groups that might
not be connected to that initial group. We need to make sure that weak ties are
willing to carry the information to make sure it moves between different islands. While word of mouth may carry buzz and understanding the science behind word of
mouth may help us make our messages more contagious, where messages
go depend on the network. We need to understand how
network structures and types of ties shape the spread
of information and influence. And how that drives why all
sorts of things catch on.
Lecture 27 - Course Conclusion
Looking at the world around us,
we see all sorts of things catch on. It might be a product at the grocery
store, it might be a service at work, or it might be a behavior at home that
we've never seen our kids doing before. Why do some things succeed and
others fail? And how by understanding that science
can we get our own products, ideas, and behaviors to be more successful? Whether you work at a for profit or
a nonprofit, a large business or a small business, whether you just
wanna be more successful at home and at work,
these tools will help you do that. We've talked about how to make ideas
stickier, how to gain more influence, how to leverage the power of social networks,
and how to generate more word of mouth. It's not random luck or chance why some
things become popular and others fail. There's a science. By understanding that science we
can craft contagious content and we get our products and
ideas to be more successful. It was wonderful teaching you and I hope
you enjoy applying these exciting ideas.