DT&SC 9-2: Diffusion of Innovations through Social Networks - YouTube

" How do innovations diffuse in and among societies?

The basic body of literature is referred to as “The Diffusion of Innovation”

• Everett Rogers (one of the most cited communication scholars) talks about the Diffusion

of Innovation

There are stages for the Diffusion of Technology

1.

The knowledge Stage- you get to know the technology 2.

Persuasion Stage-You are persuaded with the benefits of the technology and also acknowledge

the cons.

This also can be a very social process in some sense because you see others adopting

it and embracing the technology, so you start to evaluate and research it

3.

Decision Stage- Decide to adopt the technology 4.

Implementation Stage- Start to use the technology on a regular basis

5.

Confirmation Stage- You recognize the benefits of using the innovation and use it on a daily

and you start to promote the innovation to others.

Directly or indirectly • As a result of this interaction of the

individual process of adoption and a social.

If you take a bird’s side view- you start to see that these interesting wave like diffusions

of these innovations where some innovations run through these four stages really on, these

are innovators.

Other ones run through second, early adopters, etc.

• From this bird’s eye view you can really see this wavelike process.

CASE STUDY EXAMPLE: Leadership Lessons from Dancing Guy by Derek Sivers

• You have an innovator dancing in the midst of a group of people sitting.

Then you see the first follower join him and embrace him.

• The leader has the follower become an innovator himself.

• They are both innovating together or they both can be two lone wolfs

• Then a third person joins, he is also innovating

• Three is a crowd and now we have these three innovators, that would be considered

early adopters and we begin to see more early adopters and then you have more and more people

join.

• These early adopters quickly grow into a mass and then we start to see the tipping

point…

• People are now hurrying to join and become adopters because they want to be part of the

in crowd.

At this point people start to migrate to this early majority and more people want to join

the innovation and then we start to see how quickly it continues to grow.

• Originally the diffusion was slow.

People were joining very slowly but then all of a sudden, you see that more and more people

are joining extremely quickly.

• Once the innovation is good, it reaches more than half of the entire society or the

people siting on the hill • Once this happens, you have reached half

of society, the rhythm might slow down a little bit and diffusion might be a little slow again

but eventually it reaches all of society because most people are already part of the movement.

• Some people may still refuse to join and may never join but still it takes over society

and it becomes a normal practice.

• In this video we saw the innovation diffuse through society in a very social process,

which is exactly how the diffusion of innovation happens.

IDEALIZED DIFFUSION OF INNOVATION THROUGH SOCIAL NETWORKS

• Where does this diffusion wave come from?

• It comes from the fact that innovations diffuse through social networks.

• We start with two innovators: then we see how two people or innovators were infected

with a disease.

• We start with two people who were infected with this disease (cumulative adopters) out

of 100, the “adopters”, meaning that 98 people are not infected or have not adopted

the innovation, Non Adopters.

• We see the Rate of Adoption is slow at 0.01 or 1%

• We see that these 2 people have infected 2 more people (19.6) so you have a total of

4 New Adopters • So we assume that stays at a constant

rate of infecting people (Rate of Adoption) of 0.01 or 1%.

• However, these 4 people have now infected 2 more people and now there are 8 people infected

or there are 8 New Adopters.

• And this goes on and on and once you reach 100 everyone is infected or everyone has adopted

this technology and then these 100 people are infecting 0 people even if its at the

same 0.01 rate because everyone is already infected.

• You can look at them number of people that are infected in each round or the cumulative

nature of this diffusion.

• The curve on the graph basically measures the number of people that have adopted the

technology at each New Adopter period.

• The red curve you see starts with very few people, then you reach a peak (24.94)

with around 25 people adopting the new technology and then it goes down again.

This is because at this peak you have already infected 50% or half of society and at this

point it can only go down again because you have no ore than 100 and that is why it slows

down because once it reaches everyone or all of society, no one else can be infected.

In this case, the innovation diffuses completely throughout society.

• You can also look at the new adopters at each stage on the blue line and then you

can keep the line going upwards until it reaches 100.

• The reason there are two curves is because there are two ways of presenting this trend.

• The blue line represents the Cumulative nature of the diffusion while the red line

represents how the line will look if you graph it from the New Adopters column on the graph.

• These are the two ways of looking at it.

One is, how many new adopters are added at each point.

(Red Line) • Or you can look at it as how many cumulative

people you have at each point.

(Blue Line) "6:00-11:59: ...kind of like effect or convince,

persuade that another 98% is a very stable rate of 1% that means that in fact two new

people adopt this innovation.

So in total we have now four people, right?

The two earlier ones and now the other ones.

The math doesn't work our.

Now these four in fact the remaining 96 people of our network also had a constant rate of

1%, we assume a very simple constant rate.

That means that now four people are infected.

In total you have 8 people.

I want you to notice how that now increased from infecting two to infecting four.

Now you start with almost 8 people and in the next draw you would infect 7 more people.

So it increased from two to four to seven and the total numbers of adopters if i sum

them up and you have two different ways of looking at it now.

You can look at the number of people that are infected in each round or at the cumulative

nature.

At the end you will reach 100 when everybody is infected and has adopted that technology.

So let's look at this curve here.

This curve where you measure the number of people that adopt the technology in each period.

You get this red curve here, basically a hump like this, you start with very few you have

two you have four then you have seven and then you increase to 12.

And then you reach a peak here with 25 people adopting it and then it goes down again.

why does that happen?

It happens because at the point of this peak you have already infected, you have persuaded

50% of society.

Half of that can only go down because at the end you cannot have more than 100 so at the

end when you reach 100 no additional people can be added if the innovation diffuses completely

throughout the entire society.

And that's why it only can go down here.

Now you can also look at it this way and add up the new adopters at each stage and get

this cumulative logic which results into this curve here which results very early, 2,4,

7 and so forth and then you reach 100 here.

So this is one hundred percent and you cannot go further than this.

So this is why you get these two curves.

I don't want you to get confused with this.

you have two ways of presenting this.

One looks at this column here and the other looks at cumulative logic of this column here.

So I say it again, you have two way of looking at it.

One is you ask how many new adopters are added at each point.

So you start with 2,4 and so forth--you reach a maximum.

And at the end you cannot add anyone anymore because everybody has already adopted it so

this needs to be zero again.

And in the other question you look here at this blue column and you ask how many people

have already adopted it since the beginning.

And here at the end you have to reach 100.

because at the end 100% in this idealized case have adopted it.

Social networks usually never come in such a regular form of a grid.

What makes social networks interesting is that they have an intricate structure to them

that we can then study.

For example here you see a network that has been used to study the spread of a disease.

Contagious effect which work very similar to the logic that we talk about when we talk

about the diffusion of innovation.

You can see here separate clusters which are important to realize because then you can

see if somebody infects other people.

You have another network that has also been used in this study.

In a health study the spread of obesity and the fact that some obese people you can find

that they are in groups together.

Non obese people are also in the same groups together.

You can see this intricate network structure that surely affects the diffusion process.

Here we have another online community.

Also another online structure.

And here you see the entire Internet which is also a very interesting structu.

And of course and nowadays plays a big role in the diffusion of information in the persuasion

stage and therefore also in the diffusion of innovation.

Let's look at a little toy model that gives us a better insight in how this diffusion

through social networks actually works.

Let's us the agent-based computer simulation software in order to simulate the diffusion

process.

Here I work with one hundred nodes on average everyone of the 100 nodes has five connections

and we have two people that are infected.

If you talk about the spread of a disease it's infected in our case these would be innovators.

They already adopted this technology.

And they spread this technology with an average rate of about 1%.

So now we can start and go and we can see at the beginning it's spreads very slowly.

So this one convinces this one this one convinces two.

and now you can see the rhythm starts to increase.

These here are the number of people that have already adopted, the number of infected people.

And this here is the people who have not adopted yet.

It's just 100 minus the red one.

You can see now the innovation spreads more and more.

Now you reach 50% SO now half of the society has already adopted.

And after this the diffusion process has to slow again.

Because there are not as many people around and there are these people in this corner

here that still have to be reached.

It takes a long time for this innovation to reach every corner of society.

Whereas in the middle it was very quick there were many people accessible.

"12:24-18:06 (end): You can do that with a different network as well, here, you can just

randomly create several networks, and then let's start running.

Sooner and a little bit faster, you have an innovator here, you have an innovator down

here, also depends where the innovators are.

And then the innovation is different patterns of diffusion depending on where innovators

are at the beginning and it depends on the network structure.

Now here we needed about 700, 800 time steps for the innovation to diffuse through the

entire society.

What happens now if we, for example, increase the density of the network?

So here now we have a more dense network, even like this, we have an even denser network.

What do you think will happen with regard to the diffusion of the innovation?

Well, let's check it out, let's press go here and let's see and...well!

What happened?

Okay, let's decrease the number of nodes we had now here.

Beginning, we had five, we have thirty, let's go down and make a very sparse network, for

example this one here and we have the-- your two innovators here and here, and let's see

what happens.

What happened now?

The innovation couldn't even spread to everybody because the network is too sparse; there are

not enough connections, so, uh, two isolated groups here and they could never be reached,

we could keep it on running and they will never-- innovation will never jump over there.

So, the diffusion logic depends a lot on the network structure, with the network density

- the number of connections per node - being one very important characteristic.

And you can notice that if you look at empirical difficult curves.

So here we have the diffusion of different technologies among U.S. households from the

1900s until 2005, and you have different technologies, here for example you have the car, it's a

diffusion curve that looks like this - in reality, of course, these curves are not as

pretty as in our very simplistic models - here you have electricity that spreads, here you

have the fixed line phone - you see it's spreading - here you have the clothes washer, and here

you have color TV, you have the microwave, you have the cell phone, so you have computers,

what do you notice about the shape of the curve, over time?

The shape always becomes steeper, so that has to do with the general tendency that social

connectedness has increased; nowadays a viral cat video can spread to millions and millions

of people in a couple of minutes because we are very highly connected, so the level of

connectivity is one important characteristic, um, that leads to the faster diffusion of

innovations.

But network density is not the *only* characteristic that affects the diffusion of innovation,

for example, here, we have a simulation by Lada Adamic also in the software - simulation

software NetLogo - and we now create a completely-- what we call a random graph, a random network,

that means that the number of links are randomly distributed among the number of nodes, and

we start with one innovator and now spread, uh, the disease.

So, we start here and you can see a different diffusion logic as the innovation diffuses

through this very intricate network structure that we randomly created which then results

into a different diffusion pattern, and can take a long time, it depends on if you-- if

you happen to have these kinds of different clusters over here and over here and it depends

on where the innovator is, and the result is a diffusion curve that has certain characteristics

that are not as-- not such a pretty 'S' form, but nevertheless you can still see a similarity.

Now we can create another network here which is called a preferential attachment network;

in this kind of network, a scale-free network, you have some very centralized nodes, for

example this one here or this one here or this one here, and you can see that this also

affects the spread of the innovation because once you reach one of these big nodes, it's

kind of like the innovation explodes and is catapulted towards the rest of the network,

and then you get this logic where nothing happens and then suddenly, whooom you get

this-- really transition in the-- in the diffusion pattern, you get a different shaped S-curve.

As another example, here you see the diffusion of *legal* innovations among the fifty states

in the United States of America in a nice animation by Bloomberg.

And you can see here, for example, the diffusion of interracial marriages, and you see here

in 1700 and 80, 90, only very few states, less than ten states have adopted it, and

you see it takes a very long time until this legal innovation diffuses throughout the United

States; until 1950s there were less than twenty states, and if you compare that with some

more recent legislation, for example same-sex marriage you can see that it reached the same

level of diffusion in a much shorter time.